Add Phase 4: gNMI streaming telemetry and traffic generator

- gNMI integration: NETCONF script to enable gRPC on all 9 routers, Telegraf container with gnmi input plugin, InfluxDB for time-series storage, 3 Grafana telemetry dashboards (utilization, errors, combined) - Traffic generator: Scapy-based dual-mode container (sender/responder) with Flask API, RFC 2544 test suite (throughput, latency, frame-loss, back-to-back), Vue 3 web UI with flow builder, test runner, real-time stats monitor, and results export - docker-compose.yml updated with influxdb, telegraf, traffic-gen, traffic-gen-ui services - Full documentation in DOCS.md sections 15-16 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-06 15:29:44 -07:00 · 2026-03-06 15:29:44 -07:00 · dcebf15bb3
commit dcebf15bb3
parent f23e222bc0
44 changed files with 3848 additions and 3 deletions
--- a/DOCS.md
+++ b/DOCS.md
@ -16,6 +16,8 @@
 12. [Troubleshooting](#12-troubleshooting)
 13. [Data Retention](#13-data-retention)
 14. [Environment Variables Reference](#14-environment-variables-reference)
+15. [gNMI Streaming Telemetry (Phase 4)](#15-gnmi-streaming-telemetry-phase-4)
+16. [Traffic Generator (Phase 4)](#16-traffic-generator-phase-4)


 ---
@ -28,7 +30,7 @@ This is a **BGP Monitoring Platform (BMP) lab stack** deployed via Docker Compos

 - Receives BMP (BGP Monitoring Protocol, RFC 7854) telemetry from routers on TCP port 5000
 - Streams BMP data through Kafka into a TimescaleDB/PostgreSQL database
- Provides **27 Grafana dashboards** (17 operational + 6 learning + 4 advanced analytics) for real-time and historical BGP analysis
+- Provides **30 Grafana dashboards** (17 operational + 6 learning + 4 advanced analytics + 3 streaming telemetry) for real-time and historical BGP analysis
 - Includes an **ExaBGP route injector** that peers with the two CORE routers and injects synthetic BGP routes, enabling testing of BGP policy, route propagation, and Grafana dashboards without needing internet connectivity
 - Provides a **Vue 3 web UI** at `:5001` for point-and-click scenario management, live route tables, and peer monitoring

@ -64,7 +66,7 @@ IOS-XR Routers (9x, AS 65020)
    PostgreSQL 14 + TimescaleDB
         |
         +---------> obmp-grafana (grafana/grafana:9.1.7)  :3000
-         |              23 dashboards, PostgreSQL datasource
+         |              30 dashboards, PostgreSQL + InfluxDB datasources
         +---------> obmp-whois (openbmp/whois:2.2.0)      :4300
                       WHOIS query server backed by the DB

@ -73,6 +75,24 @@ ExaBGP (obmp-exabgp, built locally)
  Peers eBGP to CORE-01 and CORE-02 (AS 65100 -> AS 65020)
  HTTP API on :5050 — inject/withdraw routes on demand
  Routes propagate via iBGP mesh to all 9 routers -> BMP -> DB -> Grafana
+
+gNMI Streaming Telemetry (Phase 4):
+  IOS-XR Routers (gRPC :57400)
+         |
+         v
+  obmp-telegraf (telegraf:1.28 + gnmi plugin)
+         |
+         v
+  obmp-influxdb (influxdb:2.7)  :8086
+         |
+         v
+  obmp-grafana (InfluxDB datasource -> Telemetry dashboards)
+
+Traffic Generator (Phase 4):
+  obmp-traffic-gen (python:3.11 + Scapy + Flask)  :5051
+    Dual-mode: sender (generate traffic) / responder (echo/log)
+    RFC 2544 testing, custom packet flows
+  obmp-traffic-gen-ui (Vue 3 + NGINX)  :5002
 ```

 ### Container Summary
@ -87,7 +107,11 @@ ExaBGP (obmp-exabgp, built locally)
 | obmp-grafana | grafana/grafana:9.1.7 | 3000 | Visualization |
 | obmp-whois | openbmp/whois:2.2.0 | 4300 | WHOIS query server |
 | obmp-exabgp | local build | 5050 (host net) | BGP route injector |
-| obmp-exabgp-ui | local build | 5001 (host net) | Vue 3 web control panel |
+| obmp-exabgp-ui | local build | 5001 (host net) | Route injector web UI |
+| obmp-influxdb | influxdb:2.7 | 8086 | Time-series DB for telemetry |
+| obmp-telegraf | local build | - (host net) | gNMI telemetry collector |
+| obmp-traffic-gen | local build | 5051 (host net) | Scapy traffic generator |
+| obmp-traffic-gen-ui | local build | 5002 (host net) | Traffic generator web UI |

 ---

@ -830,3 +854,215 @@ Adjust in `docker-compose.yml` under the `psql-app` service environment block.
 | Variable | Default | Description |
 |----------|---------|-------------|
 | `EXABGP_API` | `http://localhost:5050` | ExaBGP API base URL |
+
+---
+
+## 15. gNMI Streaming Telemetry (Phase 4)
+
+### Overview
+
+gNMI (gRPC Network Management Interface) adds **data-plane visibility** alongside BMP's control-plane monitoring. Telegraf collects real-time interface counters from all 9 IOS-XR routers via gNMI subscriptions and stores them in InfluxDB. Grafana queries InfluxDB for telemetry dashboards.
+
+### Architecture
+
+```
+IOS-XR Routers (9x, gRPC port 57400)
+         |
+    gNMI subscriptions (10s sample)
+         |
+         v
+  obmp-telegraf (telegraf:1.28 + gnmi input plugin)
+    host networking → reaches routers on 10.100.0.x
+         |
+         v
+  obmp-influxdb (influxdb:2.7, port 8086)
+    bucket: "telemetry", org: "openbmp"
+         |
+         v
+  obmp-grafana (InfluxDB datasource, Flux queries)
+    3 dashboards in OBMP-Telemetry folder
+```
+
+### Enabling gRPC on Routers
+
+The routers need gRPC enabled before Telegraf can collect telemetry. A NETCONF script is provided:
+
+```bash
+# From the host (requires ncclient: pip install ncclient)
+cd /home/user/obmp-docker/gnmi
+python3 gnmi_grpc_config.py
+```
+
+This connects to all 9 routers via NETCONF (port 830, credentials webui/cisco) and pushes:
+```
+grpc
+ port 57400
+ no-tls
+```
+
+**Verify on router:**
+```
+show grpc status
+```
+Expected: gRPC listening on port 57400.
+
+### Telemetry Data Collected
+
+Telegraf subscribes to two IOS-XR YANG paths at 10-second intervals:
+
+| Subscription | YANG Path | Data |
+|-------------|-----------|------|
+| interface_counters | `Cisco-IOS-XR-infra-statsd-oper:infra-statistics/interfaces/interface/latest/generic-counters` | bytes/packets in/out, errors, drops, CRC |
+| interface_rates | `Cisco-IOS-XR-infra-statsd-oper:infra-statistics/interfaces/interface/latest/data-rate` | bits/sec in/out, packet rate |
+
+### InfluxDB Access
+
+- **URL:** `http://localhost:8086`
+- **Org:** `openbmp`
+- **Bucket:** `telemetry`
+- **Token:** `openbmp-telemetry-token`
+- **Retention:** 30 days
+
+### Grafana Telemetry Dashboards
+
+Three dashboards in the **OBMP-Telemetry** folder:
+
+| Dashboard | UID | Description |
+|-----------|-----|-------------|
+| Interface Utilization | obmp-telem-01 | Input/output bytes rate, packets rate, top interfaces by throughput |
+| Interface Errors | obmp-telem-02 | CRC errors, input/output errors, drops, overruns |
+| Combined BMP + Telemetry | obmp-telem-03 | Mixed datasource — BGP peer status (PostgreSQL) alongside interface counters (InfluxDB) |
+
+All dashboards have `$router` and `$interface` template variables for filtering.
+
+### Troubleshooting gNMI
+
+```bash
+# Check Telegraf logs for gNMI connection status
+docker logs obmp-telegraf --tail 50
+
+# Verify InfluxDB has data
+curl -s -H "Authorization: Token openbmp-telemetry-token" \
+  "http://localhost:8086/api/v2/query?org=openbmp" \
+  --data-urlencode 'q=from(bucket:"telemetry") |> range(start: -5m) |> limit(n:5)'
+
+# Check InfluxDB health
+curl http://localhost:8086/health
+```
+
+---
+
+## 16. Traffic Generator (Phase 4)
+
+### Overview
+
+A portable, containerized traffic generator with a web UI for RFC 2544 testing and custom packet flows. Built with Scapy + Flask (backend) and Vue 3 + NGINX (frontend). The container supports **dual-mode operation**: sender (generate traffic) or responder (receive/echo packets).
+
+### Accessing the UI
+
+- **Web UI:** `http://localhost:5002`
+- **API:** `http://localhost:5051`
+
+### Dual-Mode Operation
+
+Set via `TRAFFIC_GEN_MODE` environment variable in `docker-compose.yml`:
+
+| Mode | Description |
+|------|-------------|
+| `sender` (default) | Generates traffic, runs RFC 2544 tests, sends custom flows |
+| `responder` | Listens for incoming test packets, echoes/timestamps them, reports receive stats |
+
+**Typical deployment:** One instance as `sender` on the host, optionally a second instance as `responder` on another endpoint. Without a responder, the sender uses ICMP echo for latency measurement (routers respond natively).
+
+### Creating Flows
+
+Use the **Flow Builder** panel (left sidebar) in the UI:
+
+| Field | Default | Description |
+|-------|---------|-------------|
+| Name | - | Human-readable flow name |
+| Destination IP | `10.100.0.100` | Target router IP |
+| Source IP | `10.40.40.202` | Host IP |
+| Protocol | UDP | UDP, TCP, or ICMP |
+| Source Port | 50000 | (UDP/TCP only) |
+| Destination Port | 5001 | (UDP/TCP only) |
+| Frame Size | 512 | Packet size in bytes |
+| Rate (pps) | 1000 | Packets per second |
+| Duration | 30 | Seconds (0 = infinite) |
+| DSCP | 0 | Differentiated Services Code Point |
+
+After creating a flow, use the **Flows** tab to Start/Stop/Delete flows.
+
+### RFC 2544 Testing
+
+Use the **Tests** tab to configure and run RFC 2544 tests:
+
+| Test Type | Description |
+|-----------|-------------|
+| **Throughput** | Binary search for maximum zero-loss forwarding rate |
+| **Latency** | Measure round-trip time at determined throughput rate |
+| **Frame Loss** | Loss percentage vs. offered load curve |
+| **Back-to-Back** | Maximum burst length at line rate with zero loss |
+
+**Parameters:**
+- **Base Flow:** Select a previously created flow as the test template
+- **Frame Sizes:** Standard sizes: 64, 128, 256, 512, 1024, 1280, 1518 bytes
+- **Trial Duration:** Per-frame-size test duration (5–300 sec)
+- **Max Rate (pps):** Upper bound for binary search
+- **Acceptable Loss %:** Threshold for pass/fail
+
+### Quick Presets
+
+Six built-in presets are available in the **Tests** tab:
+
+| Preset | Description |
+|--------|-------------|
+| quick_icmp | ICMP ping to CORE-01 at 10 pps |
+| udp_flood_small | 64-byte UDP at 5000 pps |
+| udp_flood_large | 1518-byte UDP at 1000 pps |
+| rfc2544_throughput | Full throughput test with standard frame sizes |
+| rfc2544_latency | Latency measurement with standard frame sizes |
+| tcp_session | TCP flow at 500 pps |
+
+### API Reference
+
+| Method | Path | Description |
+|--------|------|-------------|
+| GET | `/healthz` | Health check + engine status |
+| GET | `/interfaces` | Available network interfaces |
+| GET | `/mode` | Current mode (sender/responder) |
+| GET/POST | `/flows` | List / create flows |
+| GET/PUT/DELETE | `/flows/<id>` | Get / update / delete flow |
+| POST | `/flows/<id>/start` | Start sending |
+| POST | `/flows/<id>/stop` | Stop sending |
+| GET | `/flows/<id>/stats` | Real-time stats for a flow |
+| GET/POST | `/tests` | List / create RFC 2544 tests |
+| GET | `/tests/<id>` | Test details + results |
+| POST | `/tests/<id>/start` | Start test execution |
+| POST | `/tests/<id>/stop` | Abort test |
+| GET | `/tests/<id>/results` | Exportable results |
+| GET | `/presets` | Available test presets |
+| POST | `/presets/<name>` | Create flow + test from preset |
+| GET | `/stats/history` | Stats ring buffer (300 samples) |
+| GET | `/responder/stats` | Responder-mode receive stats |
+| POST | `/responder/reset` | Reset responder counters |
+
+### Integration with gNMI Telemetry
+
+The key value of combining the traffic generator with gNMI: **send traffic while watching real-time interface counters**.
+
+1. Create a UDP flow targeting a router (e.g., R9K-01 at 10.100.0.1)
+2. Open the Grafana **Interface Utilization** dashboard, select that router
+3. Start the flow — gNMI counters show traffic appearing on the interface
+4. Run an RFC 2544 throughput test — Grafana shows the stepped traffic pattern from binary search iterations
+5. Compare Scapy-reported stats with gNMI-reported counters for cross-validation
+
+The **Combined BMP + Telemetry** dashboard shows both control-plane (BMP BGP updates) and data-plane (gNMI interface counters) side by side, enabling correlation of BGP changes with traffic impact.
+
+### Environment Variables
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `TRAFFIC_GEN_API_PORT` | `5051` | Flask API listen port |
+| `TRAFFIC_GEN_MODE` | `sender` | Operating mode: `sender` or `responder` |
+| `INFLUXDB_TOKEN` | `openbmp-telemetry-token` | InfluxDB auth token (Telegraf) |
--- a/docker-compose.yml
+++ b/docker-compose.yml
@ -231,6 +231,62 @@ services:
    network_mode: host
    # Serves on port 5001 (host network, defined in nginx.conf)

+  # --- Phase 4: gNMI Streaming Telemetry ---
+
+  influxdb:
+    restart: unless-stopped
+    container_name: obmp-influxdb
+    image: influxdb:2.7
+    ports:
+      - "8086:8086"
+    volumes:
+      - ${OBMP_DATA_ROOT}/influxdb:/var/lib/influxdb2
+    environment:
+      - DOCKER_INFLUXDB_INIT_MODE=setup
+      - DOCKER_INFLUXDB_INIT_USERNAME=openbmp
+      - DOCKER_INFLUXDB_INIT_PASSWORD=openbmp123
+      - DOCKER_INFLUXDB_INIT_ORG=openbmp
+      - DOCKER_INFLUXDB_INIT_BUCKET=telemetry
+      - DOCKER_INFLUXDB_INIT_ADMIN_TOKEN=openbmp-telemetry-token
+      - DOCKER_INFLUXDB_INIT_RETENTION=30d
+
+  telegraf:
+    restart: unless-stopped
+    container_name: obmp-telegraf
+    build:
+      context: ./telegraf
+      dockerfile: Dockerfile
+    network_mode: host
+    depends_on:
+      - influxdb
+    environment:
+      - INFLUXDB_TOKEN=openbmp-telemetry-token
+
+  # --- Phase 4: Traffic Generator ---
+
+  traffic-gen:
+    restart: unless-stopped
+    container_name: obmp-traffic-gen
+    build:
+      context: ./traffic-gen
+      dockerfile: Dockerfile
+    network_mode: host
+    cap_add:
+      - NET_RAW
+      - NET_ADMIN
+    environment:
+      - TRAFFIC_GEN_API_PORT=5051
+      - TRAFFIC_GEN_MODE=sender
+
+  traffic-gen-ui:
+    restart: unless-stopped
+    container_name: obmp-traffic-gen-ui
+    build:
+      context: ./traffic-gen-ui
+      dockerfile: Dockerfile
+    network_mode: host
+    # Serves on port 5002 (host network, defined in nginx.conf)
+
  whois:
    restart: unless-stopped
    container_name: obmp-whois
--- a/gnmi/pycache/gnmi_grpc_config.cpython-310.pyc
+++ b/gnmi/pycache/gnmi_grpc_config.cpython-310.pyc
--- a/gnmi/gnmi_grpc_config.py
+++ b/gnmi/gnmi_grpc_config.py
@ -0,0 +1,138 @@
+#!/usr/bin/env python3
+"""
+gNMI gRPC Configuration Script
+===============================
+Enables gRPC dial-in telemetry on all 9 IOS-XR routers so that
+Telegraf (or any gNMI collector) can subscribe to streaming
+telemetry data.
+
+What this script applies per router:
+  - gRPC server on port 57400 with TLS disabled
+  - YANG model: Cisco-IOS-XR-man-ems-cfg
+
+Router targets:
+  CORE-01 (10.100.0.100)
+  CORE-02 (10.100.0.200)
+  R9K-01  (10.100.0.1) through R9K-07 (10.100.0.7)
+"""
+
+from ncclient import manager
+import sys
+
+GRPC_NS = 'http://cisco.com/ns/yang/Cisco-IOS-XR-man-ems-cfg'
+
+ROUTERS = [
+    ('10.100.0.100', 'CORE-01'),
+    ('10.100.0.200', 'CORE-02'),
+    ('10.100.0.1',   'R9K-01'),
+    ('10.100.0.2',   'R9K-02'),
+    ('10.100.0.3',   'R9K-03'),
+    ('10.100.0.4',   'R9K-04'),
+    ('10.100.0.5',   'R9K-05'),
+    ('10.100.0.6',   'R9K-06'),
+    ('10.100.0.7',   'R9K-07'),
+]
+
+GRPC_CONFIG_XML = """
+<config>
+  <grpc xmlns="http://cisco.com/ns/yang/Cisco-IOS-XR-man-ems-cfg">
+    <port>57400</port>
+    <no-tls/>
+  </grpc>
+</config>
+"""
+
+
+def configure_router(mgmt_ip, label):
+    """Apply gRPC configuration via NETCONF edit-config + commit."""
+    print(f"\n{'─'*60}")
+    print(f"  Configuring {label} ({mgmt_ip})")
+    print(f"{'─'*60}")
+    print(f"  Applying: gRPC port=57400 no-tls")
+
+    try:
+        with manager.connect(
+            host=mgmt_ip,
+            port=830,
+            username='webui',
+            password='cisco',
+            hostkey_verify=False,
+            device_params={'name': 'iosxr'},
+            timeout=20,
+        ) as m:
+            print("  → Applying gRPC configuration...")
