Avassa Proven Scalability: Managing 10 000 + Edge Sites

Validated performance for large-scale edge environments

Edge computing isn’t cloud computing on a smaller scale, it’s a fundamentally different challenge. Each site runs on constrained resources, often disconnected, and there may be tens of thousands of them.

That’s why scalability and autonomy are built into the Avassa Edge Platform from day one.

Scale test overview

We regularly perform large-scale tests to validate Control Tower performance in realistic environments.

Our standard setup is a three-node cluster (4 vCPU / 8 GiB RAM per node) managing simulated edge sites that fully implement the Avassa Edge Enforcer protocol.

Tested operations include:

• Creating and deleting sites
• Secure “call home” connections from edge sites
• Application deployments and removals
• Distributed telemetry queries via the built-in pub/sub bus Volga
• Client API responsiveness (used by CLI and UI)

Key results

These measurements provide a clear snapshot of how the Avassa platform performs at scale. Below are the core results from a test setup with 10,000 edge sites and a three-node Control Tower.

• 10 000 sites managed efficiently by a 3-node Control Tower
• Application deployment: ~18 minutes (≈ 0.1 s per site)
• Secure call home: ~40 minutes (≈ 0.2 s per site)
• Client API summaries: < 20 seconds for 10 000 sites
• Memory footprint: < 2 GiB total across the cluster
• Linear scalability across all measurements
• Responsive Web UI even with thousands of connected sites

The graph below shows the time it takes to create 100, 1000, 5000, and 10,000 sites.

Corresponding distribution of the call home process is shown below:

Finally, we illustrate deploying an application across all sites:

We also performed extended tests at 50,000 sites, showing continued linear behavior and predictable resource use.

Architecture designed for scale

Avassa’s scalability is the result of deliberate architectural choices:

• Lightweight central orchestration; heavy work pushed to the edge
• Secure, multiplexed QUIC transport for efficient communication
• Built-in pub/sub bus (Volga) for distributed messaging and queries
• Fully autonomous sites that continue operating when offline

Summary

These results confirm that Avassa Control Tower offers predictable performance and minimal central footprint for truly large-scale edge deployments, the kind found in retail, manufacturing, transportation, and other distributed industries.

• Read the full scalability report here.
• See the video on the UI responsiveness with thousands of sites: