Performance Issues in AWS

[double]

Hi. We did load testing of WCS servers on AWS Marketplace (c5.4xlarge).
A publish stream VP8 720p with bitrate 2.5Mbps.
We tested using WebRTC pulling 400 streams out.
At the time, the screen freezes for seconds every few seconds. no degraded streams.
Is there a way to improve this?
we were hoping to the get 1000 streams per server (c5.4xlarge).

Configuration and Tunning done :

What kind of server do you need to run a thousand WebRTC streams? | Flashphoner Streaming & Calls for Web

flashphoner.com

Before moving to production - Web Call Server 5.2 - EN - Flashphoner Documentation

docs.flashphoner.com

1. WebRTC traffic encryption hardware acceleration
2. Stream distribution optimization
3. Traffic encryption in a separate thread for each client session
4. Heap Memory, ZGC and Media ports range is set

 

[Max]

Good day.
Please update WCS to the latest build from this page as described here. This should help. Tha latest builds also includes some fixes affecting VP8 paublishing and recording.

 

[double]

In our tests after update 5.2.919, the issue is reproduced when using WebRTC pulling 400 streams out.
you can see the server statistics in the attachment.

 

[Max]

Please do the following server tuning:
1. Enable WebRTC traffic encryption hardware acceleration
2. Enable stream distribution optimization

 

[double]

already server is tuned.
on play side, increase the Nack Count and Packets Lost when 400 connections
this test
1. two WCS servers (c5.4xlarge)
2. streaming setup described in Media Devices (Demo Web)

# wcs-core.properties
-server
-XX:+UnlockDiagnosticVMOptions
-XX:+UseAES
-XX:+UseAESIntrinsics

# flashphoner.properties
webrtc_aes_crypto_provider=JCE

streaming_distributor_subgroup_enabled=true
streaming_distributor_subgroup_size=50
streaming_distributor_audio_subgroup_size=500
streaming_distributor_subgroup_queue_size=300
streaming_distributor_subgroup_queue_max_waiting_time=5000
streaming_distributor_audio_subgroup_queue_size=300
streaming_distributor_audio_subgroup_queue_max_waiting_time=5000

rtp_paced_sender=true
rtp_paced_sender_initial_rate=200000
rtp_paced_sender_increase_interval=50
rtp_paced_sender_k_up=0.9

 

[Max]

on play side, increase the Nack Count and Packets Lost when 400 connections

Seems like the channel bandwidth from server to player is not enough to play the stream properly. Try to migrate to closer datacenter or to reduce resolution/bitrate.