Slyber Inspector Overview

Updated:
Author: Erkin Khidirov

Slyber Inspector is a unified interface for functional testing, media stream inspection, and load testing of the Slyber server infrastructure. The application is delivered as a standalone executable with a built-in web interface, allowing engineers to verify the operation of communication nodes directly from the browser without writing test client scripts.

Key Tasks of the Tool

  • Network connectivity verification: rapid diagnostics of server availability over UDP, TCP, and TLS protocols before launching the product into production.
  • Authorization mechanism validation: testing both standalone temporary token generation using the HMAC SHA1 algorithm and dynamic key issuance and revocation in Redis storage.
  • Visual and audio relay control: establishing real two-way WebRTC video calls with forced routing through relay ports to verify video and audio quality.
  • Load simulation: simultaneously launching dozens of independent connection pairs to evaluate channel throughput and port allocation stability.
  • Workspace persistence: automatically saving all entered settings to the browser's local storage and retaining uploaded video clips in the IndexedDB database across page reloads.

Modular Platform Structure

On the main screen of Slyber Inspector, you can switch between three specialized ecosystem modules:

  • Slyber TURN Server: an active testbed for verifying relay candidate allocation, testing media stream throughput, switching routes on the fly, and managing sessions in Redis.
  • Slyber SFU Server: a module for testing selective routing of multi-user video conferences, switching Simulcast spatial layers, and evaluating the throughput of cascaded rooms.
  • Slyber Signaling Server: a module for verifying SDP message exchange speed, testing event delivery latency, and inspecting room states under high concurrent load.

Testbed Architecture

  1. The web interface communicates with the internal Slyber Inspector backend to verify the target server and automatically prepare credentials.
  2. When creating a test pair, the built-in signaling node connects two independent browser RTCPeerConnection objects within a single tab over WebSocket.
  3. Both peers capture different video streams and audio tracks from the local pool and transmit them to each other strictly through the external server under test.
  4. The telemetry collector polls WebRTC statistics every second and displays the actual bitrate, volume of transmitted data, and network latency for each direction.