We’re proud to announce that we’ve completed a series of high-volume stress tests of Unified Management Platform (UMP), validating its ability to support carrier-grade CPE lifecycle management at the scale required by Tier-1 telecom operators.
The tests confirmed sustained performance of up to 10,000 concurrent device management sessions per second, equivalent to 36 million sessions per hour, across a mixed estate of TR-069 (CWMP) and TR-369 (USP) devices. The results demonstrate that UMP is ready to operate in large, heterogeneous broadband environments where predictable behavior under load is critical.
The benchmarks were conducted in a cloud-based AWS environment and focused on realistic operational scenarios, including extended session durations, continuous telemetry forwarding, and concurrent CWMP and USP traffic. Throughout the tests, the platform maintained consistent and stable performance while providing full observability across device interactions and backend components.
Built for real-world operator scale
These stress tests were designed to answer a fundamental question for large operators: Can the platform scale safely as device operations grow from thousands to tens of millions?
The results confirm that UMP is engineered for carrier-grade scale, where automation, lifecycle control, and governance must work reliably across vast and diverse device fleets.
The test architecture reflected production-grade deployments and included:
-
A distributed UMP cluster capable of handling high volumes of device sessions
-
Dedicated USP traffic management via MQTT-based Artemis brokers
-
Kafka-based telemetry streaming pipelines
-
Redis and MongoDB data layers optimized for high-throughput operations
This architecture enables hybrid TR-069 and TR-369 operation, allowing operators to modernize their device estates incrementally while preserving service continuity.
From device management to operational control
Beyond raw throughput, the results validate AVSystem’s broader approach to device management. UMP treats CPE fleets as governed systems rather than individual endpoints, enabling operators to maintain an accurate operational view of device state, often described as a digital twin model.
This foundation supports:
- Controlled, campaign-based firmware upgrades
- Safe automation at scale
- Consistent configuration enforcement
- Data-driven support and service assurance workflows
While the stress tests did not include active-active high-availability configurations, they established a solid performance baseline for large-scale deployments and will serve as a reference point for further resilience engineering and production optimization.
The full stress-test report details, including the test methodology, architecture, and performance results, and is available on request.