Re: SOAP performance data and open benchmarks

Neelakantan Lakshminarayan wrote:
> Is there any statistics out there which provides guidance on the speed of
> compression versus speed of data transfer? This would vary for the modes of
> compression, of course.

I don't have statistics (as indeed it depends on the compression method,
and within compression methods, on the type of data), but the way we're
starting to measure it here is rather simple: we balance
encoding/decoding time versus network latency.

The goal of my current project is to make it so that our compression is
always better so that the problem of deciding whether you have to
compress or not goes away. That's achieved by transmitting compact
binary infosets directly.

> As a client or server how do I decide if this particular reques/response
> should be compressed or not? Is it practical to make this judgement call for
> every request/response?

It depends on your criteria, but to do it properly would require to take
a number of elements into account, probably more than you would want to.
That's one of the motivations for the above project.

SOAP::Lite has a simple option to tell it to compress only if the data
is above a certain size. I think that for a fair number of situations
that's an interesting heuristic.

--
Robin Berjon <[email protected]>
Research Engineer, Expway
7FC0 6F5F D864 EFB8 08CE  8E74 58E6 D5DB 4889 2488

You can read messages from the SOAP archive, unsubscribe from SOAP, or subscribe to other
DevelopMentor lists at http://discuss.develop.com.