W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2013

RE: bohe and delta experimentation...

From: RUELLAN Herve <Herve.Ruellan@crf.canon.fr>
Date: Thu, 17 Jan 2013 16:50:58 +0000
To: James M Snell <jasnell@gmail.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Message-ID: <6C71876BDCCD01488E70A2399529D5E52E10F3@ADELE.crf.canon.fr>
James,

> -----Original Message-----
> From: James M Snell [mailto:jasnell@gmail.com]
> Sent: mercredi 16 janvier 2013 23:07
> To: ietf-http-wg@w3.org
> Subject: bohe and delta experimentation...
> 
> After going a number of scenarios with bohe using a variety of stream-
> compression scenarios it's painfully obvious that there is really no way
> around the CRIME issue when using stream-compression. So with that, I'm
> turning my attention to the use of Roberto's delta encoding and exploring
> whether or not binary optimized values can make a significant difference (as
> opposed to simply dropping in huffman-encoded text everywhere).
> 
> 
> I'm starting with dates first...
> 

From our dataset (see [1]), about 20% of the response headers are dates. The second type of values to optimize would be integer, that correspond to roughly 15% of the response headers.

One caveat: sometimes the value of a header that should be a date is not a date. Therefore an escape mechanism must be provided to encode these values.

Another interesting thing to investigate is encoding a value relatively to a previous value, as was proposed by Willy Tarreau. This is especially efficient for URLs.

Hervé.

[1] http://www.ietf.org/proceedings/85/slides/slides-85-httpbis-1.pdf



Received on Thursday, 17 January 2013 16:51:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 17 January 2013 16:51:31 GMT