+            m.edit_config(target='candidate', config=GRPC_CONFIG_XML)
+
+            print("  → Committing...")
+            m.commit()
+            print(f"  ✓ {label} done.")
+            return True
+
+    except Exception as e:
+        print(f"  ✗ ERROR on {label}: {e}")
+        return False
+
+
+def verify_router(mgmt_ip, label):
+    """Re-read running config to confirm the grpc block is present."""
+    try:
+        with manager.connect(
+            host=mgmt_ip, port=830, username='webui', password='cisco',
+            hostkey_verify=False, device_params={'name': 'iosxr'}, timeout=10
+        ) as m:
+            filt_grpc = """<filter>
+              <grpc xmlns="http://cisco.com/ns/yang/Cisco-IOS-XR-man-ems-cfg">
+                <port/>
+                <no-tls/>
+              </grpc>
+            </filter>"""
+            r_grpc = m.get_config(source='running', filter=filt_grpc)
+
+            has_grpc = '<port>57400</port>' in str(r_grpc)
+            has_notls = '<no-tls' in str(r_grpc)
+
+            g = '✓' if has_grpc else '✗'
+            t = '✓' if has_notls else '✗'
+            status = 'OK' if (has_grpc and has_notls) else 'INCOMPLETE'
+            print(f"  {label:8s}  grpc-port={g}  no-tls={t}  [{status}]")
+
+    except Exception as e:
+        print(f"  {label:8s}  verify error: {e}")
+
+
+def main():
+    print("gNMI gRPC Configuration Script")
+    print("================================")
+    print(f"Targets: all {len(ROUTERS)} routers")
+    print()
+
+    results = []
+    for mgmt_ip, label in ROUTERS:
+        ok = configure_router(mgmt_ip, label)
+        results.append((mgmt_ip, label, ok))
+
+    # Summary
+    print(f"\n{'='*60}")
+    print("Post-apply verification")
+    print('='*60)
+    print(f"  {'Router':8s}  {'gRPC Port':9s}  {'No-TLS':6s}  Status")
+    for mgmt_ip, label, ok in results:
+        if ok:
+            verify_router(mgmt_ip, label)
+        else:
+            print(f"  {label:8s}  skipped (apply failed)")
+
+    failed = [label for _, label, ok in results if not ok]
+    print()
+    if failed:
+        print(f"FAILED: {', '.join(failed)}")
+        sys.exit(1)
+    else:
+        print("All routers configured successfully.")
+        print()
+        print("gRPC is now listening on port 57400 (no TLS) on all routers.")
+        print("Next: start Telegraf with gNMI input plugin to begin collecting telemetry.")
+
+
+if __name__ == '__main__':
+    main()
--- a/obmp-grafana/dashboards/Telemetry-3001/combined_bmp_telemetry.json
+++ b/obmp-grafana/dashboards/Telemetry-3001/combined_bmp_telemetry.json
@ -0,0 +1,132 @@
+{
+  "annotations": {"list": [{"builtIn": 1,"datasource": {"type": "datasource","uid": "grafana"},"enable": true,"hide": true,"iconColor": "rgba(0, 211, 255, 1)","name": "Annotations & Alerts","type": "dashboard"}]},
+  "description": "Combined view of BMP control-plane data (from PostgreSQL) and gNMI data-plane telemetry (from InfluxDB). Correlate BGP peer state with interface traffic patterns.",
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 1,
+  "id": null,
+  "links": [],
+  "templating": {
+    "list": [
+      {
+        "current": {},
+        "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+        "definition": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "hide": 0,
+        "includeAll": true,
+        "label": "Router",
+        "multi": true,
+        "name": "router",
+        "options": [],
+        "query": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "refresh": 2,
+        "regex": "",
+        "type": "query"
+      }
+    ]
+  },
+  "panels": [
+    {
+      "datasource": {"type": "postgres","uid": "obmp_postgres"},
+      "description": "Current BGP peer status from the OpenBMP PostgreSQL database. Shows peer address, name, and session state.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "thresholds"},
+          "custom": {"align": "auto","displayMode": "auto","filterable": true,"inspect": true},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null}]}
+        },
+        "overrides": [
+          {"matcher": {"id": "byName","options": "state"},"properties": [{"id": "custom.displayMode","value": "color-background-solid"},{"id": "mappings","value": [{"options": {"down": {"color": "red","index": 1,"text": "DOWN"},"up": {"color": "green","index": 0,"text": "UP"}},"type": "value"}]}]},
+          {"matcher": {"id": "byName","options": "peer_addr"},"properties": [{"id": "custom.width","value": 160}]},
+          {"matcher": {"id": "byName","options": "name"},"properties": [{"id": "custom.width","value": 200}]}
+        ]
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 0},
+      "id": 1,
+      "options": {"footer": {"fields": "","reducer": ["sum"],"show": false},"showHeader": true,"sortBy": [{"desc": false,"displayName": "state"}]},
+      "targets": [
+        {
+          "datasource": {"type": "postgres","uid": "obmp_postgres"},
+          "format": "table",
+          "rawSql": "SELECT\n  p.peer_addr,\n  COALESCE(p.name, p.peer_addr::text) AS name,\n  p.state,\n  p.peer_as AS \"AS\",\n  p.router_hash_id IS NOT NULL AS \"BMP Active\",\n  p.timestamp AS \"Last State Change\"\nFROM bgp_peers p\nWHERE p.isprepolicy = true\nORDER BY p.state, p.peer_addr",
+          "refId": "A"
+        }
+      ],
+      "title": "BGP Peer Status",
+      "type": "table"
+    },
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Interface traffic rates from gNMI streaming telemetry. Shows bytes per second for each interface across selected routers.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "line","fillOpacity": 10,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "none"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "red","value": 80}]},
+          "unit": "Bps"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 10},
+      "id": 2,
+      "options": {"legend": {"calcs": ["mean","max"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r._field == \"bytes_received\" or r._field == \"bytes_sent\")\n  |> derivative(unit: 1s, nonNegative: true)\n  |> map(fn: (r) => ({r with _value: if r._value < 0.0 then 0.0 else r._value}))",
+          "refId": "A"
+        }
+      ],
+      "title": "Interface Traffic",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "postgres","uid": "obmp_postgres"},
+      "description": "BGP update activity over time from the OpenBMP PostgreSQL database. Shows peer event transitions and update counts for correlation with traffic patterns.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "bars","fillOpacity": 50,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "normal"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null}]},
+          "unit": "short"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 20},
+      "id": 3,
+      "options": {"legend": {"calcs": ["sum"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "postgres","uid": "obmp_postgres"},
+          "format": "time_series",
+          "rawSql": "SELECT\n  $__timeGroupAlias(e.timestamp, '1m'),\n  COALESCE(p.name, p.peer_addr::text) AS metric,\n  COUNT(*) AS \"value\"\nFROM peer_event_log e\nJOIN bgp_peers p ON p.hash_id = e.peer_hash_id\nWHERE $__timeFilter(e.timestamp)\nGROUP BY 1, 2\nORDER BY 1",
+          "refId": "A"
+        }
+      ],
+      "title": "BGP Update Activity",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "datasource","uid": "grafana"},
+      "gridPos": {"h": 6,"w": 24,"x": 0,"y": 30},
+      "id": 4,
+      "options": {
+        "code": {"language": "plaintext","showLineNumbers": false,"showMiniMap": false},
+        "content": "## Combined BMP + Telemetry View\n\nThis dashboard integrates two complementary data sources to provide a unified network monitoring view:\n\n### Control Plane (BMP via PostgreSQL)\n- **BGP Peer Status** -- Real-time BGP session state from BMP (OpenBMP)\n- **BGP Update Activity** -- Session transitions and update events from `peer_event_log`\n\n### Data Plane (gNMI via InfluxDB)\n- **Interface Traffic** -- Streaming telemetry byte rates collected via gNMI at 10-second intervals\n\n### Correlation Use Cases\n- A BGP peer flap (control plane) should correlate with a traffic shift on affected interfaces (data plane)\n- Sustained high interface utilization (data plane) may precede BGP session resets due to congestion\n- Compare the number of active BGP peers with interface traffic to validate routing convergence",
+        "mode": "markdown"
+      },
+      "title": "About",
+      "type": "text"
+    }
+  ],
+  "schemaVersion": 39,
+  "style": "dark",
+  "tags": ["obmp-telemetry"],
+  "time": {"from": "now-1h","to": "now"},
+  "timepicker": {},
+  "timezone": "browser",
+  "title": "Combined BMP + Telemetry View",
+  "uid": "obmp-telem-03",
+  "version": 1
+}
--- a/obmp-grafana/dashboards/Telemetry-3001/interface_errors.json
+++ b/obmp-grafana/dashboards/Telemetry-3001/interface_errors.json
@ -0,0 +1,134 @@
+{
+  "annotations": {"list": [{"builtIn": 1,"datasource": {"type": "datasource","uid": "grafana"},"enable": true,"hide": true,"iconColor": "rgba(0, 211, 255, 1)","name": "Annotations & Alerts","type": "dashboard"}]},
+  "description": "Interface error and drop counters collected via gNMI streaming telemetry. Helps identify interfaces with packet loss or physical layer issues.",
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 1,
+  "id": null,
+  "links": [],
+  "templating": {
+    "list": [
+      {
+        "current": {},
+        "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+        "definition": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "hide": 0,
+        "includeAll": true,
+        "label": "Router",
+        "multi": true,
+        "name": "router",
+        "options": [],
+        "query": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "refresh": 2,
+        "regex": "",
+        "type": "query"
+      },
+      {
+        "current": {},
+        "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+        "definition": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> keep(columns: [\"name\"])\n  |> distinct(column: \"name\")\n  |> sort()",
+        "hide": 0,
+        "includeAll": true,
+        "label": "Interface",
+        "multi": true,
+        "name": "interface",
+        "options": [],
+        "query": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> keep(columns: [\"name\"])\n  |> distinct(column: \"name\")\n  |> sort()",
+        "refresh": 2,
+        "regex": "",
+        "type": "query"
+      }
+    ]
+  },
+  "panels": [
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Interface error counters over time: input errors, output errors, and CRC errors. A rising trend indicates physical or configuration issues.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "line","fillOpacity": 10,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "none"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]},
+          "unit": "short"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 0},
+      "id": 1,
+      "options": {"legend": {"calcs": ["mean","max","last"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"input_errors\" or r._field == \"output_errors\" or r._field == \"crc_errors\")\n  |> derivative(unit: 1s, nonNegative: true)",
+          "refId": "A"
+        }
+      ],
+      "title": "Interface Errors",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Interface drop counters over time: input drops and output drops. Drops indicate congestion or queue overflow.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "line","fillOpacity": 10,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "none"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]},
+          "unit": "short"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 10},
+      "id": 2,
+      "options": {"legend": {"calcs": ["mean","max","last"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"input_drops\" or r._field == \"output_drops\")\n  |> derivative(unit: 1s, nonNegative: true)",
+          "refId": "A"
+        }
+      ],
+      "title": "Interface Drops",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Summary table showing the latest error and drop counter values per interface. Useful for quickly identifying problematic interfaces.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "thresholds"},
+          "custom": {"align": "auto","displayMode": "auto","filterable": true,"inspect": true},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]}
+        },
+        "overrides": [
+          {"matcher": {"id": "byName","options": "input_errors"},"properties": [{"id": "custom.displayMode","value": "color-background-solid"},{"id": "thresholds","value": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]}}]},
+          {"matcher": {"id": "byName","options": "output_errors"},"properties": [{"id": "custom.displayMode","value": "color-background-solid"},{"id": "thresholds","value": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]}}]},
+          {"matcher": {"id": "byName","options": "input_drops"},"properties": [{"id": "custom.displayMode","value": "color-background-solid"},{"id": "thresholds","value": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]}}]},
+          {"matcher": {"id": "byName","options": "output_drops"},"properties": [{"id": "custom.displayMode","value": "color-background-solid"},{"id": "thresholds","value": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "yellow","value": 1},{"color": "red","value": 100}]}}]}
+        ]
+      },
+      "gridPos": {"h": 12,"w": 24,"x": 0,"y": 20},
+      "id": 3,
+      "options": {"footer": {"fields": "","reducer": ["sum"],"show": false},"showHeader": true,"sortBy": [{"desc": true,"displayName": "input_errors"}]},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"input_errors\" or r._field == \"output_errors\" or r._field == \"crc_errors\" or r._field == \"input_drops\" or r._field == \"output_drops\")\n  |> last()\n  |> pivot(rowKey: [\"_time\"], columnKey: [\"_field\"], valueColumn: \"_value\")\n  |> keep(columns: [\"source\", \"name\", \"input_errors\", \"output_errors\", \"crc_errors\", \"input_drops\", \"output_drops\"])\n  |> sort(columns: [\"input_errors\"], desc: true)",
+          "refId": "A"
+        }
+      ],
+      "title": "Error Summary Table",
+      "type": "table"
+    }
+  ],
+  "schemaVersion": 39,
+  "style": "dark",
+  "tags": ["obmp-telemetry"],
+  "time": {"from": "now-1h","to": "now"},
+  "timepicker": {},
+  "timezone": "browser",
+  "title": "Interface Errors & Drops",
+  "uid": "obmp-telem-02",
+  "version": 1
+}
--- a/obmp-grafana/dashboards/Telemetry-3001/interface_utilization.json
+++ b/obmp-grafana/dashboards/Telemetry-3001/interface_utilization.json
@ -0,0 +1,141 @@
+{
+  "annotations": {"list": [{"builtIn": 1,"datasource": {"type": "datasource","uid": "grafana"},"enable": true,"hide": true,"iconColor": "rgba(0, 211, 255, 1)","name": "Annotations & Alerts","type": "dashboard"}]},
+  "description": "Interface utilization metrics collected via gNMI streaming telemetry from IOS-XR routers. Shows byte rates, packet rates, and top interfaces by traffic volume.",
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 1,
+  "id": null,
+  "links": [],
+  "templating": {
+    "list": [
+      {
+        "current": {},
+        "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+        "definition": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "hide": 0,
+        "includeAll": true,
+        "label": "Router",
+        "multi": true,
+        "name": "router",
+        "options": [],
+        "query": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> keep(columns: [\"source\"])\n  |> distinct(column: \"source\")\n  |> sort()",
+        "refresh": 2,
+        "regex": "",
+        "type": "query"
+      },
+      {
+        "current": {},
+        "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+        "definition": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> keep(columns: [\"name\"])\n  |> distinct(column: \"name\")\n  |> sort()",
+        "hide": 0,
+        "includeAll": true,
+        "label": "Interface",
+        "multi": true,
+        "name": "interface",
+        "options": [],
+        "query": "from(bucket: \"telemetry\")\n  |> range(start: -1h)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> keep(columns: [\"name\"])\n  |> distinct(column: \"name\")\n  |> sort()",
+        "refresh": 2,
+        "regex": "",
+        "type": "query"
+      }
+    ]
+  },
+  "panels": [
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Rate of bytes received and sent per interface, computed as the derivative of cumulative counters. Unit: bytes per second.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "line","fillOpacity": 10,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "none"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "red","value": 80}]},
+          "unit": "Bps"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 0},
+      "id": 1,
+      "options": {"legend": {"calcs": ["mean","max"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"bytes_received\" or r._field == \"bytes_sent\")\n  |> derivative(unit: 1s, nonNegative: true)\n  |> map(fn: (r) => ({r with _value: if r._value < 0.0 then 0.0 else r._value}))",
+          "refId": "A"
+        }
+      ],
+      "title": "Input/Output Bytes Rate",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Rate of packets received and sent per interface, computed as the derivative of cumulative counters. Unit: packets per second.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","barAlignment": 0,"drawStyle": "line","fillOpacity": 10,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineInterpolation": "linear","lineWidth": 1,"pointSize": 5,"scaleDistribution": {"type": "linear"},"showPoints": "never","spanNulls": false,"stacking": {"group": "A","mode": "none"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null},{"color": "red","value": 80}]},
+          "unit": "pps"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 10},
+      "id": 2,
+      "options": {"legend": {"calcs": ["mean","max"],"displayMode": "table","placement": "bottom"},"tooltip": {"mode": "multi","sort": "desc"}},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"packets_received\" or r._field == \"packets_sent\")\n  |> derivative(unit: 1s, nonNegative: true)\n  |> map(fn: (r) => ({r with _value: if r._value < 0.0 then 0.0 else r._value}))",
+          "refId": "A"
+        }
+      ],
+      "title": "Input/Output Packets Rate",
+      "type": "timeseries"
+    },
+    {
+      "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+      "description": "Top interfaces ranked by total bytes (received + sent) over the selected time range.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {"mode": "palette-classic"},
+          "custom": {"axisBorderShow": false,"axisCenteredZero": false,"axisLabel": "","axisPlacement": "auto","fillOpacity": 80,"gradientMode": "none","hideFrom": {"legend": false,"tooltip": false,"viz": false},"lineWidth": 1,"scaleDistribution": {"type": "linear"},"thresholdsStyle": {"mode": "off"}},
+          "mappings": [],
+          "thresholds": {"mode": "absolute","steps": [{"color": "green","value": null}]},
+          "unit": "decbytes"
+        }
+      },
+      "gridPos": {"h": 10,"w": 24,"x": 0,"y": 20},
+      "id": 3,
+      "options": {"barRadius": 0,"barWidth": 0.6,"fullHighlight": false,"groupWidth": 0.7,"legend": {"calcs": [],"displayMode": "list","placement": "bottom"},"orientation": "horizontal","showValue": "auto","stacking": "none","tooltip": {"mode": "single","sort": "none"},"xTickLabelRotation": 0},
+      "targets": [
+        {
+          "datasource": {"type": "influxdb","uid": "obmp_influxdb"},
+          "query": "from(bucket: \"telemetry\")\n  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)\n  |> filter(fn: (r) => r._measurement == \"interface_counters\")\n  |> filter(fn: (r) => r.source =~ /${router:regex}/)\n  |> filter(fn: (r) => r.name =~ /${interface:regex}/)\n  |> filter(fn: (r) => r._field == \"bytes_received\" or r._field == \"bytes_sent\")\n  |> derivative(unit: 1s, nonNegative: true)\n  |> group(columns: [\"source\", \"name\", \"_field\"])\n  |> sum()\n  |> group(columns: [\"source\", \"name\"])\n  |> sum()\n  |> group()\n  |> sort(columns: [\"_value\"], desc: true)\n  |> limit(n: 20)",
+          "refId": "A"
+        }
+      ],
+      "title": "Top Interfaces by Traffic",
+      "type": "barchart"
+    },
+    {
+      "datasource": {"type": "datasource","uid": "grafana"},
+      "gridPos": {"h": 4,"w": 24,"x": 0,"y": 30},
+      "id": 4,
+      "options": {
+        "code": {"language": "plaintext","showLineNumbers": false,"showMiniMap": false},
+        "content": "## Interface Utilization Dashboard\n\nThis dashboard displays real-time interface utilization metrics collected via **gNMI streaming telemetry** from IOS-XR routers.\n\n- **Data source:** InfluxDB (Telegraf gNMI input plugin)\n- **YANG model:** `Cisco-IOS-XR-infra-statsd-oper`\n- **Subscription path:** `/infra-statistics/interfaces/interface/latest/generic-counters`\n- **Sample interval:** 10 seconds\n\nUse the **Router** and **Interface** template variables at the top to filter the view.",
+        "mode": "markdown"
+      },
+      "title": "About",
+      "type": "text"
+    }
+  ],
+  "schemaVersion": 39,
+  "style": "dark",
+  "tags": ["obmp-telemetry"],
+  "time": {"from": "now-1h","to": "now"},
+  "timepicker": {},
+  "timezone": "browser",
+  "title": "Interface Utilization",
+  "uid": "obmp-telem-01",
+  "version": 1
+}
--- a/obmp-grafana/provisioning/dashboards/openbmp-dashboards.yml
+++ b/obmp-grafana/provisioning/dashboards/openbmp-dashboards.yml
@ -133,4 +133,15 @@ providers:
    allowUiUpdates: true
    options:
      path: /var/lib/grafana/dashboards/Learning
+      foldersFromFilesStructure: false
+  - name: 'OpenBMP-Telemetry'
+    orgId: 1
+    folder: 'OBMP-Telemetry'
+    folderUid: '3001'
+    type: file
+    disableDeletion: false
+    updateIntervalSeconds: 30
+    allowUiUpdates: true
+    options:
+      path: /var/lib/grafana/dashboards/Telemetry-3001
      foldersFromFilesStructure: false
--- a/obmp-grafana/provisioning/datasources/influxdb-ds.yml
+++ b/obmp-grafana/provisioning/datasources/influxdb-ds.yml
@ -0,0 +1,16 @@
+apiVersion: 1
+
+datasources:
+  - name: InfluxDB-Telemetry
+    uid: obmp_influxdb
+    type: influxdb
+    access: proxy
+    url: http://obmp-influxdb:8086
+    jsonData:
+      version: Flux
+      organization: openbmp
+      defaultBucket: telemetry
+    secureJsonData:
+      token: openbmp-telemetry-token
+    isDefault: false
+    editable: true
--- a/telegraf/Dockerfile
+++ b/telegraf/Dockerfile
@ -0,0 +1,2 @@
+FROM telegraf:1.28-alpine
+COPY telegraf.conf /etc/telegraf/telegraf.conf
--- a/telegraf/telegraf.conf
+++ b/telegraf/telegraf.conf
@ -0,0 +1,66 @@
+# Telegraf Configuration for gNMI Streaming Telemetry
+# Collects interface counters and data rates from IOS-XR routers
+
+[global_tags]
+
+[agent]
+  interval = "10s"
+  round_interval = true
+  metric_batch_size = 1000
+  metric_buffer_limit = 10000
+  collection_jitter = "0s"
+  flush_interval = "10s"
+  flush_jitter = "0s"
+  precision = "0s"
+
+###############################################################################
+#                            INPUT PLUGINS                                    #
+###############################################################################
+
+[[inputs.gnmi]]
+  addresses = [
+    "10.100.0.100:57400",
+    "10.100.0.200:57400",
+    "10.100.0.1:57400",
+    "10.100.0.2:57400",
+    "10.100.0.3:57400",
+    "10.100.0.4:57400",
+    "10.100.0.5:57400",
+    "10.100.0.6:57400",
+    "10.100.0.7:57400"
+  ]
+  username = "webui"
+  password = "cisco"
+
+  ## Do not verify the server certificate
+  enable_tls = false
+
+  ## gNMI encoding requested (one of: "proto", "json", "json_ietf", "bytes")
+  encoding = "proto"
+
+  ## Redial in case of failures after
+  redial = "10s"
+
+  [[inputs.gnmi.subscription]]
+    name = "interface_counters"
+    origin = "Cisco-IOS-XR-infra-statsd-oper"
+    path = "/infra-statistics/interfaces/interface/latest/generic-counters"
+    subscription_mode = "sample"
+    sample_interval = "10s"
+
+  [[inputs.gnmi.subscription]]
+    name = "interface_rates"
+    origin = "Cisco-IOS-XR-infra-statsd-oper"
+    path = "/infra-statistics/interfaces/interface/latest/data-rate"
+    subscription_mode = "sample"
+    sample_interval = "10s"
+
+###############################################################################
+#                            OUTPUT PLUGINS                                   #
+###############################################################################
+
+[[outputs.influxdb_v2]]
+  urls = ["http://localhost:8086"]
+  token = "${INFLUXDB_TOKEN}"
+  organization = "openbmp"
+  bucket = "telemetry"
--- a/traffic-gen-ui/Dockerfile
+++ b/traffic-gen-ui/Dockerfile
@ -0,0 +1,12 @@
+FROM node:20-alpine AS build
+WORKDIR /app
+COPY package.json ./
+RUN npm install
+COPY . .
