[Zope] Tomcat vrs Zope Round One

Scott Boyd sboyd@futures.com
Thu, 27 Apr 2000 17:10:21 -0700 (PDT)


How dynamic is your test page? Is the page re-rendered with differing
content on each request? If the content doesn't change Resin will just
hand-out a cached page. 

Oh, and which system did you run Resin on? (The slow or the fast one.)

Curiously,

S.

On Thu, 27 Apr 2000, J. Atwood wrote:

> A really interesting suggestion... And so I did..
> 
> Server Software:        Resin/1.1
> Server Port:            8051
> 
> Document Path:          /test2.jsp
> Document Length:        158 bytes
> 
> Concurrency Level:      25
> Time taken for tests:   2.798 seconds
> Complete requests:      1000
> Failed requests:        0
> Total transferred:      384000 bytes
> HTML transferred:       158000 bytes
> Requests per second:    357.40
> Transfer rate:          137.24 kb/s received
> 
> Connnection Times (ms)
>               min   avg   max
> Connect:        0     0    13
> Processing:    17    33   112
> Total:         17    33   125
> 
> I hit it a bunch of times... tried it again and again..
> 
> I went back and tried Tomcat and got the same (old) results. Well, on the
> very quick and narrow Resin seems to be much faster...
> 
> And so it continues...
> 
> J
> 
> > From: Scott Boyd <sboyd@futures.com>
> > Date: Thu, 27 Apr 2000 16:16:42 -0700 (PDT)
> > To: "J. Atwood" <Jatwood@bwanazulia.com>
> > Subject: Re: [Zope] Tomcat vrs Zope Round One
> > 
> > 
> > Do this test against Resin (www.caucho.com) another JSP engine. I'd be
> > interested in if that was faster, because Resin has it's own caching
> > abilities like Zope.
> > 
> > S.
> > 
> > 
> > On Thu, 27 Apr 2000, J. Atwood wrote:
> > 
> >> I just set up a Jakarta Tomcat installation (for someone else!) and was able
> >> to do some quick benchmarking on two *very* similar machines and here is
> >> what I found.
> >> 
> >> Zope Machine: RH 6.1, 300 MHz, 256 MB RAM, 8 GB EIDE Drive - Zope. 2.1.4
> >> 
> >> Tomcat Machine: RH 6.1, 500 MHz, 512 MB RAM, 3 x 9.1 SCSI in RAID - Jakarta
> >> Tomcat 3.1
> >> 
> >> Using AB (ApacheBench) and hitting a test JSP page that I replicated in Zope
> >> (exactly the same words, etc).
> >> 
> >> Tomcat:
> >> Document Path:          /test.jsp
> >> Document Length:        168 bytes
> >> 
> >> Concurrency Level:      100
> >> Time taken for tests:   5.306 seconds
> >> Complete requests:      100
> >> Failed requests:        0
> >> Total transferred:      57024 bytes
> >> HTML transferred:       16800 bytes
> >> Requests per second:    18.85
> >> Transfer rate:          10.75 kb/s received
> >> 
> >> Connnection Times (ms)
> >> min   avg   max
> >> Connect:       63   121   177
> >> Processing:  3473  3911  5063
> >> Total:       3536  4032  5240
> >> 
> >> Zope:
> >> 
> >> Document Path:          /test.html
> >> Document Length:        157 bytes
> >> 
> >> Concurrency Level:      100
> >> Time taken for tests:   2.141 seconds
> >> Complete requests:      100
> >> Failed requests:        0
> >> Total transferred:      36360 bytes
> >> HTML transferred:       15857 bytes
> >> Requests per second:    46.71
> >> Transfer rate:          16.98 kb/s received
> >> 
> >> Connnection Times (ms)
> >> min   avg   max
> >> Connect:        5     8    12
> >> Processing:    94  1091  2117
> >> Total:         99  1099  2129
> >> 
> >> ------
> >> 
> >> I thought... wow... so I ran it again, and again and again. Each time the
> >> Tomcat stats went down as Zope stayed constant.
> >> 
> >> Now I know that this does not mean that Zope is the fastest thing on the
> >> planet but it is an interesting comparison. I mean, 2 x as fast on a machine
> >> that is half as fast. Strange but true. So much for "Java Speed"
> >> 
> >> I decided to do it again.. this time with 25/1000 figuring Tomcat would
> >> build up and start really flying. Here is what I got.
> >> 
> >> Tomcat:
> >> 
> >> Concurrency Level:      25
> >> Time taken for tests:   41.252 seconds
> >> Complete requests:      1000
> >> Failed requests:        0
> >> Total transferred:      582555 bytes
> >> HTML transferred:       177000 bytes
> >> Requests per second:    24.24
> >> Transfer rate:          14.12 kb/s received
> >> 
> >> Connnection Times (ms)
> >> min   avg   max
> >> Connect:        0     0     5
> >> Processing:   183  1015  2300
> >> Total:        183  1015  2305
> >> 
> >> Zope:
> >> Concurrency Level:      25
> >> Time taken for tests:   19.360 seconds
> >> Complete requests:      1000
> >> Failed requests:        0
> >> Total transferred:      360000 bytes
> >> HTML transferred:       157000 bytes
> >> Requests per second:    51.65
> >> Transfer rate:          18.60 kb/s received
> >> 
> >> Connnection Times (ms)
> >> min   avg   max
> >> Connect:        0     0     2
> >> Processing:    71   478   540
> >> Total:         71   478   542
> >> 
> >> 
> >> Yupp... better for both but Zope still was able to serve up the same content
> >> on a slower machine twice as fast.
> >> 
> >> Next, I guess, would be to write some code to actually "do" something on the
> >> machine and see what it could do. Also, I should level the playing field by
> >> installing Zope on the faster machine and seeing what it can do there.
> >> 
> >> Comments?
> >> 
> >> J
> >> 
> >> 
> >> 
> >> _______________________________________________
> >> Zope maillist  -  Zope@zope.org
> >> http://lists.zope.org/mailman/listinfo/zope
> >> **   No cross posts or HTML encoding!  **
> >> (Related lists -
> >> http://lists.zope.org/mailman/listinfo/zope-announce
> >> http://lists.zope.org/mailman/listinfo/zope-dev )
> >> 
>