Re: More benchmarking stuff

Chris Lilley, Computer Graphics Unit (lilley@v5.cgu.mcc.ac.uk)
Mon, 25 Jul 1994 11:52:31 +0200

In message <9407231902.AA01851@thyme.finesse.com> Mary Morris said:

> Personally, I'd like to see some benchmarks on things like
> using server includes and other processing type items. It would
> be nice to know how to develop efficent web pages, and to know
> what the tradeoffs between efficiency and low maintenance are.

Yeah, what she said. This is all done by guesswork at the moment. We don't
do server includes in case there is a real performance hit - but we don't
know.

We don't do fancy indexes because our server root is in AFS and it is
rumoured to be slow... Knowing the actual performance penalties for
particular things would enable much more rational decisions to be taken when
considering what snazzy thing to do.

Such benchmarks would need careful design though. For example, a system
which parses all html files to look for includes will perform differently
from one that marks including and non-including files as distinct, only
processing the latter. And systems of the second type will perform
differently if they can cache the result of the include so subsequent
accesses involve no processing.

So benchmarks will only allow comparison between version x of server Y and
version a of server B. In itself that would be handy, though.

--
Chris Lilley
+--------------------------------------------------------------------------+
|Technical Author, ITTI Computer Graphics & Visualisation Training Project |
+--------------------------------------------------------------------------+
| Computer Graphics Unit, | Internet: C.C.Lilley@mcc.ac.uk |
| Manchester Computing Centre, | Janet: C.C.Lilley@uk.ac.mcc |
| Oxford Road, | Voice: +44 61 275 6045 |
| Manchester, UK. M13 9PL | Fax: +44 61 275 6040 |
| X400: /I=c/S=lilley/O=manchester-computing-centre/PRMD=UK.AC/ADMD= /C=GB/|
| <A HREF="http://info.mcc.ac.uk/CGU/staff/lilley/lilley.html">my page</A> |
+--------------------------------------------------------------------------+