+RUN npm run build
+
+FROM nginx:alpine
+COPY --from=build /app/dist /usr/share/nginx/html
+COPY nginx.conf /etc/nginx/conf.d/default.conf
+EXPOSE 5002
+CMD ["nginx", "-g", "daemon off;"]
--- a/traffic-gen-ui/index.html
+++ b/traffic-gen-ui/index.html
@ -0,0 +1,12 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8" />
+  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+  <title>Traffic Generator</title>
+</head>
+<body>
+  <div id="app"></div>
+  <script type="module" src="/src/main.js"></script>
+</body>
+</html>
--- a/traffic-gen-ui/nginx.conf
+++ b/traffic-gen-ui/nginx.conf
@ -0,0 +1,15 @@
+server {
+    listen 5002;
+    root /usr/share/nginx/html;
+    index index.html;
+
+    location /api/ {
+        proxy_pass http://localhost:5051/;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+    }
+
+    location / {
+        try_files $uri $uri/ /index.html;
+    }
+}
--- a/traffic-gen-ui/package.json
+++ b/traffic-gen-ui/package.json
@ -0,0 +1,17 @@
+{
+  "name": "traffic-gen-ui",
+  "version": "1.0.0",
+  "private": true,
+  "scripts": {
+    "dev": "vite",
+    "build": "vite build",
+    "preview": "vite preview"
+  },
+  "dependencies": {
+    "vue": "^3.3.0"
+  },
+  "devDependencies": {
+    "@vitejs/plugin-vue": "^4.2.0",
+    "vite": "^4.4.0"
+  }
+}
--- a/traffic-gen-ui/src/App.vue
+++ b/traffic-gen-ui/src/App.vue
@ -0,0 +1,313 @@
+<template>
+  <div class="app-layout">
+    <!-- HEADER -->
+    <header class="app-header">
+      <div class="header-title">
+        <span class="logo-icon">&#9889;</span>
+        <h1>Traffic Generator</h1>
+      </div>
+      <StatusBar :health="health" :api-error="apiError" />
+    </header>
+
+    <!-- ERROR BANNER -->
+    <div v-if="apiError" class="error-banner">
+      <span class="error-icon">&#9888;</span>
+      API unreachable: {{ apiError }} &mdash; retrying every 5s
+    </div>
+
+    <!-- MAIN CONTENT -->
+    <div class="main-content">
+      <!-- LEFT COLUMN: Flow Builder -->
+      <aside class="left-col">
+        <FlowBuilder @created="fetchFlows" @updated="fetchFlows" />
+      </aside>
+
+      <!-- RIGHT COLUMN: Tabs -->
+      <main class="right-col">
+        <div class="tabs">
+          <button
+            v-for="tab in tabs"
+            :key="tab.id"
+            class="tab-btn"
+            :class="{ active: activeTab === tab.id }"
+            @click="activeTab = tab.id"
+          >
+            {{ tab.label }}
+          </button>
+        </div>
+
+        <div class="tab-content">
+          <FlowTable v-if="activeTab === 'flows'" :flows="flows" @refresh="fetchFlows" />
+          <TestBuilder v-else-if="activeTab === 'tests'" :flows="flows" @created="fetchTests" @refresh="fetchAll" />
+          <TestRunner v-else-if="activeTab === 'runner'" :tests="tests" @refresh="fetchTests" />
+          <ResultsPanel v-else-if="activeTab === 'results'" :tests="tests" />
+          <StatsMonitor v-else-if="activeTab === 'monitor'" :flows="flows" />
+        </div>
+      </main>
+    </div>
+
+    <!-- FOOTER -->
+    <footer class="app-footer">
+      <span>Refreshing every 5s (health) / 3s (flows)</span>
+      <span class="footer-sep">|</span>
+      <a href="http://localhost:3000" target="_blank" class="footer-link">Grafana: :3000</a>
+      <span class="footer-sep">|</span>
+      <a href="http://localhost:5001" target="_blank" class="footer-link">Route Injector: :5001</a>
+    </footer>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted, onUnmounted } from 'vue'
+import { api } from './api.js'
+import StatusBar from './components/StatusBar.vue'
+import FlowBuilder from './components/FlowBuilder.vue'
+import FlowTable from './components/FlowTable.vue'
+import TestBuilder from './components/TestBuilder.vue'
+import TestRunner from './components/TestRunner.vue'
+import ResultsPanel from './components/ResultsPanel.vue'
+import StatsMonitor from './components/StatsMonitor.vue'
+
+const health = ref(null)
+const flows = ref([])
+const tests = ref([])
+const apiError = ref(null)
+const activeTab = ref('flows')
+
+const tabs = [
+  { id: 'flows', label: 'Flows' },
+  { id: 'tests', label: 'Tests' },
+  { id: 'runner', label: 'Runner' },
+  { id: 'results', label: 'Results' },
+  { id: 'monitor', label: 'Monitor' },
+]
+
+async function fetchHealth() {
+  try {
+    health.value = await api.health()
+    apiError.value = null
+  } catch (e) {
+    apiError.value = e.message
+    health.value = null
+  }
+}
+
+async function fetchFlows() {
+  try {
+    const data = await api.flows()
+    flows.value = data.flows || []
+  } catch (_) {}
+}
+
+async function fetchTests() {
+  try {
+    const data = await api.tests()
+    tests.value = data.tests || []
+  } catch (_) {}
+}
+
+async function fetchAll() {
+  await Promise.all([fetchFlows(), fetchTests()])
+}
+
+let healthTimer = null
+let dataTimer = null
+
+onMounted(() => {
+  fetchHealth()
+  fetchAll()
+  healthTimer = setInterval(fetchHealth, 5000)
+  dataTimer = setInterval(fetchAll, 3000)
+})
+
+onUnmounted(() => {
+  clearInterval(healthTimer)
+  clearInterval(dataTimer)
+})
+</script>
+
+<style>
+:root {
+  --bg: #0f1117;
+  --card-bg: #1a1f2e;
+  --border: #2d3748;
+  --accent: #4f9cf9;
+  --success: #48bb78;
+  --danger: #fc8181;
+  --warning: #f6ad55;
+  --text: #e2e8f0;
+  --muted: #718096;
+  --radius: 8px;
+}
+
+* { box-sizing: border-box; margin: 0; padding: 0; }
+
+body {
+  background: var(--bg);
+  color: var(--text);
+  font-family: 'Segoe UI', system-ui, -apple-system, sans-serif;
+  font-size: 14px;
+  line-height: 1.5;
+}
+
+button {
+  cursor: pointer;
+  font-family: inherit;
+  font-size: 13px;
+  border: none;
+  border-radius: var(--radius);
+  transition: opacity 0.15s, background 0.15s;
+}
+
+button:disabled {
+  opacity: 0.5;
+  cursor: not-allowed;
+}
+
+input, select {
+  font-family: inherit;
+  font-size: 13px;
+  background: var(--bg);
+  color: var(--text);
+  border: 1px solid var(--border);
+  border-radius: var(--radius);
+  padding: 6px 10px;
+  outline: none;
+}
+
+input:focus {
+  border-color: var(--accent);
+}
+</style>
+
+<style scoped>
+.app-layout {
+  display: grid;
+  grid-template-rows: auto auto 1fr auto;
+  min-height: 100vh;
+}
+
+.app-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  padding: 12px 20px;
+  background: var(--card-bg);
+  border-bottom: 1px solid var(--border);
+  gap: 16px;
+  flex-wrap: wrap;
+}
+
+.header-title {
+  display: flex;
+  align-items: center;
+  gap: 10px;
+}
+
+.logo-icon {
+  font-size: 22px;
+  color: var(--warning);
+}
+
+.app-header h1 {
+  font-size: 18px;
+  font-weight: 600;
+  color: var(--text);
+  letter-spacing: 0.02em;
+}
+
+.error-banner {
+  background: rgba(252, 129, 129, 0.12);
+  border-bottom: 1px solid var(--danger);
+  color: var(--danger);
+  padding: 8px 20px;
+  font-size: 13px;
+  display: flex;
+  align-items: center;
+  gap: 8px;
+}
+
+.error-icon {
+  font-size: 16px;
+}
+
+.main-content {
+  display: grid;
+  grid-template-columns: 340px 1fr;
+  overflow: hidden;
+  height: calc(100vh - 110px);
+}
+
+.left-col {
+  border-right: 1px solid var(--border);
+  overflow-y: auto;
+  padding: 16px;
+}
+
+.right-col {
+  display: flex;
+  flex-direction: column;
+  overflow: hidden;
+}
+
+.tabs {
+  display: flex;
+  gap: 2px;
+  padding: 12px 16px 0;
+  background: var(--card-bg);
+  border-bottom: 1px solid var(--border);
+}
+
+.tab-btn {
+  background: transparent;
+  color: var(--muted);
+  padding: 8px 18px;
+  border-radius: var(--radius) var(--radius) 0 0;
+  border: 1px solid transparent;
+  border-bottom: none;
+  font-weight: 500;
+}
+
+.tab-btn:hover {
+  color: var(--text);
+  background: rgba(79, 156, 249, 0.08);
+}
+
+.tab-btn.active {
+  color: var(--accent);
+  background: var(--bg);
+  border-color: var(--border);
+  border-bottom: 1px solid var(--bg);
+  margin-bottom: -1px;
+}
+
+.tab-content {
+  flex: 1;
+  overflow-y: auto;
+  padding: 16px;
+}
+
+.app-footer {
+  padding: 8px 20px;
+  background: var(--card-bg);
+  border-top: 1px solid var(--border);
+  color: var(--muted);
+  font-size: 12px;
+  display: flex;
+  align-items: center;
+  gap: 10px;
+}
+
+.footer-sep {
+  color: var(--border);
+}
+
+.footer-link {
+  color: var(--accent);
+  text-decoration: none;
+}
+
+.footer-link:hover {
+  text-decoration: underline;
+}
+</style>
--- a/traffic-gen-ui/src/api.js
+++ b/traffic-gen-ui/src/api.js
@ -0,0 +1,44 @@
+const BASE = '/api'
+
+async function req(method, path, body) {
+  const opts = { method, headers: { 'Content-Type': 'application/json' } }
+  if (body) opts.body = JSON.stringify(body)
+  const r = await fetch(BASE + path, opts)
+  if (!r.ok) throw new Error(`${method} ${path} -> ${r.status}`)
+  return r.json()
+}
+
+export const api = {
+  health:       () => req('GET', '/healthz'),
+  interfaces:   () => req('GET', '/interfaces'),
+  mode:         () => req('GET', '/mode'),
+
+  // Flows
+  flows:        () => req('GET', '/flows'),
+  createFlow:   (f) => req('POST', '/flows', f),
+  getFlow:      (id) => req('GET', `/flows/${id}`),
+  updateFlow:   (id, f) => req('PUT', `/flows/${id}`, f),
+  deleteFlow:   (id) => req('DELETE', `/flows/${id}`),
+  startFlow:    (id) => req('POST', `/flows/${id}/start`),
+  stopFlow:     (id) => req('POST', `/flows/${id}/stop`),
+  flowStats:    (id) => req('GET', `/flows/${id}/stats`),
+
+  // Tests
+  tests:        () => req('GET', '/tests'),
+  createTest:   (t) => req('POST', '/tests', t),
+  getTest:      (id) => req('GET', `/tests/${id}`),
+  startTest:    (id) => req('POST', `/tests/${id}/start`),
+  stopTest:     (id) => req('POST', `/tests/${id}/stop`),
+  testResults:  (id) => req('GET', `/tests/${id}/results`),
+
+  // Presets
+  presets:      () => req('GET', '/presets'),
+  loadPreset:   (name, overrides) => req('POST', `/presets/${name}`, overrides),
+
+  // Stats
+  statsHistory: () => req('GET', '/stats/history'),
+
+  // Responder
+  responderStats: () => req('GET', '/responder/stats'),
+  responderReset: () => req('POST', '/responder/reset'),
+}
--- a/traffic-gen-ui/src/components/FlowBuilder.vue
+++ b/traffic-gen-ui/src/components/FlowBuilder.vue
@ -0,0 +1,126 @@
+<template>
+  <div class="flow-builder">
+    <h3>{{ editing ? 'Edit Flow' : 'Create Flow' }}</h3>
+    <form @submit.prevent="submit">
+      <div class="form-row">
+        <label>Name</label>
+        <input v-model="form.name" placeholder="My Flow" />
+      </div>
+      <div class="form-row">
+        <label>Destination IP *</label>
+        <input v-model="form.dst_ip" placeholder="10.100.0.100" required />
+      </div>
+      <div class="form-row">
+        <label>Source IP</label>
+        <input v-model="form.src_ip" placeholder="auto (from interface)" />
+      </div>
+      <div class="form-row">
+        <label>Dst MAC</label>
+        <input v-model="form.dst_mac" placeholder="auto" />
+      </div>
+      <div class="form-row">
+        <label>Protocol</label>
+        <select v-model="form.protocol">
+          <option value="udp">UDP</option>
+          <option value="tcp">TCP</option>
+          <option value="icmp">ICMP</option>
+        </select>
+      </div>
+      <div v-if="form.protocol !== 'icmp'" class="form-row-pair">
+        <div class="form-row">
+          <label>Src Port</label>
+          <input v-model.number="form.src_port" type="number" min="1" max="65535" />
+        </div>
+        <div class="form-row">
+          <label>Dst Port</label>
+          <input v-model.number="form.dst_port" type="number" min="1" max="65535" />
+        </div>
+      </div>
+      <div class="form-row-pair">
+        <div class="form-row">
+          <label>Frame Size (bytes)</label>
+          <input v-model.number="form.frame_size" type="number" min="64" max="9000" />
+        </div>
+        <div class="form-row">
+          <label>Rate (pps)</label>
+          <input v-model.number="form.rate_pps" type="number" min="1" max="100000" />
+        </div>
+      </div>
+      <div class="form-row-pair">
+        <div class="form-row">
+          <label>Duration (sec)</label>
+          <input v-model.number="form.duration" type="number" min="0" />
+        </div>
+        <div class="form-row">
+          <label>DSCP</label>
+          <input v-model.number="form.dscp" type="number" min="0" max="63" />
+        </div>
+      </div>
+      <div class="form-row">
+        <label>Responder URL (optional)</label>
+        <input v-model="form.responder_url" placeholder="http://host:5053" />
+      </div>
+      <div class="form-actions">
+        <button type="submit" class="btn btn-accent" :disabled="!form.dst_ip">
+          {{ editing ? 'Update' : 'Create Flow' }}
+        </button>
+        <button v-if="editing" type="button" class="btn btn-muted" @click="$emit('cancel')">Cancel</button>
+      </div>
+    </form>
+  </div>
+</template>
+
+<script setup>
+import { reactive, watch } from 'vue'
+import { api } from '../api.js'
+
+const props = defineProps({ editFlow: Object })
+const emit = defineEmits(['created', 'updated', 'cancel'])
+
+const editing = !!props.editFlow
+
+const defaults = {
+  name: '', dst_ip: '', src_ip: '', dst_mac: '',
+  protocol: 'udp', src_port: 50000, dst_port: 5001,
+  frame_size: 512, rate_pps: 1000, duration: 30,
+  dscp: 0, responder_url: '',
+}
+
+const form = reactive({ ...defaults, ...(props.editFlow || {}) })
+
+async function submit() {
+  try {
+    const payload = { ...form }
+    if (!payload.src_ip) delete payload.src_ip
+    if (!payload.dst_mac) delete payload.dst_mac
+    if (!payload.responder_url) delete payload.responder_url
+    if (!payload.name) payload.name = `${payload.protocol.toUpperCase()} -> ${payload.dst_ip}`
+
+    if (editing) {
+      await api.updateFlow(props.editFlow.id, payload)
+      emit('updated')
+    } else {
+      await api.createFlow(payload)
+      Object.assign(form, defaults)
+      emit('created')
+    }
+  } catch (e) {
+    alert('Error: ' + e.message)
+  }
+}
+</script>
+
+<style scoped>
+.flow-builder { padding: 0; }
+h3 { font-size: 15px; margin-bottom: 12px; color: var(--accent); }
+.form-row { margin-bottom: 10px; }
+.form-row label { display: block; font-size: 11px; color: var(--muted); margin-bottom: 3px; text-transform: uppercase; letter-spacing: 0.05em; }
+.form-row input, .form-row select { width: 100%; }
+.form-row-pair { display: grid; grid-template-columns: 1fr 1fr; gap: 8px; }
+.form-actions { display: flex; gap: 8px; margin-top: 14px; }
+.btn { padding: 8px 16px; font-weight: 600; font-size: 13px; }
+.btn-accent { background: var(--accent); color: #fff; }
+.btn-accent:hover { opacity: 0.9; }
+.btn-accent:disabled { opacity: 0.4; }
+.btn-muted { background: var(--border); color: var(--text); }
+</style>
--- a/traffic-gen-ui/src/components/FlowTable.vue
+++ b/traffic-gen-ui/src/components/FlowTable.vue
@ -0,0 +1,92 @@
+<template>
+  <div class="flow-table">
+    <div v-if="!flows.length" class="empty">No flows created yet. Use the builder to create one.</div>
+    <table v-else>
+      <thead>
+        <tr>
+          <th>Name</th>
+          <th>Dst IP</th>
+          <th>Proto</th>
