tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Craig R. McClanahan" <Craig.McClana...@eng.sun.com>
Subject Re: Tomcat benchmarks from orionserver.com, Please comment. Important
Date Thu, 20 Jul 2000 17:43:40 GMT
"Roytman, Alex" wrote:

> According to Orion Tomcat 3.1 is almost 8 times slower than Orion or Resin.
> Are this fair results?

What do you mean by "fair"?

Are you wondering whether the people running the test are lying about
the
results?  I have *absolute* faith that the people who ran this test are
ethical
and honest, and that the results they reported are the results that
actually
occurred.

Was this a test that focused on trying to maximize the difference
between Orion
and Resin and the others?  Let's not be silly.  The facts are the facts.

Does this benchmark, or *any* benchmark, represent how *your*
application will
run on the various servers?  Only if it is representative of the
workload your
application requires.  I like to say "Hello, World benchmarks only tell
you how
fast the server can say Hello, World" :-)  IMHO the only benchmarks that
really
matters is your application itself, or a simulation of the app that
faithfully
represents it.

>
> Could someone from Jakarta team comment on it. We all like Tomcat, so please
> do say something
>

You will note one important issue -- the benchmark tested Tomcat 3.1,
not
3.2dev.  This is an entirely reasonable decision at the time the
benchmarks were
run.  My suggestion is that you try your app under 3.1 and then 3.2
(where
there's been significant improvements in performance).

If Orion and Resin are still faster, then more power to the developers
of those
servers.  They are doing a great job at providing all of us with
multiple good
answers to "what server should I run", instead of only having one
option.  I
would be really concerned if Tomcat (or any other particular server) was
the
only possible way to run servet/JSP based applications.

And, oh by the way, "let the performance improvement race continue ..." 
:-)

Craig McClanahan

Mime
View raw message