TMF628 – Performance Management API
🎯 Learning Objective: Understand TMF628 - the TM Forum Open API for Performance Management. Learn how OSS systems expose performance measurements and KPI data to analytics platforms, customer portals, and other operational systems for SLA monitoring, reporting, and capacity planning.
Key Performance Fields Explained
| Field | Description (Simplified) | Example |
|---|---|---|
| id | The unique "serial number" for this performance measurement record. | "perf-12345" |
| serviceId / resourceId | Which service or resource is being measured? | "vpn-123" or "router-mum-01" |
| metricType | What is being measured? (latency, throughput, packet loss, availability) | "latency" |
| value | The actual measurement number. | 25.5 |
| unit | The unit of measurement (ms, Mbps, %, seconds). | "ms" |
| timestamp | When the measurement was taken (UTC time). | "2026-05-15T10:00:00Z" |
| granularity | How often the measurement is collected (5min, 15min, hourly). | "5min" |
| slaCompliance | Percentage of time SLA was met (if applicable). | 99.99 |
Performance measurements are numeric values with associated units. Unlike alarms (which have severity levels like "critical"), performance metrics are continuous - they measure how good or bad the service is, not just if it's broken.
TMF628 Performance Measurement Object (JSON)
{
"id": "perf-67890",
"serviceId": "vpn-enterprise-001",
"resourceId": "router-mum-pe-01",
"metricType": "latency",
"value": 25.5,
"unit": "ms",
"timestamp": "2026-05-15T10:00:00Z",
"granularity": "5min",
"slaCompliance": 99.99
}
This example shows a latency measurement of 25.5ms for a specific enterprise VPN. The measurement was taken every 5 minutes, and the service is 99.99% compliant with its SLA (which likely requires latency under a certain threshold). The OSS platform may compare these measurements against thresholds configured separately through TMF649 or internal assurance systems.
Simple Analogy - The Health Dashboard
TMF642 (Fault/Alarm API)
Like a "Check Engine Light" in a car. Tells you something is broken right now. "Engine failure - pull over immediately!"
TMF628 (Performance API)
Like a "Health Dashboard" in a car. Shows speed, fuel efficiency, engine temperature over time. "Your fuel efficiency has dropped 15% this month."
TMF642 tells you what broke (alarms). TMF628 tells you how well the network is performing (measurements, KPIs, trends).
What is TMF628?
TMF628 is the TM Forum Open API for Performance Management. It allows OSS systems to manage performance measurement collection, production, retrieval, and notification workflows for other operational systems, analytics platforms, and customer-facing applications.
What It Exposes
- Latency (ms)
- Throughput (Mbps/Gbps)
- Packet loss (%)
- Availability (%)
- Jitter (ms)
- CPU / memory utilization (%)
Key Capabilities
- Create measurement collection jobs
- Create measurement production jobs
- Retrieve performance measurements
- Trigger ad hoc data collection
- Subscribe to performance event notifications
Who Uses It
- BSS for SLA monitoring
- Customer portals for near real-time or recent performance metrics
- Analytics platforms for trend analysis
- Orchestration for closed-loop automation
TMF628 does not directly collect performance data from routers, switches, or 5G devices. Network measurements are typically gathered through protocols and systems such as SNMP polling, streaming telemetry (gNMI), vendor EMS/NMS platforms, or cloud-native observability pipelines. TMF628 standardizes how OSS platforms expose and manage those measurements.
Key TMF628 Operations
TMF628 focuses on measurement collection and production jobs, not threshold configuration. Thresholding is handled by TMF649 (see section below).
| Operation | HTTP Method | Description |
|---|---|---|
| GET /measurementCollectionJob | GET | Retrieve configured measurement collection jobs |
| POST /measurementCollectionJob | POST | Create a new measurement collection job |
| GET /measurementProductionJob | GET | Retrieve measurement production jobs |
| POST /measurementProductionJob | POST | Create a measurement production job |
| GET /performanceMeasurement | GET | Retrieve collected performance measurements |
| POST /adHocCollection | POST | Trigger one-time performance data collection |
| POST /hub | POST | Subscribe to performance-related event notifications |
A measurement collection job defines how OSS collects performance data from network devices (interval, metrics, targets). A measurement production job defines how OSS produces/publishes aggregated measurements to downstream systems.
Common Performance Metrics Exposed by TMF628
| Metric | Description | Typical Values | SLA Impact |
|---|---|---|---|
| Latency | Packet travel time (round trip) | 10-50 ms (good), >100 ms (poor) | Real-time services (VoLTE, gaming) |
| Jitter | Variation in packet delay | < 10 ms (good), >30 ms (poor) | Voice and video quality |
| Packet Loss | Percentage of packets dropped | < 0.1% (good), >1% (poor) | All services degrade |
| Throughput | Data transfer rate | 100 Mbps, 1 Gbps, 10 Gbps | Bandwidth-dependent services |
| Availability | Uptime percentage | 99.999% (5 min/year downtime) | All services critical |
| PRB Utilization | 5G radio resource usage | < 70% (normal), >85% (congested) | Mobile broadband experience |
TMF649 - Performance Thresholding API (Related)
While TMF628 handles measurement collection and exposure, TMF649 is the TM Forum API for performance threshold management.