+          <th>Size</th>
+          <th>Rate</th>
+          <th>State</th>
+          <th>TX pps</th>
+          <th>Actions</th>
+        </tr>
+      </thead>
+      <tbody>
+        <tr v-for="f in flows" :key="f.id" :class="{ running: f.state === 'running' }">
+          <td>{{ f.name || '-' }}</td>
+          <td class="mono">{{ f.dst_ip }}</td>
+          <td>{{ f.protocol.toUpperCase() }}</td>
+          <td>{{ f.frame_size }}B</td>
+          <td>{{ f.rate_pps }} pps</td>
+          <td>
+            <span class="state-badge" :class="'state-' + f.state">{{ f.state }}</span>
+          </td>
+          <td class="mono">{{ stats[f.id]?.tx_pps || 0 }}</td>
+          <td class="actions">
+            <button v-if="f.state !== 'running'" class="btn-sm btn-go" @click="start(f.id)">Start</button>
+            <button v-else class="btn-sm btn-stop" @click="stop(f.id)">Stop</button>
+            <button class="btn-sm btn-del" @click="del(f.id)" :disabled="f.state === 'running'">Del</button>
+          </td>
+        </tr>
+      </tbody>
+    </table>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted, onUnmounted } from 'vue'
+import { api } from '../api.js'
+
+const props = defineProps({ flows: Array })
+const emit = defineEmits(['refresh'])
+const stats = ref({})
+
+let statsTimer = null
+
+async function fetchStats() {
+  for (const f of (props.flows || [])) {
+    if (f.state === 'running') {
+      try {
+        const s = await api.flowStats(f.id)
+        stats.value[f.id] = s
+      } catch (_) {}
+    }
+  }
+}
+
+onMounted(() => { statsTimer = setInterval(fetchStats, 1000) })
+onUnmounted(() => { clearInterval(statsTimer) })
+
+async function start(id) {
+  try { await api.startFlow(id); emit('refresh') } catch (e) { alert(e.message) }
+}
+async function stop(id) {
+  try { await api.stopFlow(id); emit('refresh') } catch (e) { alert(e.message) }
+}
+async function del(id) {
+  try { await api.deleteFlow(id); emit('refresh') } catch (e) { alert(e.message) }
+}
+</script>
+
+<style scoped>
+.flow-table { overflow-x: auto; }
+.empty { color: var(--muted); padding: 20px; text-align: center; }
+table { width: 100%; border-collapse: collapse; }
+th { text-align: left; font-size: 11px; color: var(--muted); text-transform: uppercase; letter-spacing: 0.05em; padding: 6px 8px; border-bottom: 1px solid var(--border); }
+td { padding: 8px; border-bottom: 1px solid rgba(45,55,72,0.5); font-size: 13px; }
+tr.running { background: rgba(79,156,249,0.05); }
+.mono { font-family: 'Cascadia Code', 'Fira Code', monospace; font-size: 12px; }
+.state-badge { font-size: 11px; padding: 2px 8px; border-radius: 10px; font-weight: 600; }
+.state-idle { background: rgba(113,128,150,0.2); color: var(--muted); }
+.state-running { background: rgba(72,187,120,0.15); color: var(--success); }
+.state-stopped { background: rgba(246,173,85,0.15); color: var(--warning); }
+.actions { display: flex; gap: 4px; }
+.btn-sm { padding: 3px 10px; font-size: 11px; font-weight: 600; border-radius: 6px; }
+.btn-go { background: var(--success); color: #fff; }
+.btn-stop { background: var(--warning); color: #000; }
+.btn-del { background: rgba(252,129,129,0.15); color: var(--danger); }
+.btn-del:disabled { opacity: 0.3; }
+</style>
--- a/traffic-gen-ui/src/components/ResultsPanel.vue
+++ b/traffic-gen-ui/src/components/ResultsPanel.vue
@ -0,0 +1,126 @@
+<template>
+  <div class="results-panel">
+    <h3>Test Results</h3>
+    <div v-if="!completedTests.length" class="empty">No completed tests yet.</div>
+
+    <div v-for="t in completedTests" :key="t.id" class="result-card">
+      <div class="result-header">
+        <strong>{{ t.type }} Test</strong>
+        <span class="result-time">{{ t.completed_at }}</span>
+        <div class="export-btns">
+          <button class="btn-sm btn-export" @click="exportJSON(t)">Export JSON</button>
+          <button class="btn-sm btn-export" @click="exportCSV(t)">Export CSV</button>
+        </div>
+      </div>
+
+      <div v-if="t.results" class="result-table">
+        <table>
+          <thead>
+            <tr>
+              <th>Frame Size (B)</th>
+              <th v-for="col in resultColumns(t)" :key="col">{{ col }}</th>
+            </tr>
+          </thead>
+          <tbody>
+            <tr v-for="(val, size) in t.results" :key="size">
+              <td>{{ size }}</td>
+              <td v-for="col in resultColumns(t)" :key="col" class="mono">
+                {{ formatVal(val, col) }}
+              </td>
+            </tr>
+          </tbody>
+        </table>
+      </div>
+
+      <div v-if="t.results" class="result-chart">
+        <div class="bar-chart">
+          <div v-for="(val, size) in t.results" :key="size" class="bar-item">
+            <div class="bar-fill" :style="{ height: barHeight(t, val) + '%' }"></div>
+            <span class="bar-label">{{ size }}B</span>
+          </div>
+        </div>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { computed } from 'vue'
+
+const props = defineProps({ tests: Array })
+
+const completedTests = computed(() =>
+  (props.tests || []).filter(t => t.state === 'complete' && t.results)
+)
+
+function resultColumns(t) {
+  if (t.type === 'throughput') return ['Max Rate (pps)', 'Throughput (Mbps)']
+  if (t.type === 'latency') return ['Min (ms)', 'Avg (ms)', 'Max (ms)', 'Jitter (ms)']
+  if (t.type === 'frame_loss') return ['Loss %']
+  if (t.type === 'back_to_back') return ['Max Burst']
+  return ['Value']
+}
+
+function formatVal(val, col) {
+  if (typeof val === 'object') {
+    if (col.includes('Rate')) return val.max_rate_pps ?? '-'
+    if (col.includes('Throughput')) return val.throughput_mbps ?? '-'
+    if (col.includes('Min')) return val.min_ms ?? '-'
+    if (col.includes('Avg')) return val.avg_ms ?? '-'
+    if (col.includes('Max') && col.includes('ms')) return val.max_ms ?? '-'
+    if (col.includes('Jitter')) return val.jitter_ms ?? '-'
+    if (col.includes('Loss')) return val.loss_pct ?? '-'
+    if (col.includes('Burst')) return val.max_burst ?? '-'
+    return JSON.stringify(val)
+  }
+  return val
+}
+
+function barHeight(t, val) {
+  const v = typeof val === 'object' ? (val.max_rate_pps || val.avg_ms || val.loss_pct || val.max_burst || 0) : val
+  return Math.min(100, Math.max(5, v / 100))
+}
+
+function exportJSON(t) {
+  const blob = new Blob([JSON.stringify(t, null, 2)], { type: 'application/json' })
+  downloadBlob(blob, `test_${t.type}_${t.id}.json`)
+}
+
+function exportCSV(t) {
+  if (!t.results) return
+  const cols = resultColumns(t)
+  let csv = 'Frame Size,' + cols.join(',') + '\n'
+  for (const [size, val] of Object.entries(t.results)) {
+    csv += size + ',' + cols.map(c => formatVal(val, c)).join(',') + '\n'
+  }
+  downloadBlob(new Blob([csv], { type: 'text/csv' }), `test_${t.type}_${t.id}.csv`)
+}
+
+function downloadBlob(blob, name) {
+  const url = URL.createObjectURL(blob)
+  const a = document.createElement('a')
+  a.href = url; a.download = name; a.click()
+  URL.revokeObjectURL(url)
+}
+</script>
+
+<style scoped>
+h3 { font-size: 15px; margin-bottom: 12px; color: var(--accent); }
+.empty { color: var(--muted); padding: 20px; text-align: center; }
+.result-card { background: var(--card-bg); border: 1px solid var(--border); border-radius: var(--radius); padding: 12px; margin-bottom: 12px; }
+.result-header { display: flex; align-items: center; gap: 12px; margin-bottom: 10px; flex-wrap: wrap; }
+.result-header strong { font-size: 14px; text-transform: capitalize; }
+.result-time { font-size: 11px; color: var(--muted); }
+.export-btns { margin-left: auto; display: flex; gap: 4px; }
+.btn-sm { padding: 3px 10px; font-size: 11px; font-weight: 600; border-radius: 6px; }
+.btn-export { background: rgba(79,156,249,0.12); color: var(--accent); }
+.btn-export:hover { background: rgba(79,156,249,0.25); }
+table { width: 100%; border-collapse: collapse; }
+th { font-size: 11px; color: var(--muted); text-align: left; padding: 4px 8px; border-bottom: 1px solid var(--border); }
+td { font-size: 13px; padding: 4px 8px; }
+.mono { font-family: monospace; }
+.bar-chart { display: flex; align-items: flex-end; gap: 8px; height: 80px; margin-top: 12px; padding: 0 8px; }
+.bar-item { flex: 1; display: flex; flex-direction: column; align-items: center; height: 100%; }
+.bar-fill { width: 100%; background: var(--accent); border-radius: 3px 3px 0 0; min-height: 4px; transition: height 0.3s; margin-top: auto; }
+.bar-label { font-size: 10px; color: var(--muted); margin-top: 4px; }
+</style>
--- a/traffic-gen-ui/src/components/StatsMonitor.vue
+++ b/traffic-gen-ui/src/components/StatsMonitor.vue
@ -0,0 +1,129 @@
+<template>
+  <div class="stats-monitor">
+    <h3>Live Statistics</h3>
+
+    <div class="flow-selector">
+      <label>Flow:</label>
+      <select v-model="selectedFlow">
+        <option value="">All Flows</option>
+        <option v-for="f in flows" :key="f.id" :value="f.id">{{ f.name || f.dst_ip }}</option>
+      </select>
+    </div>
+
+    <div class="stats-grid">
+      <div class="stat-card">
+        <div class="stat-value tx">{{ current.tx_pps || 0 }}</div>
+        <div class="stat-label">TX pps</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value rx">{{ current.rx_pps || 0 }}</div>
+        <div class="stat-label">RX pps</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value tx">{{ (current.tx_mbps || 0).toFixed(2) }}</div>
+        <div class="stat-label">TX Mbps</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value rx">{{ (current.rx_mbps || 0).toFixed(2) }}</div>
+        <div class="stat-label">RX Mbps</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value" :class="lossClass">{{ (current.loss_pct || 0).toFixed(1) }}%</div>
+        <div class="stat-label">Loss</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value">{{ current.avg_latency_ms ? current.avg_latency_ms.toFixed(1) : '-' }}</div>
+        <div class="stat-label">Avg Latency (ms)</div>
+      </div>
+    </div>
+
+    <div class="totals">
+      <span>TX Packets: {{ current.tx_packets || 0 }}</span>
+      <span>RX Packets: {{ current.rx_packets || 0 }}</span>
+      <span>TX Bytes: {{ formatBytes(current.tx_bytes || 0) }}</span>
+      <span>RX Bytes: {{ formatBytes(current.rx_bytes || 0) }}</span>
+    </div>
+
+    <div class="history-chart">
+      <div class="chart-header">TX/RX Rate History (last 60s)</div>
+      <div class="sparkline">
+        <div v-for="(s, i) in history" :key="i" class="spark-bar">
+          <div class="spark-tx" :style="{ height: sparkHeight(s.tx_pps) + 'px' }"></div>
+          <div class="spark-rx" :style="{ height: sparkHeight(s.rx_pps) + 'px' }"></div>
+        </div>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { ref, computed, onMounted, onUnmounted } from 'vue'
+import { api } from '../api.js'
+
+const props = defineProps({ flows: Array })
+
+const selectedFlow = ref('')
+const current = ref({})
+const history = ref([])
+
+const lossClass = computed(() => {
+  const l = current.value.loss_pct || 0
+  if (l > 5) return 'loss-high'
+  if (l > 0) return 'loss-med'
+  return 'loss-ok'
+})
+
+let timer = null
+
+async function fetchStats() {
+  try {
+    if (selectedFlow.value) {
+      const s = await api.flowStats(selectedFlow.value)
+      current.value = s
+    } else {
+      const h = await api.statsHistory()
+      if (h.current) current.value = h.current
+      if (h.history) history.value = h.history.slice(-60)
+    }
+  } catch (_) {}
+}
+
+function sparkHeight(val) {
+  if (!val) return 0
+  const max = Math.max(...history.value.map(s => Math.max(s.tx_pps || 0, s.rx_pps || 0)), 1)
+  return Math.max(1, (val / max) * 40)
+}
+
+function formatBytes(b) {
+  if (b < 1024) return b + ' B'
+  if (b < 1048576) return (b / 1024).toFixed(1) + ' KB'
+  if (b < 1073741824) return (b / 1048576).toFixed(1) + ' MB'
+  return (b / 1073741824).toFixed(2) + ' GB'
+}
+
+onMounted(() => { fetchStats(); timer = setInterval(fetchStats, 1000) })
+onUnmounted(() => { clearInterval(timer) })
+</script>
+
+<style scoped>
+h3 { font-size: 15px; margin-bottom: 12px; color: var(--accent); }
+.flow-selector { margin-bottom: 12px; display: flex; align-items: center; gap: 8px; }
+.flow-selector label { font-size: 12px; color: var(--muted); }
+.flow-selector select { flex: 1; }
+.stats-grid { display: grid; grid-template-columns: repeat(3, 1fr); gap: 10px; margin-bottom: 12px; }
+.stat-card { background: var(--card-bg); border: 1px solid var(--border); border-radius: var(--radius); padding: 10px; text-align: center; }
+.stat-value { font-size: 22px; font-weight: 700; font-family: monospace; }
+.stat-value.tx { color: var(--accent); }
+.stat-value.rx { color: var(--success); }
+.stat-label { font-size: 11px; color: var(--muted); margin-top: 2px; }
+.loss-ok { color: var(--success); }
+.loss-med { color: var(--warning); }
+.loss-high { color: var(--danger); }
+.totals { display: flex; gap: 16px; font-size: 12px; color: var(--muted); margin-bottom: 16px; flex-wrap: wrap; }
+.history-chart { background: var(--card-bg); border: 1px solid var(--border); border-radius: var(--radius); padding: 12px; }
+.chart-header { font-size: 12px; color: var(--muted); margin-bottom: 8px; }
+.sparkline { display: flex; align-items: flex-end; gap: 1px; height: 50px; }
+.spark-bar { flex: 1; display: flex; flex-direction: column; justify-content: flex-end; gap: 1px; }
+.spark-tx { background: var(--accent); border-radius: 1px; min-width: 2px; }
+.spark-rx { background: var(--success); border-radius: 1px; min-width: 2px; }
+</style>
--- a/traffic-gen-ui/src/components/StatusBar.vue
+++ b/traffic-gen-ui/src/components/StatusBar.vue
@ -0,0 +1,37 @@
+<template>
+  <div class="status-bar">
+    <div class="status-badges">
+      <span class="badge" :class="connected ? 'badge-ok' : 'badge-err'">
+        {{ connected ? 'Connected' : 'Offline' }}
+      </span>
+      <span v-if="health" class="badge badge-info">
+        Mode: {{ health.mode || 'sender' }}
+      </span>
+      <span v-if="health" class="badge badge-info">
+        Flows: {{ health.active_flows || 0 }}
+      </span>
+      <span v-if="health" class="badge badge-info">
+        Tests: {{ health.active_tests || 0 }}
+      </span>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { computed } from 'vue'
+const props = defineProps({ health: Object, apiError: String })
+const connected = computed(() => !props.apiError && props.health)
+</script>
+
+<style scoped>
+.status-bar { display: flex; align-items: center; gap: 8px; }
+.status-badges { display: flex; gap: 6px; flex-wrap: wrap; }
+.badge {
+  font-size: 11px; padding: 3px 10px; border-radius: 12px;
+  font-weight: 600; letter-spacing: 0.03em;
+}
+.badge-ok { background: rgba(72,187,120,0.15); color: var(--success); }
+.badge-err { background: rgba(252,129,129,0.15); color: var(--danger); animation: pulse 1.5s infinite; }
+.badge-info { background: rgba(79,156,249,0.12); color: var(--accent); }
+@keyframes pulse { 0%,100% { opacity: 1; } 50% { opacity: 0.5; } }
+</style>
--- a/traffic-gen-ui/src/components/TestBuilder.vue
+++ b/traffic-gen-ui/src/components/TestBuilder.vue
