RE: [squid-users] How to optimize squid config for live flash video streaming?

From: Andrew Krupiczka <andrewk_at_vbrick.com>
Date: Mon, 23 Jul 2012 13:28:27 +0000

Hello Everyone,

Just one comment and question.

1) we're using the squid to optimize HLS (http based) video streaming exactly via caching, so you can deliver more efficiently a cached video segments instead of directly communicating with an origin server(s).
2) does anyone know, if/when collapsed forwarding feature is going to be incorporated into Squid 3.x?

Best regards,

Andrew

-----Original Message-----
From: Amos Jeffries [mailto:squid3_at_treenet.co.nz]
Sent: Friday, July 20, 2012 10:54 PM
To: squid-users_at_squid-cache.org
Subject: Re: [squid-users] How to optimize squid config for live flash video streaming?

On 21/07/2012 5:58 a.m., J DJ wrote:
> My squid proxy is not caching (using the "cache deny all" in
> squid.conf), and I was wondering if anyone had any suggestions or
> resources on how to optimize it for live flash video streaming?

There is no optimization for live streaming without caching.
The Squid collapsed-forwarding feature for splitting single response to multiple clients relies on responses being cacheable (even if not
stored) so that multiple clients are *allowed* to receive them. Live streams are usually marked as non-cacheable, and Flash video is even worse since each stream has different framing intervals depending on when each client connected with adaptive bitrate frames depending on each clients individual end-to-end latency.

This assumes the stream is an HTTP data stream. A lot of streams these days are ICY protocol, even if displayed by a Flash script interpreter.

Amos
Received on Mon Jul 23 2012 - 13:28:35 MDT

This archive was generated by hypermail 2.2.0 : Tue Jul 24 2012 - 12:00:02 MDT