W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2014

Re: Call for Consensus: Remove "reference set" from HPACK (to address #552)

From: Greg Wilkins <gregw@intalio.com>
Date: Thu, 17 Jul 2014 00:22:00 +1000
Message-ID: <CAH_y2NHOpVqM2Pvh3P+61ERfjQqgBrObQc+0+iw1y7eUneYFjg@mail.gmail.com>
To: Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
Note also that those numbers are from a Jetty encoder that had a encoding
strategy designed with a RefSet in mind.   It makes some choices about
which encoding to use - eg it doesn't put set-cookie in the header table as
it is unlikely to be sent again.

So I'm making no representation about the absolute numbers as the choice of
encoding may not be optimal, and for things like Date it will depend on the
speed of a server if it is worthwhile putting them in the header table or
not.  Although I note my numbers are similar to others previously published.

The only representation I'm making with this data is a very limited one
about the relative values with/without various features - that being that
they don't make a lot of difference.

cheers





On 16 July 2014 23:58, Greg Wilkins <gregw@intalio.com> wrote:

>
> On 16 July 2014 23:30, Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com> wrote:
>
>> ​From the table, h2-12 shows best efficiency to me.
>
>
> Well to be precise h2-12-noRefSet-noCopy-static1 showed the best
> efficiency, but now h2-12-noRefSet-noCopy-static1-date5 does.
>
> However, I think the data set is way too small to be claiming any of these
> (other than date5) gives the best result.   Looking at the results from the
> individual stories in the test data, they are each +/- 5%. I have simply
> combined the results from all stories and there is no rigour in that as
> some are probably more frequent that others.
>
> So reporting the figures to two decimal places was not a valid thing for
> me to do and the following table probably more accurately represents the
> inaccuracies of this test:
>
> HPACK/HEADER TABLE0 1024409616384 h2-1264%47% 42%42% h2-12-noRefSet64%47%
> 43%42%h2-12-noRefSet-noCopy 64%47%42% 43%h2-12-noRefSet-noCopy-static164%
> 47%42%41% h2-12-noRefSet-noCopy-static1-date554%39% 35%34%
>
> Ie, none of these variations (except date5) give any significant savings
> on this data set.   Hence I think the simplest is the best:
> h2-12-noRefSet-noCopy-static1
>
> But I would be very interested to see a concrete proposal for
> h2-12-noRefSet-noCopy-static1-date5
>
> cheers
>
>
>
>
> --
> Greg Wilkins <gregw@intalio.com>
> http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that
> scales
> http://www.webtide.com  advice and support for jetty and cometd.
>



-- 
Greg Wilkins <gregw@intalio.com>
http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that scales
http://www.webtide.com  advice and support for jetty and cometd.
Received on Wednesday, 16 July 2014 14:22:29 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:09 UTC