@ -0,0 +1,126 @@
+<template>
+  <div class="test-builder">
+    <h3>RFC 2544 Test</h3>
+
+    <div class="form-row">
+      <label>Test Type</label>
+      <select v-model="form.type">
+        <option value="throughput">Throughput (binary search for max rate)</option>
+        <option value="latency">Latency (measure RTT)</option>
+        <option value="frame_loss">Frame Loss (loss vs rate curve)</option>
+        <option value="back_to_back">Back-to-Back (max burst)</option>
+      </select>
+    </div>
+
+    <div class="form-row">
+      <label>Base Flow</label>
+      <select v-model="form.flow_id">
+        <option value="" disabled>Select a flow...</option>
+        <option v-for="f in flows" :key="f.id" :value="f.id">
+          {{ f.name || f.dst_ip }} ({{ f.protocol }})
+        </option>
+      </select>
+    </div>
+
+    <div class="form-row">
+      <label>Frame Sizes</label>
+      <div class="frame-sizes">
+        <label v-for="s in standardSizes" :key="s" class="checkbox-label">
+          <input type="checkbox" :value="s" v-model="form.frame_sizes" /> {{ s }}
+        </label>
+      </div>
+    </div>
+
+    <div class="form-row-pair">
+      <div class="form-row">
+        <label>Trial Duration (sec)</label>
+        <input v-model.number="form.trial_duration" type="number" min="5" max="300" />
+      </div>
+      <div class="form-row">
+        <label>Max Rate (pps)</label>
+        <input v-model.number="form.max_rate_pps" type="number" min="10" max="100000" />
+      </div>
+    </div>
+
+    <div class="form-row">
+      <label>Acceptable Loss %</label>
+      <input v-model.number="form.acceptable_loss_pct" type="number" min="0" max="100" step="0.1" />
+    </div>
+
+    <button class="btn btn-accent" @click="create" :disabled="!form.flow_id">
+      Create & Run Test
+    </button>
+
+    <div class="presets-section">
+      <h4>Quick Presets</h4>
+      <div class="preset-list">
+        <button v-for="(p, name) in presets" :key="name" class="btn-preset" @click="loadPreset(name)">
+          <strong>{{ name }}</strong>
+          <span>{{ p.description }}</span>
+        </button>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { reactive, ref, onMounted } from 'vue'
+import { api } from '../api.js'
+
+const props = defineProps({ flows: Array })
+const emit = defineEmits(['created', 'refresh'])
+
+const standardSizes = [64, 128, 256, 512, 1024, 1280, 1518]
+const presets = ref({})
+
+const form = reactive({
+  type: 'throughput',
+  flow_id: '',
+  frame_sizes: [64, 512, 1518],
+  trial_duration: 30,
+  max_rate_pps: 10000,
+  acceptable_loss_pct: 0.0,
+})
+
+onMounted(async () => {
+  try { const r = await api.presets(); presets.value = r.presets || r } catch (_) {}
+})
+
+async function create() {
+  try {
+    const test = await api.createTest({ ...form })
+    await api.startTest(test.id)
+    emit('created')
+  } catch (e) { alert(e.message) }
+}
+
+async function loadPreset(name) {
+  try {
+    await api.loadPreset(name, {})
+    emit('refresh')
+  } catch (e) { alert(e.message) }
+}
+</script>
+
+<style scoped>
+h3 { font-size: 15px; margin-bottom: 12px; color: var(--accent); }
+h4 { font-size: 13px; margin: 16px 0 8px; color: var(--muted); }
+.form-row { margin-bottom: 10px; }
+.form-row label { display: block; font-size: 11px; color: var(--muted); margin-bottom: 3px; text-transform: uppercase; letter-spacing: 0.05em; }
+.form-row input, .form-row select { width: 100%; }
+.form-row-pair { display: grid; grid-template-columns: 1fr 1fr; gap: 8px; }
+.frame-sizes { display: flex; flex-wrap: wrap; gap: 8px; }
+.checkbox-label { font-size: 12px; display: flex; align-items: center; gap: 4px; color: var(--text); cursor: pointer; }
+.btn { padding: 8px 16px; font-weight: 600; font-size: 13px; width: 100%; margin-top: 8px; }
+.btn-accent { background: var(--accent); color: #fff; }
+.btn-accent:disabled { opacity: 0.4; }
+.preset-list { display: flex; flex-direction: column; gap: 6px; }
+.btn-preset {
+  display: flex; flex-direction: column; align-items: flex-start;
+  padding: 8px 12px; background: var(--card-bg); border: 1px solid var(--border);
+  border-radius: var(--radius); text-align: left;
+}
+.btn-preset:hover { border-color: var(--accent); }
+.btn-preset strong { font-size: 12px; color: var(--accent); }
+.btn-preset span { font-size: 11px; color: var(--muted); }
+</style>
--- a/traffic-gen-ui/src/components/TestRunner.vue
+++ b/traffic-gen-ui/src/components/TestRunner.vue
@ -0,0 +1,97 @@
+<template>
+  <div class="test-runner">
+    <div v-if="!tests.length" class="empty">No tests created yet. Use the Test Builder tab.</div>
+    <div v-for="t in tests" :key="t.id" class="test-card" :class="'state-' + t.state">
+      <div class="test-header">
+        <div>
+          <strong>{{ t.type }}</strong>
+          <span class="test-state">{{ t.state }}</span>
+        </div>
+        <div class="test-actions">
+          <button v-if="t.state === 'idle'" class="btn-sm btn-go" @click="start(t.id)">Start</button>
+          <button v-if="t.state === 'running'" class="btn-sm btn-stop" @click="stop(t.id)">Stop</button>
+        </div>
+      </div>
+
+      <div v-if="t.state === 'running'" class="progress-section">
+        <div class="progress-label">Running {{ t.type }} test...</div>
+        <div class="progress-bar">
+          <div class="progress-fill" :style="{ width: progressPct(t) + '%' }"></div>
+        </div>
+      </div>
+
+      <div v-if="t.state === 'complete' && t.results" class="results-preview">
+        <table>
+          <thead>
+            <tr>
+              <th>Frame Size</th>
+              <th v-if="t.type === 'throughput'">Max Rate (pps)</th>
+              <th v-if="t.type === 'latency'">Avg Latency (ms)</th>
+              <th v-if="t.type === 'frame_loss'">Loss @ Max %</th>
+              <th v-if="t.type === 'back_to_back'">Max Burst</th>
+            </tr>
+          </thead>
+          <tbody>
+            <tr v-for="(val, size) in t.results" :key="size">
+              <td>{{ size }} B</td>
+              <td v-if="t.type === 'throughput'">{{ val.max_rate_pps || val }}</td>
+              <td v-if="t.type === 'latency'">{{ val.avg_ms || val }}</td>
+              <td v-if="t.type === 'frame_loss'">{{ val.loss_pct || val }}%</td>
+              <td v-if="t.type === 'back_to_back'">{{ val.max_burst || val }}</td>
+            </tr>
+          </tbody>
+        </table>
+      </div>
+
+      <div class="test-meta">
+        <span>Created: {{ t.created_at || '-' }}</span>
+        <span v-if="t.started_at">Started: {{ t.started_at }}</span>
+        <span v-if="t.completed_at">Completed: {{ t.completed_at }}</span>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { api } from '../api.js'
+
+const props = defineProps({ tests: Array })
+const emit = defineEmits(['refresh'])
+
+function progressPct(t) {
+  if (!t.results || !t.frame_sizes) return 20
+  const done = Object.keys(t.results).length
+  return Math.min(95, (done / t.frame_sizes.length) * 100)
+}
+
+async function start(id) {
+  try { await api.startTest(id); emit('refresh') } catch (e) { alert(e.message) }
+}
+async function stop(id) {
+  try { await api.stopTest(id); emit('refresh') } catch (e) { alert(e.message) }
+}
+</script>
+
+<style scoped>
+.empty { color: var(--muted); padding: 20px; text-align: center; }
+.test-card { background: var(--card-bg); border: 1px solid var(--border); border-radius: var(--radius); padding: 12px; margin-bottom: 10px; }
+.test-card.state-running { border-color: var(--accent); }
+.test-card.state-complete { border-color: var(--success); }
+.test-header { display: flex; justify-content: space-between; align-items: center; margin-bottom: 8px; }
+.test-header strong { font-size: 14px; text-transform: capitalize; }
+.test-state { font-size: 11px; padding: 2px 8px; border-radius: 10px; margin-left: 8px; font-weight: 600; background: rgba(113,128,150,0.2); color: var(--muted); }
+.state-running .test-state { background: rgba(79,156,249,0.15); color: var(--accent); }
+.state-complete .test-state { background: rgba(72,187,120,0.15); color: var(--success); }
+.test-actions { display: flex; gap: 4px; }
+.btn-sm { padding: 3px 10px; font-size: 11px; font-weight: 600; border-radius: 6px; }
+.btn-go { background: var(--success); color: #fff; }
+.btn-stop { background: var(--warning); color: #000; }
+.progress-section { margin: 8px 0; }
+.progress-label { font-size: 12px; color: var(--muted); margin-bottom: 4px; }
+.progress-bar { height: 6px; background: var(--border); border-radius: 3px; overflow: hidden; }
+.progress-fill { height: 100%; background: var(--accent); border-radius: 3px; transition: width 0.5s; }
+.results-preview table { width: 100%; border-collapse: collapse; margin-top: 8px; }
+.results-preview th { font-size: 11px; color: var(--muted); text-align: left; padding: 4px 8px; border-bottom: 1px solid var(--border); }
+.results-preview td { font-size: 13px; padding: 4px 8px; font-family: monospace; }
+.test-meta { display: flex; gap: 12px; margin-top: 8px; font-size: 11px; color: var(--muted); }
+</style>
--- a/traffic-gen-ui/src/main.js
+++ b/traffic-gen-ui/src/main.js
@ -0,0 +1,3 @@
+import { createApp } from 'vue'
+import App from './App.vue'
+createApp(App).mount('#app')
--- a/traffic-gen-ui/vite.config.js
+++ b/traffic-gen-ui/vite.config.js
@ -0,0 +1,14 @@
+import { defineConfig } from 'vite'
+import vue from '@vitejs/plugin-vue'
+
+export default defineConfig({
+  plugins: [vue()],
+  server: {
+    proxy: {
+      '/api': {
+        target: 'http://localhost:5051',
+        rewrite: path => path.replace(/^\/api/, '')
+      }
+    }
+  }
+})
--- a/traffic-gen/Dockerfile
+++ b/traffic-gen/Dockerfile
@ -0,0 +1,8 @@
+FROM python:3.11-slim
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    tcpreplay libpcap-dev procps && rm -rf /var/lib/apt/lists/*
+RUN pip install --no-cache-dir flask scapy psutil
+COPY . /traffic-gen/
+WORKDIR /traffic-gen
+EXPOSE 5051
+CMD ["python3", "server.py"]
--- a/traffic-gen/pycache/server.cpython-310.pyc
+++ b/traffic-gen/pycache/server.cpython-310.pyc
--- a/traffic-gen/engine/init.py
+++ b/traffic-gen/engine/init.py
@ -0,0 +1 @@
+# traffic-gen engine package
--- a/traffic-gen/engine/pycache/init.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/init.cpython-310.pyc
--- a/traffic-gen/engine/pycache/packet_builder.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/packet_builder.cpython-310.pyc
--- a/traffic-gen/engine/pycache/responder.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/responder.cpython-310.pyc
--- a/traffic-gen/engine/pycache/rfc2544.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/rfc2544.cpython-310.pyc
--- a/traffic-gen/engine/pycache/sender.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/sender.cpython-310.pyc
--- a/traffic-gen/engine/pycache/stats.cpython-310.pyc
+++ b/traffic-gen/engine/pycache/stats.cpython-310.pyc
--- a/traffic-gen/engine/packet_builder.py
+++ b/traffic-gen/engine/packet_builder.py
@ -0,0 +1,120 @@
+"""
+Packet Builder - constructs Scapy packets from flow configuration.
+
+Each generated packet embeds:
+  - Magic bytes b'TGEN' (4 bytes)
+  - Sequence number (4 bytes, big-endian)
+  - Sender timestamp in nanoseconds (8 bytes, big-endian)
+  - Padding to reach requested frame_size
+"""
+
+import struct
+import time
+
+from scapy.all import (
+    Ether, IP, UDP, TCP, ICMP, Dot1Q, Raw, conf,
+)
+
+MAGIC = b'TGEN'
+HEADER_LEN = 4 + 4 + 8  # magic + seq + timestamp_ns
+
+
+def _build_payload(seq: int, frame_size: int, header_overhead: int) -> Raw:
+    """Build payload with magic bytes, sequence number, timestamp placeholder,
+    and padding to reach the desired frame_size."""
+    timestamp_ns = time.time_ns()
+    header = MAGIC + struct.pack('!I', seq) + struct.pack('!Q', timestamp_ns)
+    # frame_size includes Ethernet header (14) + FCS (4) in standard accounting,
+    # but Scapy doesn't add FCS, so we target frame_size - 4 total bytes on wire.
+    # header_overhead accounts for Ether + IP + L4 headers already present.
+    pad_len = max(0, frame_size - 4 - header_overhead - HEADER_LEN)
+    return Raw(load=header + (b'\x00' * pad_len))
+
+
+def stamp_payload(payload_bytes: bytes, seq: int) -> bytes:
+    """Re-stamp an existing payload with a new sequence number and fresh timestamp."""
+    timestamp_ns = time.time_ns()
+    return (
+        MAGIC
+        + struct.pack('!I', seq)
+        + struct.pack('!Q', timestamp_ns)
+        + payload_bytes[HEADER_LEN:]
+    )
+
+
+def parse_payload(payload_bytes: bytes):
+    """Extract (seq, timestamp_ns) from a TGEN payload, or None if invalid."""
+    if len(payload_bytes) < HEADER_LEN:
+        return None
+    if payload_bytes[:4] != MAGIC:
+        return None
+    seq = struct.unpack('!I', payload_bytes[4:8])[0]
+    timestamp_ns = struct.unpack('!Q', payload_bytes[8:16])[0]
+    return seq, timestamp_ns
+
+
+def build_packet(flow_config: dict, seq: int = 0):
+    """Build a Scapy packet from a flow configuration dict.
+
+    Required keys:
+        dst_ip, protocol
+
+    Optional keys:
+        src_mac, dst_mac, src_ip, src_port, dst_port, dscp, vlan_id, frame_size
+    """
+    protocol = flow_config.get('protocol', 'udp').lower()
+    frame_size = flow_config.get('frame_size', 512)
+
+    # --- Layer 2 ---
+    src_mac = flow_config.get('src_mac', 'auto')
+    dst_mac = flow_config.get('dst_mac')
+
+    ether_kwargs = {}
+    if src_mac and src_mac != 'auto':
+        ether_kwargs['src'] = src_mac
+    if dst_mac:
+        ether_kwargs['dst'] = dst_mac
+
+    pkt = Ether(**ether_kwargs)
+    header_overhead = 14  # Ethernet
+
+    # --- VLAN ---
+    vlan_id = flow_config.get('vlan_id')
+    if vlan_id is not None:
+        pkt = pkt / Dot1Q(vlan=int(vlan_id))
+        header_overhead += 4
+
+    # --- Layer 3 ---
+    ip_kwargs = {'dst': flow_config['dst_ip']}
+    src_ip = flow_config.get('src_ip')
+    if src_ip:
+        ip_kwargs['src'] = src_ip
+
+    dscp = flow_config.get('dscp', 0)
+    if dscp:
+        ip_kwargs['tos'] = int(dscp) << 2
+
+    pkt = pkt / IP(**ip_kwargs)
+    header_overhead += 20  # IP (no options)
+
+    # --- Layer 4 ---
+    if protocol == 'udp':
+        src_port = flow_config.get('src_port', 12000)
+        dst_port = flow_config.get('dst_port', 5001)
+        pkt = pkt / UDP(sport=int(src_port), dport=int(dst_port))
+        header_overhead += 8
+    elif protocol == 'tcp':
+        src_port = flow_config.get('src_port', 12000)
+        dst_port = flow_config.get('dst_port', 80)
+        pkt = pkt / TCP(sport=int(src_port), dport=int(dst_port), flags='S')
+        header_overhead += 20
+    elif protocol == 'icmp':
+        pkt = pkt / ICMP()
+        header_overhead += 8
+    else:
+        raise ValueError(f'Unsupported protocol: {protocol}')
+
+    # --- Payload ---
+    pkt = pkt / _build_payload(seq, frame_size, header_overhead)
+
+    return pkt
--- a/traffic-gen/engine/responder.py
+++ b/traffic-gen/engine/responder.py
@ -0,0 +1,185 @@
+"""
+Responder - listens for TGEN-tagged packets and collects receive statistics.
