04-13-2016 10:41 AM - edited 04-13-2016 11:05 AM
It looks like all the lag, buffering, stuttering, and low resolution happens when they move your stream to a new server in the cluster.
Here the TCP session gets reset, followed by audio cutouts, buffering, etc.
2016-04-13 12:45:49.333829000 192.168.1.92 188.8.131.52 TCP 54 742083 62803→80 [RST, ACK] Seq=27682 Ack=29107321 Win=0 Len=0
Then, a new server IP gets used shortly after:
2016-04-13 12:45:56.575340000 192.168.1.92 184.108.40.206 HTTP 248 742228 GET /LIVE/1009/hls/ae/FNews_HD/20160413T163805-2400-5073763.ts HTTP/1.1
Frankly, I'm surprised the stream is using TCP instead of UDP. UDP is more fault tolerant for things like video and audio.
2016-04-13 12:58:31.671846000 220.127.116.11 192.168.1.92 TCP 1514 1000448 [TCP segment of a reassembled PDU]
Furthermore, Adobe Primetime now supports HTML5 video, which uses UDP by default. Flash is outdated, unstable and the whole world of browsers is ditching it in favor of HTML5.
04-13-2016 05:52 PM
Watching now, during primetime proves that this update had no impact whatsover. Constant audio skipping, buffering, etc.
01-19-2017 11:20 AM - edited 01-19-2017 11:21 AM
Still broken! Almost a year has passed. When does Time Warner plan on admitting that their CDN platform that serves content to web browsers is undersized, overutilized, and is running on inferior and broken technology?
Here's my experience for the past 48 hours. Not to mention constant buffering, skipping, stopping, repeating.
01-28-2017 07:46 PM