TMF628 (This Article)
- Performance measurement jobs
- Collecting and exposing KPIs
- Ad hoc data collection
- Performance event notifications
TMF649 (Thresholding)
- Create performance thresholds
- Manage threshold rules
- Raise threshold crossing alerts (TCAs)
- Monitor breaches
OSS performance systems collect measurements, while TMF628 standardizes how those measurements and jobs are exposed. TMF649 defines thresholds on those measurements and raises alerts when thresholds are crossed. In practice, OSS platforms often implement both.
TMF628 vs Other TMF APIs
TMF639
Resource Inventory
"What resources exist?"
TMF638
Service Inventory
"What services are active?"
TMF628
Performance Management
"How well are they performing?"
TMF649
Performance Thresholding
"When should we alert?"
TMF642
Alarm Management
"What broke?"
TMF639 + TMF638 = WHAT exists. TMF628 = HOW WELL it's performing. TMF649 = WHEN to alert. TMF642 = WHAT broke.
Connection to BSS - From Network to Revenue
SLA Dashboards
BSS and customer portals can use TMF628 to retrieve near real-time and historical SLA-related performance measurements
SLA Credits
Performance breach (via TMF649) triggers automatic credit calculation in BSS
Capacity Planning
Performance trends help BSS plan upselling opportunities
Proactive Alerts
Performance degradation (via TMF649 thresholds) triggers proactive customer notifications
With TMF628 + TMF649: Automated performance monitoring → SLA breach auto-detected → BSS auto-credits customer → retention improves
Key Terms You Must Know
TM Forum Open API for Performance Management (measurements, collection jobs)
TM Forum Open API for Performance Thresholding (thresholds, TCAs)
Defines how OSS collects performance data from network devices
Defines how OSS produces/publishes aggregated measurements
A single KPI value at a specific time (e.g., latency = 25ms)
One-time, on-demand performance data collection
Contractual performance commitments (e.g., latency < 30ms)
Alert when a performance metric exceeds a threshold (handled by TMF649)
Frequency of performance data collection (5 min, 15 min, hourly)
Packet travel time from source to destination (ms)
Percentage of packets dropped during transmission
Percentage of radio resources being consumed in a 4G/5G cell
Common Questions
Q1. What is TMF628?
TMF628 is the TM Forum Open API for Performance Management. It exposes network performance measurements (KPIs) from OSS to BSS, customer portals, and analytics platforms.
Q2. What are the key fields in a TMF628 performance measurement?
Key fields include id, serviceId/resourceId, metricType, value, unit, timestamp, granularity, and slaCompliance. These tell you what was measured, the numeric value, when it was measured, and how it relates to SLAs.
Q3. What is the difference between TMF628 and TMF649?
TMF628 handles performance measurement collection and exposure (jobs, measurements). TMF649 handles threshold definition and threshold crossing alerts (TCAs). They work together: TMF628 collects, TMF649 alerts on thresholds.
Q4. How is TMF628 different from TMF642?
TMF628 is for performance metrics (how well the network is performing). TMF642 is for fault/alarm management (what broke). Performance degradation often precedes faults.
Q5. What are the key operations in TMF628?
GET/POST measurementCollectionJob, GET/POST measurementProductionJob, GET performanceMeasurement, POST adHocCollection, and POST hub for event notifications.
Q6. What is the difference between a measurement collection job and a measurement production job?
Collection job defines how OSS collects data from network devices (interval, metrics, targets). Production job defines how OSS produces/publishes aggregated measurements to downstream systems.
Q7. How does TMF628 support SLA management?
TMF628 exposes real-time and historical KPIs. TMF649 monitors these KPIs against SLA thresholds. When breached, BSS can auto-credit customers or trigger alerts.
Q8. What is the business value of TMF628?
Enables automated SLA monitoring, proactive customer notifications, auto-crediting for breaches, capacity planning, and upselling based on usage trends.
📌 Key Takeaways:
- TMF628 = Performance Management API - exposes network measurements to BSS and analytics
- Key fields: id, serviceId/resourceId, metricType, value, unit, timestamp, granularity, slaCompliance
- TMF649 = Performance Thresholding API - defines thresholds and raises TCAs (separate API)
- TMF628 vs TMF642: TMF628 = "How well is it performing?" (metrics). TMF642 = "What broke?" (alarms)
- Key operations: measurementCollectionJob, measurementProductionJob, performanceMeasurement, adHocCollection, hub
- Common KPIs: Latency, throughput, packet loss, availability, jitter, PRB utilization
- Integration: Works with TMF639 (resource inventory) and TMF638 (service inventory) for complete visibility
- Business Value: Automated SLA monitoring, proactive customer alerts, revenue assurance, capacity planning