+
+Two sub-modes:
+  - echo: swaps src/dst MAC and IP, sends packet back with receive timestamp
+  - log:  records rx stats only, exposed via API
+"""
+
+import logging
+import struct
+import threading
+import time
+
+from scapy.all import (
+    AsyncSniffer, Ether, IP, Raw, send, conf,
+)
+
+from engine.packet_builder import MAGIC, HEADER_LEN, parse_payload
+
+log = logging.getLogger(__name__)
+conf.verb = 0
+
+
+class Responder:
+    """Listens for TGEN packets on an interface and collects stats."""
+
+    def __init__(self, mode: str = 'log'):
+        """
+        Args:
+            mode: 'echo' to reflect packets back, 'log' to only record stats.
+        """
+        self._mode = mode
+        self._lock = threading.Lock()
+        self._sniffer = None
+        self._running = False
+
+        # Stats
+        self._rx_packets = 0
+        self._rx_bytes = 0
+        self._latency_samples = []   # list of (latency_ms,)
+        self._seen_seqs = set()
+        self._last_seq = -1
+        self._out_of_order = 0
+        self._duplicates = 0
+
+    # ------------------------------------------------------------------
+    # Control
+    # ------------------------------------------------------------------
+
+    def start(self, interface: str = None):
+        """Start sniffing for TGEN packets."""
+        if self._running:
+            log.warning('Responder already running')
+            return
+
+        bpf_filter = 'ip'  # broad filter; we check magic in callback
+        kwargs = {
+            'prn': self._handle_packet,
+            'store': False,
+            'filter': bpf_filter,
+        }
+        if interface:
+            kwargs['iface'] = interface
+
+        self._sniffer = AsyncSniffer(**kwargs)
+        self._sniffer.start()
+        self._running = True
+        log.info('Responder started on interface=%s mode=%s', interface or 'all', self._mode)
+
+    def stop(self):
+        """Stop sniffing."""
+        if self._sniffer and self._running:
+            self._sniffer.stop()
+            self._running = False
+            log.info('Responder stopped')
+
+    def is_running(self) -> bool:
+        return self._running
+
+    # ------------------------------------------------------------------
+    # Stats
+    # ------------------------------------------------------------------
+
+    def get_stats(self) -> dict:
+        with self._lock:
+            latency = {}
+            if self._latency_samples:
+                vals = self._latency_samples
+                latency = {
+                    'min_ms': round(min(vals), 3),
+                    'max_ms': round(max(vals), 3),
+                    'avg_ms': round(sum(vals) / len(vals), 3),
+                    'jitter_ms': round(
+                        sum(abs(vals[i] - vals[i - 1]) for i in range(1, len(vals))) / max(1, len(vals) - 1),
+                        3
+                    ) if len(vals) > 1 else 0.0,
+                    'samples': len(vals),
+                }
+            return {
+                'rx_packets': self._rx_packets,
+                'rx_bytes': self._rx_bytes,
+                'out_of_order': self._out_of_order,
+                'duplicates': self._duplicates,
+                'latency': latency,
+                'running': self._running,
+            }
+
+    def reset_stats(self):
+        with self._lock:
+            self._rx_packets = 0
+            self._rx_bytes = 0
+            self._latency_samples = []
+            self._seen_seqs.clear()
+            self._last_seq = -1
+            self._out_of_order = 0
+            self._duplicates = 0
+        log.info('Responder stats reset')
+
+    # ------------------------------------------------------------------
+    # Packet handling
+    # ------------------------------------------------------------------
+
+    def _handle_packet(self, pkt):
+        """Process a received packet, checking for TGEN magic bytes."""
+        if not pkt.haslayer(Raw):
+            return
+
+        payload = bytes(pkt[Raw].load)
+        parsed = parse_payload(payload)
+        if parsed is None:
+            return
+
+        seq, sender_ts_ns = parsed
+        rx_time_ns = time.time_ns()
+        pkt_len = len(bytes(pkt))
+
+        # Compute one-way latency (only meaningful if clocks are synced)
+        latency_ms = (rx_time_ns - sender_ts_ns) / 1_000_000
+
+        with self._lock:
+            self._rx_packets += 1
+            self._rx_bytes += pkt_len
+
+            # Duplicate detection
+            if seq in self._seen_seqs:
+                self._duplicates += 1
+            else:
+                self._seen_seqs.add(seq)
+                # Keep set bounded
+                if len(self._seen_seqs) > 100000:
+                    # Remove oldest entries (approximate)
+                    to_remove = sorted(self._seen_seqs)[:50000]
+                    self._seen_seqs -= set(to_remove)
+
+            # Out-of-order detection
+            if seq < self._last_seq and seq not in self._seen_seqs:
+                self._out_of_order += 1
+            self._last_seq = seq
+
+            # Record latency (only if plausible: 0 < latency < 60s)
+            if 0 < latency_ms < 60000:
+                self._latency_samples.append(latency_ms)
+                if len(self._latency_samples) > 10000:
+                    self._latency_samples = self._latency_samples[-5000:]
+
+        # Echo mode: swap and send back
+        if self._mode == 'echo' and pkt.haslayer(Ether) and pkt.haslayer(IP):
+            try:
+                echo_pkt = pkt.copy()
+                # Swap MACs
+                echo_pkt[Ether].src, echo_pkt[Ether].dst = pkt[Ether].dst, pkt[Ether].src
+                # Swap IPs
+                echo_pkt[IP].src, echo_pkt[IP].dst = pkt[IP].dst, pkt[IP].src
+                # Append receive timestamp to payload
+                rx_ts_bytes = struct.pack('!Q', rx_time_ns)
+                echo_pkt[Raw].load = payload + rx_ts_bytes
+                # Clear checksums so Scapy recalculates
+                del echo_pkt[IP].chksum
+                if echo_pkt.haslayer('UDP'):
+                    del echo_pkt['UDP'].chksum
+                elif echo_pkt.haslayer('TCP'):
+                    del echo_pkt['TCP'].chksum
+                send(echo_pkt[IP], verbose=0)
+            except Exception as e:
+                log.debug('Echo send error: %s', e)
--- a/traffic-gen/engine/rfc2544.py
+++ b/traffic-gen/engine/rfc2544.py
@ -0,0 +1,336 @@
+"""
+RFC 2544 test implementations:
+  - ThroughputTest: binary search for max zero-loss throughput
+  - LatencyTest: measure latency at a given rate
+  - FrameLossTest: measure loss at decreasing rates
+  - BackToBackTest: find max burst length with zero loss
+"""
+
+import logging
+import threading
+import time
+from typing import Dict, List, Optional
+
+from scapy.all import send, sr, conf, IP, ICMP
+
+from engine.packet_builder import build_packet, parse_payload
+
+log = logging.getLogger(__name__)
+conf.verb = 0
+
+
+class _BaseTest:
+    """Base class for RFC 2544 tests."""
+
+    def __init__(self, test_id: str, flow_config: dict, frame_sizes: List[int],
+                 trial_duration: float = 60, max_rate_pps: int = 10000,
+                 acceptable_loss_pct: float = 0.0):
+        self.test_id = test_id
+        self.flow_config = dict(flow_config)
+        self.frame_sizes = frame_sizes
+        self.trial_duration = trial_duration
+        self.max_rate_pps = max_rate_pps
+        self.acceptable_loss_pct = acceptable_loss_pct
+
+        self.state = 'idle'      # idle -> running -> complete/error
+        self.results = {}
+        self.error = None
+        self.started_at = None
+        self.completed_at = None
+
+        self._thread: Optional[threading.Thread] = None
+        self._stop_event = threading.Event()
+        self._lock = threading.Lock()
+
+    def start(self):
+        if self.state == 'running':
+            return
+        self._stop_event.clear()
+        self.state = 'running'
+        self.started_at = time.strftime('%Y-%m-%dT%H:%M:%SZ', time.gmtime())
+        self.results = {}
+        self.error = None
+        self._thread = threading.Thread(target=self._run_safe, daemon=True,
+                                        name=f'test-{self.test_id[:8]}')
+        self._thread.start()
+
+    def stop(self):
+        self._stop_event.set()
+        if self._thread and self._thread.is_alive():
+            self._thread.join(timeout=self.trial_duration + 5)
+        if self.state == 'running':
+            self.state = 'error'
+            self.error = 'Cancelled by user'
+
+    def _run_safe(self):
+        try:
+            self._run()
+            if self.state == 'running':
+                self.state = 'complete'
+        except Exception as e:
+            log.error('Test %s error: %s', self.test_id[:8], e)
+            self.state = 'error'
+            self.error = str(e)
+        finally:
+            self.completed_at = time.strftime('%Y-%m-%dT%H:%M:%SZ', time.gmtime())
+
+    def _run(self):
+        raise NotImplementedError
+
+    def _is_stopped(self) -> bool:
+        return self._stop_event.is_set()
+
+    def _send_trial(self, frame_size: int, rate_pps: int, duration: float):
+        """Send packets at a given rate for a duration. Returns (tx_count, rx_count, latencies)."""
+        flow = dict(self.flow_config)
+        flow['frame_size'] = frame_size
+
+        interval = 1.0 / rate_pps if rate_pps > 0 else 1.0
+        tx_count = 0
+        rx_count = 0
+        latencies = []
+        protocol = flow.get('protocol', 'udp').lower()
+
+        start = time.time()
+        seq = 0
+
+        while (time.time() - start) < duration and not self._is_stopped():
+            pkt = build_packet(flow, seq=seq)
+            seq += 1
+
+            if protocol == 'icmp':
+                # Use sr() for ICMP to get responses
+                answered, _ = sr(pkt[IP], timeout=1, verbose=0)
+                tx_count += 1
+                for sent_pkt, recv_pkt in answered:
+                    rx_count += 1
+                    rtt_ms = (recv_pkt.time - sent_pkt.sent_time) * 1000
+                    latencies.append(rtt_ms)
+            else:
+                send(pkt[IP], verbose=0)
+                tx_count += 1
+
+            # Rate limiting
+            elapsed = time.time() - start
+            expected_sent = elapsed * rate_pps
+            if tx_count > expected_sent:
+                sleep_time = (tx_count - expected_sent) / rate_pps
+                if sleep_time > 0:
+                    self._stop_event.wait(min(sleep_time, 0.1))
+
+        # For non-ICMP, we can't easily measure rx without a responder.
+        # rx_count stays 0 for UDP/TCP unless a responder is configured.
+        return tx_count, rx_count, latencies
+
+    def get_info(self) -> dict:
+        return {
+            'test_id': self.test_id,
+            'type': self.__class__.__name__,
+            'state': self.state,
+            'results': self.results,
+            'error': self.error,
+            'started_at': self.started_at,
+            'completed_at': self.completed_at,
+        }
+
+
+class ThroughputTest(_BaseTest):
+    """Binary search for maximum throughput with acceptable loss."""
+
+    def _run(self):
+        for fs in self.frame_sizes:
+            if self._is_stopped():
+                break
+
+            low = 0
+            high = self.max_rate_pps
+            best_rate = 0
+            convergence_threshold = max(1, int(self.max_rate_pps * 0.01))
+
+            log.info('Throughput test: frame_size=%d, searching [%d, %d] pps', fs, low, high)
+
+            while (high - low) > convergence_threshold and not self._is_stopped():
+                mid = (low + high) // 2
+                if mid == 0:
+                    break
+
+                tx, rx, _ = self._send_trial(fs, mid, self.trial_duration)
+
+                if tx == 0:
+                    loss_pct = 100.0
+                else:
+                    # For ICMP we have rx; for UDP assume zero loss if no responder
+                    protocol = self.flow_config.get('protocol', 'udp').lower()
+                    if protocol == 'icmp':
+                        loss_pct = ((tx - rx) / tx) * 100
+                    else:
+                        # Without responder, assume success (user should use responder for accurate test)
+                        loss_pct = 0.0
+
+                log.info('  frame=%d rate=%d tx=%d rx=%d loss=%.2f%%',
+                         fs, mid, tx, rx, loss_pct)
+
+                if loss_pct <= self.acceptable_loss_pct:
+                    best_rate = mid
+                    low = mid + 1
+                else:
+                    high = mid - 1
+
+            self.results[str(fs)] = {
+                'max_throughput_pps': best_rate,
+                'frame_size': fs,
+            }
+            log.info('Throughput result: frame_size=%d -> %d pps', fs, best_rate)
+
+
+class LatencyTest(_BaseTest):
+    """Measure latency at a specified rate."""
+
+    def _run(self):
+        rate = self.flow_config.get('rate_pps', 100)
+
+        for fs in self.frame_sizes:
+            if self._is_stopped():
+                break
+
+            log.info('Latency test: frame_size=%d at %d pps for %ds', fs, rate, self.trial_duration)
+            _, _, latencies = self._send_trial(fs, rate, self.trial_duration)
+
+            if latencies:
+                avg_ms = sum(latencies) / len(latencies)
+                min_ms = min(latencies)
+                max_ms = max(latencies)
+                jitter_ms = (
+                    sum(abs(latencies[i] - latencies[i - 1]) for i in range(1, len(latencies)))
+                    / max(1, len(latencies) - 1)
+                ) if len(latencies) > 1 else 0.0
+
+                self.results[str(fs)] = {
+                    'frame_size': fs,
+                    'min_ms': round(min_ms, 3),
+                    'avg_ms': round(avg_ms, 3),
+                    'max_ms': round(max_ms, 3),
+                    'jitter_ms': round(jitter_ms, 3),
+                    'samples': len(latencies),
+                }
+            else:
+                self.results[str(fs)] = {
+                    'frame_size': fs,
+                    'min_ms': None, 'avg_ms': None, 'max_ms': None,
+                    'jitter_ms': None, 'samples': 0,
+                    'note': 'No responses received (use ICMP or configure responder)',
+                }
+            log.info('Latency result: frame_size=%d -> %s', fs, self.results[str(fs)])
+
+
+class FrameLossTest(_BaseTest):
+    """Measure frame loss at decreasing rates (100%, 90%, 80%, ...)."""
+
+    def _run(self):
+        for fs in self.frame_sizes:
+            if self._is_stopped():
+                break
+
+            results_for_size = []
+            for pct in range(100, 0, -10):
+                if self._is_stopped():
+                    break
+
+                rate = int(self.max_rate_pps * pct / 100)
+                if rate == 0:
+                    continue
+
+                log.info('FrameLoss test: frame_size=%d rate=%d (%d%%)', fs, rate, pct)
+                tx, rx, _ = self._send_trial(fs, rate, self.trial_duration)
+
+                protocol = self.flow_config.get('protocol', 'udp').lower()
+                if tx > 0 and protocol == 'icmp':
+                    loss_pct = ((tx - rx) / tx) * 100
+                else:
+                    loss_pct = 0.0  # Cannot measure without responder for non-ICMP
+
+                results_for_size.append({
+                    'rate_pct': pct,
+                    'rate_pps': rate,
+                    'tx_packets': tx,
+                    'rx_packets': rx,
+                    'loss_pct': round(loss_pct, 3),
+                })
+
+            self.results[str(fs)] = results_for_size
+
+
+class BackToBackTest(_BaseTest):
+    """Find maximum burst length with zero loss."""
+
+    def _run(self):
+        for fs in self.frame_sizes:
+            if self._is_stopped():
+                break
+
+            low = 1
+            high = self.max_rate_pps  # Use max_rate_pps as max burst length
+            best_burst = 0
+            convergence = max(1, high // 100)
+
+            log.info('BackToBack test: frame_size=%d searching burst [%d, %d]', fs, low, high)
+
+            while (high - low) > convergence and not self._is_stopped():
+                mid = (low + high) // 2
+                if mid == 0:
+                    break
+
+                flow = dict(self.flow_config)
+                flow['frame_size'] = fs
+                protocol = flow.get('protocol', 'udp').lower()
+
+                # Send burst of 'mid' packets as fast as possible
+                tx_count = 0
+                rx_count = 0
+                for seq in range(mid):
+                    if self._is_stopped():
+                        break
+                    pkt = build_packet(flow, seq=seq)
+                    if protocol == 'icmp':
+                        answered, _ = sr(pkt[IP], timeout=0.5, verbose=0)
+                        tx_count += 1
+                        rx_count += len(answered)
+                    else:
+                        send(pkt[IP], verbose=0)
+                        tx_count += 1
+
+                if protocol == 'icmp' and tx_count > 0:
+                    loss_pct = ((tx_count - rx_count) / tx_count) * 100
+                else:
+                    loss_pct = 0.0  # Can't measure without responder
+
+                log.info('  burst=%d tx=%d rx=%d loss=%.2f%%', mid, tx_count, rx_count, loss_pct)
+
+                if loss_pct <= self.acceptable_loss_pct:
+                    best_burst = mid
+                    low = mid + 1
+                else:
+                    high = mid - 1
+
+            self.results[str(fs)] = {
+                'frame_size': fs,
+                'max_burst_frames': best_burst,
+            }
+            log.info('BackToBack result: frame_size=%d -> %d frames', fs, best_burst)
+
+
+# Factory function
+TEST_TYPES = {
+    'throughput': ThroughputTest,
+    'latency': LatencyTest,
+    'frame_loss': FrameLossTest,
+    'back_to_back': BackToBackTest,
+}
+
+
+def create_test(test_id: str, test_type: str, flow_config: dict, **kwargs):
+    """Create an RFC 2544 test instance by type name."""
+    cls = TEST_TYPES.get(test_type)
+    if cls is None:
+        raise ValueError(f'Unknown test type: {test_type}. Available: {list(TEST_TYPES)}')
+    return cls(test_id=test_id, flow_config=flow_config, **kwargs)
--- a/traffic-gen/engine/sender.py
+++ b/traffic-gen/engine/sender.py
@ -0,0 +1,247 @@
+"""
+FlowSender - manages traffic generation with background threads per flow.
