Red5 Documentation

Release 5.0.0

  • Switched default HLS behavior to use Disk Writes to alleviate some memory utilization issues
  • Fixed – HLS can’t play live stream with Disk Writes set to true
  • Fixed – “Zombie” streams are not being reaped
  • Fixed – RTMP Publisher memory leak increases over time and ultimately crashes the server
  • Fixed – VP8 publish failure in WebRTC (affects Opera and some Android Chrome browsers)
  • Fixed – VP8 transcoding not working
  • Fixed – RTC video only to RTSP subscriber issues
  • Fixed – Internet Explorer Flash broadcast to Windows Chrome quality issues
  • Fixed – Underscores in stream names break HLS
  • Fixed – appDisconnect Event is called twice for RTC publisher
  • Fixed – Opera RTC publish and subscribe issues
  • Fixed – Autoscaling – Azure controller unable to create instances for all VM types
  • Fixed – Front End to handle UTF-8
  • Fixed – Newer versions of OBS Failed to Publish

Autoscaling API v3.0 – New Features:

  • Streaming failover for multi-origin broadcasts
  • New Cauldron Transcoder Node to generate multiple bitrates with option to transcode VP8
  • New Relay Node option to allow for even larger scale and connectivity between regions
  • Multi-bitrate stream provisioning
  • WebRTC adaptive bitrate subscribing (ABR Support for WebRTC clients only, mobile SDK plus other protocol support coming soon)
  • VP8 Transcoding on Cauldron Node to alleviate load on edge servers

Release 5.0.0 Server Performance Metrics

Tests were run against an AWS c5.large instance (2 CPUs with 4GB memory, 2GB allocated to java_heap). We used our RTMP Bee, RTSP Bee, and RTC Bee clients to do load testing.

Publishing a 256kbps stream via RTMP, we were able to achieve the following while still maintaining quality of stream:

  • 400 WebRTC subscribers
  • 1,800 RTSP (mobile) subscribers
  • 1,000 RTMP subscribers

The same server type can support approximately 75-80 480p RTMP publishers