RE: Tomcat benchmarks from orionserver.com, Please comment. Impor tant

["Roytman, Alex" <ro...@peacetech.com>]

<I think that performance is over-stated here, though.  If performance was
<our number-one goal we would not be using Java at all"

Sorry, can't agree with it. Performance is important
And we are not talking about Java vs. C++ we are talking about different
implementations of the same specs. 
I trust, as Tim Kientzle said, Tomcat will catch up with other
implementations as far as performance is concerned. I wonder if 3.2 is
already better than 3.1? :-) 

-----Original Message-----
From: Tice, Ed [mailto:ETice@bluestone.com]
Sent: Wednesday, July 19, 2000 3:13 PM
To: 'James Cook '; 'tomcat-user@jakarta.apache.org '; 'kientzle@acm.org
'
Subject: RE: Tomcat benchmarks from orionserver.com, Please comment.
Impor tant


I think that performance is over-stated here, though.  If performance was
our number-one goal we would not be using Java at all.  We would write CGIs
in assembly language.  Tomcat has great load-balancing (if using integrated
with apache and mod_jserv) so you can throw harware (which is cheap!) at the
problem.

Ed 

-----Original Message-----
From: James Cook
To: tomcat-user@jakarta.apache.org; kientzle@acm.org
Sent: 7/19/00 3:03 PM
Subject: Re: Tomcat benchmarks from orionserver.com, Please comment.
Important

This is not meant as a slam against the very capable developers of
Tomcat.

It really is a testament to these other opensource projects (resin
inparticular), that a single developer or two can produce such capable
products in such accelerated timeframes. Resin is mainly developed by
one or
two persons, and didn't have the advantage(?) of an existing codebase.
They
pumped out a very respectable servlet 2.2 compliant server months ago.

jim

----- Original Message -----
From: "Tim Kientzle" <ki...@acm.org>
To: <to...@jakarta.apache.org>
Sent: Wednesday, July 19, 2000 2:22 PM
Subject: Re: Tomcat benchmarks from orionserver.com, Please comment.
Important


> I personally find it neither surprising nor disturbing that Tomcat 3.1
> is slower than competing products right now.  Tomcat is a very
> new project, and as with any new project, the first focus is on
> stability and features.  Once the system generally does what people
> need it to do, the developers can then move on to improve performance.
>
> Given the number of people using Tomcat, performance will improve.
> I'm comfortable with that; upgrading versions of Tomcat in the
> future will be much easier for me than switching between very
> different servers.  If you're building a new site and
> expect to spend your first year building traffic, then Tomcat
> might be a good choice; by the time you need it, Tomcat's
> performance will be much better than it is today.
>
> The best thing you can do to help is to use Tomcat, understand
> how it works, and pass your suggestions (and fixes if you can)
> back to the Tomcat developers.  Open-source projects thrive when
> there is an active base of user/developers.  Right now, it appears
> Tomcat has a real need for people to work on documentation, so
> if you think you understand how to use Tomcat, volunteer to write
> something.
>
> Also keep in mind that Apache is somewhat unique among open-source
> projects in being significantly faster than most of it's commercial
> counterparts.  Since Tomcat is the reference implementation, it will
> probably have to focus on features and portability, sometimes to the
> detriment of performance.  It's quite possible that commercial
> servlet engines will always be somewhat faster than Tomcat, although
> that margin should shrink considerably once Tomcat is past its
> infancy.
>
> - Tim
>
> "Roytman, Alex" wrote:
> >
> > According to Orion Tomcat 3.1 is almost 8 times slower than Orion or
Resin.
> > Are this fair results?
> > Could someone from Jakarta team comment on it. We all like Tomcat,
so
please
> > do say something
> >
> > form Orion web site:
> > Requests per second under the four different stress levels:
> >
> > Server  1 x 1   4 x 1   8 x 2 (16)      16 x 4(64)
> > Orion   615.00  622.36  597.1   579.89
> > Tomcat  84.19   86.59   74.48   68.24
> > JRun            386.17  387.62  370.1   306.34
> > Resin   606.01  598.77  588.74  575.49
> > Weblogic        122.55  124.28  125.73  116.81
> >
> > see http://www.orionserver.com/benchmarks/benchmark.html
> >
> > -----Original Message-----
> > From: Chon-Chon Tang [mailto:ztang@weber.lcs.mit.edu]
> > Sent: Tuesday, July 18, 2000 4:35 PM
> > To: tomcat-user@jakarta.apache.org
> > Subject: Re: Tomcat benchmarks from orionserver.com
> >
> > Hmm, first of all, apologies if this is the wrong list for this
discussion
> > (re-direct if a better one exists).
> >
> > Caucho also has done quite a bit of back-patting on the performance
of
> > Resin on JSP/Servlet content serving.  Orion's performance seems
similarly
> > impressive... both appear to be on the order of 5-8 times faster
than
> > Tomcat, and in fact even superior to most (substantially more
expensive)
> > commercial products.
> >
> > Has anyone out there had practical experiences with these less
publicized
> > servers?  Since Resin is also standards-meeting (both JSP and
Servlet),
> > why might I choose Tomcat over Resin on the short-term?  This isn't
to
> > demean the value of Jakarta and it's development efforts; I'm just
trying
> > to evaluate a product for the near term.
> >
> > On Tue, 18 Jul 2000, Aleksey Tsalolikhin wrote:
> >
> > > FYI, orionserver.com's benchmarks of Tomcat and several other
> > > JSP engines, dated 17 July 2000:
> > >
> > > http://www.orionserver.com/benchmarks/benchmark.html
> > >
> > > Yours,
> > > Aleksey Tsalolikhin
> > >
> > >
> > >