+"""
+
+import logging
+import shutil
+import threading
+import time
+import urllib.request
+import json
+
+from scapy.all import send, sendpfast, sr, conf
+
+from engine.packet_builder import build_packet, stamp_payload
+
+log = logging.getLogger(__name__)
+
+# Suppress Scapy verbosity globally
+conf.verb = 0
+
+HAS_TCPREPLAY = shutil.which('tcpreplay') is not None
+
+
+class FlowSender:
+    """Manages sending threads for multiple flows."""
+
+    def __init__(self):
+        self._lock = threading.Lock()
+        self._flows = {}          # flow_id -> flow_config dict
+        self._threads = {}        # flow_id -> Thread
+        self._stop_events = {}    # flow_id -> Event
+        self._stats = {}          # flow_id -> {tx_packets, tx_bytes, ...}
+
+    # ------------------------------------------------------------------
+    # Flow CRUD
+    # ------------------------------------------------------------------
+
+    def add_flow(self, flow_id: str, flow_config: dict):
+        with self._lock:
+            self._flows[flow_id] = flow_config
+            self._stats[flow_id] = {
+                'tx_packets': 0, 'tx_bytes': 0,
+                'rx_packets': 0, 'rx_bytes': 0,
+                'latency_samples': [],
+            }
+
+    def get_flow(self, flow_id: str):
+        with self._lock:
+            return self._flows.get(flow_id)
+
+    def get_all_flows(self):
+        with self._lock:
+            return dict(self._flows)
+
+    def update_flow(self, flow_id: str, updates: dict):
+        with self._lock:
+            if flow_id not in self._flows:
+                return False
+            self._flows[flow_id].update(updates)
+            return True
+
+    def remove_flow(self, flow_id: str):
+        self.stop(flow_id)
+        with self._lock:
+            self._flows.pop(flow_id, None)
+            self._stats.pop(flow_id, None)
+
+    # ------------------------------------------------------------------
+    # Start / Stop
+    # ------------------------------------------------------------------
+
+    def start(self, flow_id: str):
+        with self._lock:
+            if flow_id not in self._flows:
+                raise KeyError(f'Flow {flow_id} not found')
+            if flow_id in self._threads and self._threads[flow_id].is_alive():
+                return  # already running
+            self._flows[flow_id]['state'] = 'running'
+            self._stats[flow_id] = {
+                'tx_packets': 0, 'tx_bytes': 0,
+                'rx_packets': 0, 'rx_bytes': 0,
+                'latency_samples': [],
+            }
+            stop_event = threading.Event()
+            self._stop_events[flow_id] = stop_event
+            t = threading.Thread(
+                target=self._send_loop,
+                args=(flow_id, stop_event),
+                daemon=True,
+                name=f'sender-{flow_id[:8]}',
+            )
+            self._threads[flow_id] = t
+            t.start()
+
+    def stop(self, flow_id: str):
+        with self._lock:
+            ev = self._stop_events.pop(flow_id, None)
+            if ev:
+                ev.set()
+            t = self._threads.pop(flow_id, None)
+            if flow_id in self._flows:
+                self._flows[flow_id]['state'] = 'stopped'
+        if t and t.is_alive():
+            t.join(timeout=5)
+
+    def is_running(self, flow_id: str) -> bool:
+        with self._lock:
+            t = self._threads.get(flow_id)
+            return t is not None and t.is_alive()
+
+    # ------------------------------------------------------------------
+    # Stats
+    # ------------------------------------------------------------------
+
+    def get_stats(self, flow_id: str) -> dict:
+        with self._lock:
+            s = self._stats.get(flow_id, {})
+            return dict(s)
+
+    def get_all_stats(self) -> dict:
+        with self._lock:
+            return {fid: dict(s) for fid, s in self._stats.items()}
+
+    # ------------------------------------------------------------------
+    # Internal send loop
+    # ------------------------------------------------------------------
+
+    def _send_loop(self, flow_id: str, stop_event: threading.Event):
+        with self._lock:
+            flow = dict(self._flows[flow_id])
+
+        rate_pps = flow.get('rate_pps', 1000)
+        duration = flow.get('duration', 30)
+        protocol = flow.get('protocol', 'udp').lower()
+        responder_url = flow.get('responder_url')
+        use_icmp_sr = (protocol == 'icmp' and not responder_url)
+
+        # Build template packet
+        pkt_template = build_packet(flow, seq=0)
+        pkt_bytes_len = len(bytes(pkt_template))
+
+        # Calculate sleep interval: send in batches for efficiency
+        batch_size = max(1, min(rate_pps // 10, 100))
+        interval = batch_size / rate_pps if rate_pps > 0 else 1.0
+
+        seq = 0
+        start_time = time.time()
+        last_responder_poll = 0
+        log.info('Flow %s: starting send loop at %d pps for %ds', flow_id[:8], rate_pps, duration)
+
+        try:
+            while not stop_event.is_set():
+                elapsed = time.time() - start_time
+                if duration and elapsed >= duration:
+                    break
+
+                if use_icmp_sr:
+                    # ICMP mode: use sr() to measure latency from responses
+                    pkt = build_packet(flow, seq=seq)
+                    answered, _ = sr(pkt[pkt.firstlayer().payload.__class__],
+                                     timeout=1, verbose=0)
+                    with self._lock:
+                        stats = self._stats.get(flow_id)
+                        if stats:
+                            stats['tx_packets'] += 1
+                            stats['tx_bytes'] += pkt_bytes_len
+                            for sent_pkt, recv_pkt in answered:
+                                rtt_ms = (recv_pkt.time - sent_pkt.sent_time) * 1000
+                                stats['rx_packets'] += 1
+                                stats['rx_bytes'] += len(bytes(recv_pkt))
+                                stats['latency_samples'].append(rtt_ms)
+                                # Keep only last 1000 samples
+                                if len(stats['latency_samples']) > 1000:
+                                    stats['latency_samples'] = stats['latency_samples'][-1000:]
+                    seq += 1
+                    # Rate limit for ICMP
+                    sleep_time = (1.0 / rate_pps) - (time.time() - start_time - elapsed)
+                    if sleep_time > 0:
+                        stop_event.wait(sleep_time)
+                else:
+                    # UDP/TCP mode: send batches
+                    packets = []
+                    for _ in range(batch_size):
+                        pkt = build_packet(flow, seq=seq)
+                        packets.append(pkt)
+                        seq += 1
+
+                    try:
+                        if HAS_TCPREPLAY and rate_pps >= 1000:
+                            sendpfast(packets, pps=rate_pps, loop=0)
+                        else:
+                            for p in packets:
+                                send(p[p.firstlayer().payload.__class__], verbose=0)
+                    except Exception as e:
+                        # Fallback: basic send
+                        log.debug('Send error (falling back): %s', e)
+                        for p in packets:
+                            try:
+                                send(p[p.firstlayer().payload.__class__], verbose=0)
+                            except Exception:
+                                pass
+
+                    with self._lock:
+                        stats = self._stats.get(flow_id)
+                        if stats:
+                            stats['tx_packets'] += len(packets)
+                            stats['tx_bytes'] += pkt_bytes_len * len(packets)
+
+                    # Poll responder for rx stats periodically
+                    if responder_url and (time.time() - last_responder_poll) >= 2.0:
+                        self._poll_responder(flow_id, responder_url)
+                        last_responder_poll = time.time()
+
+                    # Rate limit
+                    stop_event.wait(interval)
+
+        except Exception as e:
+            log.error('Flow %s: send loop error: %s', flow_id[:8], e)
+        finally:
+            with self._lock:
+                if flow_id in self._flows:
+                    self._flows[flow_id]['state'] = 'stopped'
+            # Final responder poll
+            if responder_url:
+                self._poll_responder(flow_id, responder_url)
+            log.info('Flow %s: send loop finished. seq=%d', flow_id[:8], seq)
+
+    def _poll_responder(self, flow_id: str, responder_url: str):
+        """Poll a responder's /responder/stats endpoint for rx metrics."""
+        try:
+            url = responder_url.rstrip('/') + '/responder/stats'
+            req = urllib.request.Request(url, method='GET')
+            req.add_header('Accept', 'application/json')
+            with urllib.request.urlopen(req, timeout=2) as resp:
+                data = json.loads(resp.read().decode())
+            with self._lock:
+                stats = self._stats.get(flow_id)
+                if stats:
+                    stats['rx_packets'] = data.get('rx_packets', 0)
+                    stats['rx_bytes'] = data.get('rx_bytes', 0)
+                    lat = data.get('latency', {})
+                    if lat.get('avg_ms') is not None:
+                        stats['latency_samples'].append(lat['avg_ms'])
+                        if len(stats['latency_samples']) > 1000:
+                            stats['latency_samples'] = stats['latency_samples'][-1000:]
+        except Exception as e:
+            log.debug('Responder poll error for flow %s: %s', flow_id[:8], e)
--- a/traffic-gen/engine/stats.py
+++ b/traffic-gen/engine/stats.py
@ -0,0 +1,119 @@
+"""
+StatsCollector - ring buffer of per-flow traffic statistics.
+
+Stores the last 300 samples (5 minutes at 1-second intervals) per flow,
+with derived rates and loss calculations.
+"""
+
+import threading
+import time
+from collections import defaultdict, deque
+
+
+class StatsCollector:
+    """Collects and stores per-flow traffic statistics in a ring buffer."""
+
+    def __init__(self, max_samples: int = 300):
+        self._lock = threading.Lock()
+        self._max_samples = max_samples
+        # flow_id -> deque of sample dicts
+        self._history = defaultdict(lambda: deque(maxlen=self._max_samples))
+        # flow_id -> previous counters for rate calculation
+        self._prev = {}
+
+    def record(self, flow_id: str, tx_packets: int, tx_bytes: int,
+               rx_packets: int = 0, rx_bytes: int = 0,
+               latency: dict = None):
+        """Record a stats sample for a flow.
+
+        Args:
+            flow_id: Flow identifier.
+            tx_packets: Cumulative transmitted packets.
+            tx_bytes: Cumulative transmitted bytes.
+            rx_packets: Cumulative received packets.
+            rx_bytes: Cumulative received bytes.
+            latency: Optional dict with min_ms, avg_ms, max_ms, jitter_ms.
+        """
+        now = time.time()
+
+        with self._lock:
+            prev = self._prev.get(flow_id)
+
+            if prev is not None:
+                dt = now - prev['time']
+                if dt > 0:
+                    d_tx_pkts = tx_packets - prev['tx_packets']
+                    d_tx_bytes = tx_bytes - prev['tx_bytes']
+                    d_rx_pkts = rx_packets - prev['rx_packets']
+                    d_rx_bytes = rx_bytes - prev['rx_bytes']
+
+                    tx_pps = d_tx_pkts / dt
+                    rx_pps = d_rx_pkts / dt
+                    tx_mbps = (d_tx_bytes * 8) / (dt * 1_000_000)
+                    rx_mbps = (d_rx_bytes * 8) / (dt * 1_000_000)
+                else:
+                    tx_pps = rx_pps = tx_mbps = rx_mbps = 0.0
+            else:
+                tx_pps = rx_pps = tx_mbps = rx_mbps = 0.0
+
+            # Loss calculation
+            loss_pct = 0.0
+            if tx_packets > 0 and rx_packets > 0:
+                loss_pct = max(0.0, ((tx_packets - rx_packets) / tx_packets) * 100)
+
+            sample = {
+                'timestamp': now,
+                'tx_packets': tx_packets,
+                'tx_bytes': tx_bytes,
+                'rx_packets': rx_packets,
+                'rx_bytes': rx_bytes,
+                'tx_pps': round(tx_pps, 2),
+                'rx_pps': round(rx_pps, 2),
+                'tx_mbps': round(tx_mbps, 4),
+                'rx_mbps': round(rx_mbps, 4),
+                'loss_pct': round(loss_pct, 3),
+            }
+
+            if latency:
+                sample['latency'] = latency
+
+            self._history[flow_id].append(sample)
+            self._prev[flow_id] = {
+                'time': now,
+                'tx_packets': tx_packets,
+                'tx_bytes': tx_bytes,
+                'rx_packets': rx_packets,
+                'rx_bytes': rx_bytes,
+            }
+
+    def get_history(self, flow_id: str, count: int = 60) -> list:
+        """Get the last N samples for a flow."""
+        with self._lock:
+            history = self._history.get(flow_id)
+            if not history:
+                return []
+            samples = list(history)
+            return samples[-count:]
+
+    def get_latest(self, flow_id: str) -> dict:
+        """Get the most recent sample for a flow."""
+        with self._lock:
+            history = self._history.get(flow_id)
+            if not history:
+                return {}
+            return dict(history[-1])
+
+    def get_all_latest(self) -> dict:
+        """Get the most recent sample for every flow."""
+        with self._lock:
+            result = {}
+            for flow_id, history in self._history.items():
+                if history:
+                    result[flow_id] = dict(history[-1])
+            return result
+
+    def remove_flow(self, flow_id: str):
+        """Remove all data for a flow."""
+        with self._lock:
+            self._history.pop(flow_id, None)
+            self._prev.pop(flow_id, None)
--- a/traffic-gen/presets/init.py
+++ b/traffic-gen/presets/init.py
@ -0,0 +1,75 @@
+"""
+Built-in test presets for the traffic generator.
+"""
+
+PRESETS = {
+    'quick_icmp': {
+        'description': 'Quick ICMP ping test to verify connectivity',
+        'flow': {
+            'protocol': 'icmp',
+            'frame_size': 64,
+            'rate_pps': 10,
+            'duration': 10,
+        },
+    },
+    'udp_flood_small': {
+        'description': 'UDP flood with 64-byte frames at 1000 pps',
+        'flow': {
+            'protocol': 'udp',
+            'dst_port': 5001,
+            'frame_size': 64,
+            'rate_pps': 1000,
+            'duration': 30,
+        },
+    },
+    'udp_flood_large': {
+        'description': 'UDP flood with 1518-byte frames at 500 pps',
+        'flow': {
+            'protocol': 'udp',
+            'dst_port': 5001,
+            'frame_size': 1518,
+            'rate_pps': 500,
+            'duration': 30,
+        },
+    },
+    'rfc2544_throughput': {
+        'description': 'RFC 2544 throughput test across standard frame sizes',
+        'flow': {
+            'protocol': 'udp',
+            'dst_port': 5001,
+            'frame_size': 64,
+            'rate_pps': 10000,
+            'duration': 60,
+        },
+        'test': {
+            'type': 'throughput',
+            'frame_sizes': [64, 128, 256, 512, 1024, 1280, 1518],
+            'trial_duration': 60,
+            'acceptable_loss_pct': 0.0,
+        },
+    },
+    'rfc2544_latency': {
+        'description': 'RFC 2544 latency test at moderate rate',
+        'flow': {
+            'protocol': 'icmp',
+            'frame_size': 64,
+            'rate_pps': 100,
+            'duration': 30,
+        },
+        'test': {
+            'type': 'latency',
+            'frame_sizes': [64, 512, 1518],
+            'trial_duration': 30,
+        },
+    },
+    'tcp_session': {
+        'description': 'TCP SYN flood at 100 pps (for testing ACL/firewall)',
+        'flow': {
+            'protocol': 'tcp',
+            'dst_port': 80,
+            'frame_size': 64,
+            'rate_pps': 100,
+            'duration': 30,
+        },
+    },
+}
--- a/traffic-gen/presets/pycache/init.cpython-310.pyc
+++ b/traffic-gen/presets/pycache/init.cpython-310.pyc
--- a/traffic-gen/requirements.txt
+++ b/traffic-gen/requirements.txt
@ -0,0 +1,3 @@
+flask
+scapy
+psutil
--- a/traffic-gen/server.py
+++ b/traffic-gen/server.py
@ -0,0 +1,656 @@
+#!/usr/bin/env python3
+"""
+Traffic Generator API Server
+
+Scapy-based traffic generator with a Flask API. Runs in two modes
+controlled by TRAFFIC_GEN_MODE env var:
+  - sender (default): generates traffic, runs RFC 2544 tests
+  - responder: listens for test packets, echoes/timestamps, reports rx stats
+
+API endpoints:
+  GET  /healthz                - health + active flows/tests count
+  GET  /interfaces             - list available network interfaces
+  GET  /mode                   - current mode
+
+  Sender mode:
+    GET    /flows              - list all flows
+    POST   /flows              - create flow
+    GET    /flows/<id>         - get flow details + stats
+    PUT    /flows/<id>         - update flow (only if idle)
+    DELETE /flows/<id>         - delete flow
+    POST   /flows/<id>/start   - start sending
+    POST   /flows/<id>/stop    - stop sending
+    GET    /flows/<id>/stats   - real-time stats
+
+    GET    /tests              - list all tests
+    POST   /tests              - create RFC 2544 test
+    GET    /tests/<id>         - test details + results
+    POST   /tests/<id>/start   - start test
+    POST   /tests/<id>/stop    - abort test
+    GET    /tests/<id>/results - exportable results
+
+    GET    /presets             - list presets
+    POST   /presets/<name>     - create flow/test from preset
+
+    GET    /stats/history      - historical stats for all flows
+
+  Responder mode:
+    GET    /responder/stats    - receive statistics
+    POST   /responder/reset    - reset receive counters
+"""
+
+import sys
+import os
+import json
+import threading
+import time
+import logging
+import uuid
+
+import psutil
+from flask import Flask, request, jsonify
+
+# --- logging to stderr so stdout stays clean ---
+logging.basicConfig(
+    stream=sys.stderr,
+    level=logging.INFO,
+    format='%(asctime)s [TGEN] %(levelname)s %(message)s',
+)
+log = logging.getLogger(__name__)
+
+app = Flask(__name__)
+
+# ---------------------------------------------------------------------------
+# Global state (set in main())
+# ---------------------------------------------------------------------------
+MODE = 'sender'  # or 'responder'
+
+# Sender-mode globals
+_sender = None        # FlowSender instance
+_stats_collector = None  # StatsCollector instance
+_flows_meta = {}      # flow_id -> metadata dict (includes config)
+_flows_lock = threading.Lock()
+
+_tests = {}           # test_id -> RFC 2544 test instance
+_tests_lock = threading.Lock()
+
+# Responder-mode globals
+_responder = None     # Responder instance
+
+
+# ---------------------------------------------------------------------------
+# Helper
+# ---------------------------------------------------------------------------
+
+def _now_iso():
+    return time.strftime('%Y-%m-%dT%H:%M:%SZ', time.gmtime())
+
+
+def _flow_response(flow_id: str) -> dict:
+    """Build a serializable flow dict."""
