We are running WSE on our AWS EC2 instance, using it to capture / record live RTMP. In our case we are capturing two live streams of the same live content (originating from two separate encoders) with unique stream keys. We have been using this workflow successfully for multiple events for almost two years.
To monitor the inbound streams at the entry point to WSE, we copy the “test stream” URL from the incoming live streams control panel and paste it into a standard JW Player test page. At our last live event, both streams had buffering issues at the WSE entry point, i.e. the inbound test stream link showed classic ‘spinning wheel’ buffering delays in the JW Player. We’ve never seen any buffering previously, capturing streams of exactly the same spec’s.
The two streams also buffered at different rates/amounts. Neither recording actually lost any data but by about twenty minutes after the start of the recordings, one recorded stream was delayed by about six minutes more than it’s (otherwise synchronous) partner. Other streams of the same live content from the same encoding facility were streaming smoothly on other platforms, without any buffering issues. We eventually stopped the recording and the streaming transmission and after restarting, and by the end of the 90 minute program, the buffering had stopped. And FWIW - We had tested both streams for more than an hour the evening before, from exactly the same signal path and had zero buffering issues.
Has anyone else experienced this issue? We would appreciate any guidance on settings or best practices to avoid this serious issue on future events. Thanks!
Hello Bill,
It’s unfortunate to hear that you experienced buffering issues at the entry point of the Wowza Streaming Engine (WSE). While it’s difficult to provide a specific solution without further investigation, I can offer some general guidelines and best practices that might help address the issue:
-
Network Considerations: Ensure that your network infrastructure has sufficient bandwidth and is properly configured to handle the incoming RTMP streams. Verify that there are no network bottlenecks, congestion, or packet loss issues that could cause buffering delays. You may want to work closely with your network team to analyze network performance and address any potential issues.
-
Server Resources: Check the resource utilization of your AWS EC2 instance running WSE during the event. Ensure that the instance has enough CPU, memory, and disk I/O capacity to handle the incoming streams and perform the necessary encoding and recording tasks.
-
Wowza Streaming Engine Configuration: Review your WSE configuration settings to ensure they are optimized for your streaming requirements.
-
Encoding Considerations: Confirm that your encoders are properly configured and delivering the RTMP streams with optimal settings for the target playback devices and network conditions. Double-check the encoding parameters such as codec, bitrate, and keyframe interval to ensure they are appropriate for your streaming requirements.
-
Load Testing and Pre-Event Testing: Perform load testing on your streaming setup to simulate high-traffic conditions and identify any bottlenecks or issues before the actual live event. Additionally, conduct thorough pre-event testing, including streaming for extended periods, to ensure smooth and uninterrupted playback without buffering.
-
Wowza Support: Reach out to Wowza Support via ticket with detailed information about the issue you encountered. Wowza Support can provide more specific guidance based on your setup, logs, and any specific error messages you may have encountered.
Thank you.
Regards, Umair