+    with _flows_lock:
+        meta = _flows_meta.get(flow_id)
+        if meta is None:
+            return None
+    result = dict(meta)
+    if _sender:
+        result['is_running'] = _sender.is_running(flow_id)
+        result['stats'] = _sender.get_stats(flow_id)
+    return result
+
+
+# ---------------------------------------------------------------------------
+# Common endpoints
+# ---------------------------------------------------------------------------
+
+@app.route('/healthz')
+def health():
+    with _flows_lock:
+        active_flows = sum(
+            1 for fid in _flows_meta
+            if _sender and _sender.is_running(fid)
+        ) if MODE == 'sender' else 0
+    with _tests_lock:
+        active_tests = sum(
+            1 for t in _tests.values() if t.state == 'running'
+        ) if MODE == 'sender' else 0
+    resp = {
+        'status': 'ok',
+        'mode': MODE,
+        'active_flows': active_flows,
+        'active_tests': active_tests,
+    }
+    if MODE == 'responder' and _responder:
+        resp['responder_running'] = _responder.is_running()
+    return jsonify(resp)
+
+
+@app.route('/interfaces')
+def list_interfaces():
+    ifaces = []
+    addrs = psutil.net_if_addrs()
+    stats = psutil.net_if_stats()
+    for name, addr_list in addrs.items():
+        info = {'name': name, 'ip': None, 'mac': None, 'is_up': False}
+        for addr in addr_list:
+            if addr.family.name == 'AF_INET':
+                info['ip'] = addr.address
+            elif addr.family.name == 'AF_PACKET':
+                info['mac'] = addr.address
+        if name in stats:
+            info['is_up'] = stats[name].isup
+        ifaces.append(info)
+    return jsonify({'interfaces': ifaces})
+
+
+@app.route('/mode')
+def get_mode():
+    return jsonify({'mode': MODE})
+
+
+# ---------------------------------------------------------------------------
+# Sender-mode: Flow endpoints
+# ---------------------------------------------------------------------------
+
+@app.route('/flows', methods=['GET'])
+def list_flows():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _flows_lock:
+        flow_ids = list(_flows_meta.keys())
+    flows = [_flow_response(fid) for fid in flow_ids]
+    return jsonify({'count': len(flows), 'flows': flows})
+
+
+@app.route('/flows', methods=['POST'])
+def create_flow():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    data = request.get_json(force=True)
+
+    # Validate required fields
+    if 'dst_ip' not in data:
+        return jsonify({'error': 'dst_ip is required'}), 400
+    if 'protocol' not in data:
+        return jsonify({'error': 'protocol is required'}), 400
+    if data['protocol'].lower() not in ('udp', 'tcp', 'icmp'):
+        return jsonify({'error': 'protocol must be udp, tcp, or icmp'}), 400
+
+    flow_id = str(uuid.uuid4())
+    flow_config = {
+        'id': flow_id,
+        'name': data.get('name', f'flow-{flow_id[:8]}'),
+        'src_mac': data.get('src_mac', 'auto'),
+        'dst_mac': data.get('dst_mac'),
+        'src_ip': data.get('src_ip'),
+        'dst_ip': data['dst_ip'],
+        'protocol': data['protocol'].lower(),
+        'src_port': data.get('src_port'),
+        'dst_port': data.get('dst_port'),
+        'frame_size': int(data.get('frame_size', 512)),
+        'rate_pps': int(data.get('rate_pps', 1000)),
+        'duration': int(data.get('duration', 30)),
+        'dscp': int(data.get('dscp', 0)),
+        'vlan_id': data.get('vlan_id'),
+        'state': 'idle',
+        'responder_url': data.get('responder_url'),
+        'created_at': _now_iso(),
+    }
+
+    with _flows_lock:
+        _flows_meta[flow_id] = flow_config
+    _sender.add_flow(flow_id, dict(flow_config))
+
+    log.info('Created flow %s (%s -> %s, %s)',
+             flow_id[:8], flow_config.get('src_ip', 'auto'),
+             flow_config['dst_ip'], flow_config['protocol'])
+    return jsonify(_flow_response(flow_id)), 201
+
+
+@app.route('/flows/<flow_id>', methods=['GET'])
+def get_flow(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    result = _flow_response(flow_id)
+    if result is None:
+        return jsonify({'error': 'Flow not found'}), 404
+    return jsonify(result)
+
+
+@app.route('/flows/<flow_id>', methods=['PUT'])
+def update_flow(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    with _flows_lock:
+        meta = _flows_meta.get(flow_id)
+        if meta is None:
+            return jsonify({'error': 'Flow not found'}), 404
+        if meta.get('state') == 'running':
+            return jsonify({'error': 'Cannot update a running flow'}), 409
+
+    data = request.get_json(force=True)
+    updatable = [
+        'name', 'src_mac', 'dst_mac', 'src_ip', 'dst_ip', 'protocol',
+        'src_port', 'dst_port', 'frame_size', 'rate_pps', 'duration',
+        'dscp', 'vlan_id', 'responder_url',
+    ]
+
+    with _flows_lock:
+        for key in updatable:
+            if key in data:
+                _flows_meta[flow_id][key] = data[key]
+        _sender.update_flow(flow_id, dict(_flows_meta[flow_id]))
+
+    return jsonify(_flow_response(flow_id))
+
+
+@app.route('/flows/<flow_id>', methods=['DELETE'])
+def delete_flow(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    with _flows_lock:
+        if flow_id not in _flows_meta:
+            return jsonify({'error': 'Flow not found'}), 404
+
+    _sender.remove_flow(flow_id)
+    if _stats_collector:
+        _stats_collector.remove_flow(flow_id)
+
+    with _flows_lock:
+        _flows_meta.pop(flow_id, None)
+
+    log.info('Deleted flow %s', flow_id[:8])
+    return jsonify({'deleted': flow_id})
+
+
+@app.route('/flows/<flow_id>/start', methods=['POST'])
+def start_flow(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    with _flows_lock:
+        if flow_id not in _flows_meta:
+            return jsonify({'error': 'Flow not found'}), 404
+        _flows_meta[flow_id]['state'] = 'running'
+
+    try:
+        _sender.start(flow_id)
+    except KeyError:
+        return jsonify({'error': 'Flow not found in sender'}), 404
+
+    log.info('Started flow %s', flow_id[:8])
+    return jsonify(_flow_response(flow_id))
+
+
+@app.route('/flows/<flow_id>/stop', methods=['POST'])
+def stop_flow(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    with _flows_lock:
+        if flow_id not in _flows_meta:
+            return jsonify({'error': 'Flow not found'}), 404
+        _flows_meta[flow_id]['state'] = 'stopped'
+
+    _sender.stop(flow_id)
+
+    log.info('Stopped flow %s', flow_id[:8])
+    return jsonify(_flow_response(flow_id))
+
+
+@app.route('/flows/<flow_id>/stats', methods=['GET'])
+def flow_stats(flow_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    with _flows_lock:
+        if flow_id not in _flows_meta:
+            return jsonify({'error': 'Flow not found'}), 404
+
+    stats = _sender.get_stats(flow_id)
+    latest = _stats_collector.get_latest(flow_id) if _stats_collector else {}
+    return jsonify({'flow_id': flow_id, 'counters': stats, 'rates': latest})
+
+
+# ---------------------------------------------------------------------------
+# Sender-mode: Test endpoints
+# ---------------------------------------------------------------------------
+
+@app.route('/tests', methods=['GET'])
+def list_tests():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _tests_lock:
+        tests = [t.get_info() for t in _tests.values()]
+    return jsonify({'count': len(tests), 'tests': tests})
+
+
+@app.route('/tests', methods=['POST'])
+def create_test():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    from engine.rfc2544 import create_test as _create_test
+
+    data = request.get_json(force=True)
+    flow_id = data.get('flow_id')
+    test_type = data.get('type')
+
+    if not flow_id or not test_type:
+        return jsonify({'error': 'flow_id and type are required'}), 400
+
+    with _flows_lock:
+        flow_meta = _flows_meta.get(flow_id)
+        if flow_meta is None:
+            return jsonify({'error': 'Flow not found'}), 404
+
+    test_id = str(uuid.uuid4())
+    kwargs = {
+        'frame_sizes': data.get('frame_sizes', [64, 512, 1518]),
+        'trial_duration': float(data.get('trial_duration', 60)),
+        'max_rate_pps': int(data.get('max_rate_pps', flow_meta.get('rate_pps', 10000))),
+        'acceptable_loss_pct': float(data.get('acceptable_loss_pct', 0.0)),
+    }
+
+    try:
+        test = _create_test(test_id, test_type, dict(flow_meta), **kwargs)
+    except ValueError as e:
+        return jsonify({'error': str(e)}), 400
+
+    with _tests_lock:
+        _tests[test_id] = test
+
+    log.info('Created test %s (type=%s, flow=%s)', test_id[:8], test_type, flow_id[:8])
+    return jsonify(test.get_info()), 201
+
+
+@app.route('/tests/<test_id>', methods=['GET'])
+def get_test(test_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _tests_lock:
+        test = _tests.get(test_id)
+    if test is None:
+        return jsonify({'error': 'Test not found'}), 404
+    return jsonify(test.get_info())
+
+
+@app.route('/tests/<test_id>/start', methods=['POST'])
+def start_test(test_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _tests_lock:
+        test = _tests.get(test_id)
+    if test is None:
+        return jsonify({'error': 'Test not found'}), 404
+    test.start()
+    log.info('Started test %s', test_id[:8])
+    return jsonify(test.get_info())
+
+
+@app.route('/tests/<test_id>/stop', methods=['POST'])
+def stop_test(test_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _tests_lock:
+        test = _tests.get(test_id)
+    if test is None:
+        return jsonify({'error': 'Test not found'}), 404
+    test.stop()
+    log.info('Stopped test %s', test_id[:8])
+    return jsonify(test.get_info())
+
+
+@app.route('/tests/<test_id>/results', methods=['GET'])
+def test_results(test_id):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+    with _tests_lock:
+        test = _tests.get(test_id)
+    if test is None:
+        return jsonify({'error': 'Test not found'}), 404
+    return jsonify({
+        'test_id': test_id,
+        'type': test.__class__.__name__,
+        'state': test.state,
+        'results': test.results,
+        'started_at': test.started_at,
+        'completed_at': test.completed_at,
+    })
+
+
+# ---------------------------------------------------------------------------
+# Sender-mode: Presets
+# ---------------------------------------------------------------------------
+
+@app.route('/presets', methods=['GET'])
+def list_presets():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    from presets import PRESETS
+    out = {}
+    for name, preset in PRESETS.items():
+        out[name] = {
+            'description': preset.get('description', ''),
+            'has_test': 'test' in preset,
+        }
+    return jsonify({'presets': out})
+
+
+@app.route('/presets/<name>', methods=['POST'])
+def load_preset(name):
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    from presets import PRESETS
+    from engine.rfc2544 import create_test as _create_test
+
+    if name not in PRESETS:
+        return jsonify({'error': f'Unknown preset: {name}. Available: {list(PRESETS)}'}), 404
+
+    preset = PRESETS[name]
+    data = request.get_json(silent=True) or {}
+
+    # The preset flow needs a dst_ip; caller can override
+    flow_data = dict(preset['flow'])
+    flow_data['dst_ip'] = data.get('dst_ip', flow_data.get('dst_ip', '10.0.0.1'))
+    if 'src_ip' in data:
+        flow_data['src_ip'] = data['src_ip']
+    if 'responder_url' in data:
+        flow_data['responder_url'] = data['responder_url']
+
+    # Validate protocol present
+    if 'protocol' not in flow_data:
+        flow_data['protocol'] = 'udp'
+
+    flow_id = str(uuid.uuid4())
+    flow_config = {
+        'id': flow_id,
+        'name': data.get('name', f'{name}-{flow_id[:8]}'),
+        'src_mac': flow_data.get('src_mac', 'auto'),
+        'dst_mac': flow_data.get('dst_mac'),
+        'src_ip': flow_data.get('src_ip'),
+        'dst_ip': flow_data['dst_ip'],
+        'protocol': flow_data['protocol'],
+        'src_port': flow_data.get('src_port'),
+        'dst_port': flow_data.get('dst_port'),
+        'frame_size': int(flow_data.get('frame_size', 512)),
+        'rate_pps': int(flow_data.get('rate_pps', 1000)),
+        'duration': int(flow_data.get('duration', 30)),
+        'dscp': int(flow_data.get('dscp', 0)),
+        'vlan_id': flow_data.get('vlan_id'),
+        'state': 'idle',
+        'responder_url': flow_data.get('responder_url'),
+        'created_at': _now_iso(),
+    }
+
+    with _flows_lock:
+        _flows_meta[flow_id] = flow_config
+    _sender.add_flow(flow_id, dict(flow_config))
+
+    result = {'flow': _flow_response(flow_id)}
+
+    # Optionally create a test
+    if 'test' in preset:
+        test_cfg = preset['test']
+        test_id = str(uuid.uuid4())
+        kwargs = {
+            'frame_sizes': test_cfg.get('frame_sizes', [64, 512, 1518]),
+            'trial_duration': float(test_cfg.get('trial_duration', 60)),
+            'max_rate_pps': int(test_cfg.get('max_rate_pps', flow_config.get('rate_pps', 10000))),
+            'acceptable_loss_pct': float(test_cfg.get('acceptable_loss_pct', 0.0)),
+        }
+        test = _create_test(test_id, test_cfg['type'], dict(flow_config), **kwargs)
+        with _tests_lock:
+            _tests[test_id] = test
+        result['test'] = test.get_info()
+
+    log.info('Loaded preset %s -> flow %s', name, flow_id[:8])
+    return jsonify(result), 201
+
+
+# ---------------------------------------------------------------------------
+# Sender-mode: Stats history
+# ---------------------------------------------------------------------------
+
+@app.route('/stats/history', methods=['GET'])
+def stats_history():
+    if MODE != 'sender':
+        return jsonify({'error': 'Not in sender mode'}), 400
+
+    count = int(request.args.get('count', 60))
+    with _flows_lock:
+        flow_ids = list(_flows_meta.keys())
+
+    history = {}
+    for fid in flow_ids:
+        history[fid] = _stats_collector.get_history(fid, count)
+    return jsonify({'history': history})
+
+
+# ---------------------------------------------------------------------------
+# Responder-mode endpoints
+# ---------------------------------------------------------------------------
+
+@app.route('/responder/stats', methods=['GET'])
+def responder_stats():
+    if MODE != 'responder' or _responder is None:
+        return jsonify({'error': 'Not in responder mode'}), 400
+    return jsonify(_responder.get_stats())
+
+
+@app.route('/responder/reset', methods=['POST'])
+def responder_reset():
+    if MODE != 'responder' or _responder is None:
+        return jsonify({'error': 'Not in responder mode'}), 400
+    _responder.reset_stats()
+    return jsonify({'status': 'reset'})
+
+
+# ---------------------------------------------------------------------------
+# Stats collection loop (sender mode)
+# ---------------------------------------------------------------------------
+
+def _stats_loop(stop_event: threading.Event):
+    """Runs every 1s, recording per-flow stats into the StatsCollector."""
+    while not stop_event.is_set():
+        try:
+            all_stats = _sender.get_all_stats()
+            for flow_id, s in all_stats.items():
+                latency = None
+                samples = s.get('latency_samples', [])
+                if samples:
+                    latency = {
+                        'min_ms': round(min(samples), 3),
+                        'max_ms': round(max(samples), 3),
+                        'avg_ms': round(sum(samples) / len(samples), 3),
+                    }
+                _stats_collector.record(
+                    flow_id,
+                    tx_packets=s.get('tx_packets', 0),
+                    tx_bytes=s.get('tx_bytes', 0),
+                    rx_packets=s.get('rx_packets', 0),
+                    rx_bytes=s.get('rx_bytes', 0),
+                    latency=latency,
+                )
+        except Exception as e:
+            log.debug('Stats loop error: %s', e)
+        stop_event.wait(1.0)
+
+
+# ---------------------------------------------------------------------------
+# Main
+# ---------------------------------------------------------------------------
+
+def main():
+    global MODE, _sender, _stats_collector, _responder
+
+    MODE = os.environ.get('TRAFFIC_GEN_MODE', 'sender').lower()
+    api_port = int(os.environ.get('TRAFFIC_GEN_PORT', 5051))
+    listen_iface = os.environ.get('TRAFFIC_GEN_INTERFACE', None)
+    responder_sub_mode = os.environ.get('TRAFFIC_GEN_RESPONDER_MODE', 'log')  # echo or log
+
+    log.info('Traffic Generator starting in %s mode', MODE)
+
+    if MODE == 'sender':
+        from engine.sender import FlowSender
+        from engine.stats import StatsCollector
+
+        _sender = FlowSender()
+        _stats_collector = StatsCollector()
+
+        # Start stats collection loop
+        stats_stop = threading.Event()
+        stats_thread = threading.Thread(
+            target=_stats_loop, args=(stats_stop,), daemon=True, name='stats-loop'
+        )
+        stats_thread.start()
+
+        # Start Flask in background thread
+        flask_thread = threading.Thread(
+            target=lambda: app.run(host='0.0.0.0', port=api_port, threaded=True),
+            daemon=True,
+        )
+        flask_thread.start()
+        log.info('Flask API listening on port %d', api_port)
+
+        # Main thread: keep alive
+        try:
+            while True:
+                time.sleep(1)
+        except KeyboardInterrupt:
+            log.info('Shutting down sender...')
+            stats_stop.set()
+
+    elif MODE == 'responder':
+        from engine.responder import Responder
+
+        _responder = Responder(mode=responder_sub_mode)
+        _responder.start(interface=listen_iface)
+
+        # Start Flask in background thread
+        flask_thread = threading.Thread(
+            target=lambda: app.run(host='0.0.0.0', port=api_port, threaded=True),
+            daemon=True,
+        )
+        flask_thread.start()
+        log.info('Flask API listening on port %d', api_port)
+
+        # Main thread: keep alive
+        try:
+            while True:
+                time.sleep(1)
+        except KeyboardInterrupt:
+            log.info('Shutting down responder...')
+            _responder.stop()
+
+    else:
+        log.error('Unknown mode: %s. Use "sender" or "responder".', MODE)
+        sys.exit(1)
+
+
+if __name__ == '__main__':
+    main()