Re: tomcat5/tomcat4 comparison
Just wanted to add using jk2 with a unix pipe. I am suprised to find that on Linux at least, the pipe is not as fast as sockets. -walter +--+ | Tomcat | Apache| Connector| Mean | | Version | Version | | Connection | +==+ | 4.1.30 | 1.3.29| webapp | 73.81 | | | +---+ | | | mod_jk | 40.59 | | +---+ | | 2.0.49| webapp | 74.19 | | | +---+ | | | jk2(ajp) | 40.16 | | | +---+ | | | jk2(coyote) | 29.31 | | +---+ | | none | http(coyote) | 18.98 | +--+ | 5.0.19 | 2.0.49| jk2(unix)| 29.75 | | | +---+ | | | jk2(socket) | 27.35 | | +---+ | | none | http(coyote) | 18.76 | +--+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: tomcat5/tomcat4 comparison
Posting my findings for Redhat Linux 7.3 on a Intel Zeon 2.4 with a GB of ram. My kernel is not SMP even though the Zeon has hyperthreading. I also ran ab on the same machine as apache and tomcat. I was interested in seeing the WARP connector maintained, so I also include those numbers in my tests. I find that WARP works with both Apache 1.3.29 and 2.0.49. It does not however work with tomcat 5. I also found that it does not have the performance of the jk2 connector. This was told to me, but I didn't believe it until I tested it myself. I found that the warp connector was removed from tomcat 5, and I did not work on porting it. +--+ | Tomcat | Apache| Connector| Mean | | Version | Version | | Connection | +==+ | 4.1.30 | 1.3.29| webapp | 73.81 | | +---+ | | 2.0.49| webapp | 74.19 | | | +---+ | | | jk2(ajp) | 40.16 | | | +---+ | | | jk2(coyote) | 29.31 | | +---+ | | none | http(coyote) | 18.98 | +--+ | 5.0.19 | 2.0.49| jk2 | 27.35 | | +---+ | | none | http(coyote) | 18.76 | +--+ All my tests were using the HelloWorldExample servlet. I did not perform any static page tests. The example servlet changed URLs between 4.1 and 5.0 tomcat versions. I used the ab version from my 1.3.29 apache. I don't know if there were any changes between it and the one in 2.0.49. Tomcat 4.1.30 - ab -n 1 -c 30 http://www/examples/servlet/HelloWorldExample Tomcat 5.0.19 - ab -n 1 -c 30 http://www/servlets-examples/servlet/HelloWorldExample These numbers did not match my expectations, but then I am not familiar with the code in each module. It shows that using the Http Connector is faster than going through apache for the dynamic content. I would expect dynamic content to be the reverse, so consider what percentage of your pages/images are dynamic when choosing which to use. As a side note, I believe that since webapp works with 2.0.49, it does support threads. Because it doesn't have the performance of jk2, I won't be trying to get it out of deprication. I know that my tests don't compare directly to the other post, and it doesn't look like there is as much of a change in performance as the other tests showed between versions. The HelloWorldExample servlet was 1.2% faster using 5.0.19 compared to 4.1.30 with the HTTP Connector. It was 7.2% faster when comparing 5.0.19 and 4.1.30 using the jk2 Connector. jk2 4.1.30 100.0 % http 4.1.30 154.4 % jk2 5.0.19 107.2 % http 5.0.19 156.2 % -walter Apache 1.3.29/Tomcat 4.1.30/mod_webapp 1.20 Total transferred: 5442176 bytes HTML transferred: 3931572 bytes Requests per second:406.47 [#/sec] (mean) Time per request: 73.81 [ms] (mean) Time per request: 2.46 [ms] (mean, across all concurrent requests) Transfer rate: 221.21 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/mod_webapp 1.20 Total transferred: 5641692 bytes HTML transferred: 3931179 bytes Requests per second:404.38 [#/sec] (mean) Time per request: 74.19 [ms] (mean) Time per request: 2.47 [ms] (mean, across all concurrent requests) Transfer rate: 228.14 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/jk2 2.04(org.apache.ajp.tomcat4.Ajp13Connector) Total transferred: 5575013 bytes HTML transferred: 3933537 bytes Requests per second:746.99 [#/sec] (mean) Time per request: 40.16 [ms] (mean) Time per request: 1.34 [ms] (mean, across all concurrent requests) Transfer rate: 416.45 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/jk2 2.04(org.apache.coyote.tomcat4.CoyoteConnector) Total transferred: 5910030 bytes HTML transferred: 4056885 bytes Requests per second:1023.44 [#/sec] (mean) Time per request: 29.31 [ms] (mean) Time per request: 0.98 [ms] (mean, across all concurrent requests) Transfer rate: 604.85 [Kbytes/sec] received Tomcat 4.1.30/Coyote Http Total transferred: 5541060 bytes HTML transferred: 4058100 bytes Requests per second:1581.03 [#/sec] (mean) Time per request: 18.98 [ms] (mean) Time per request: 0.63 [ms] (mean, across all concurrent requests) Transfer rate: 876.06 [Kbytes/sec] received Tomcat 5.0.14/Coyote Http Total transferred: 5273150 bytes HTML transferred: 3598975 bytes Requests per second:1599.49 [#/sec] (mean) Time per request: 18.76 [ms] (mean)
RE: tomcat5/tomcat4 comparison
Hi, Great stuff, thanks for posting. Now if we could only get people to search the archives before posting questions ;) Yoav Shapira Millennium Research Informatics -Original Message- From: Walter Truitt [mailto:[EMAIL PROTECTED] Sent: Wednesday, March 31, 2004 3:24 PM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: tomcat5/tomcat4 comparison Posting my findings for Redhat Linux 7.3 on a Intel Zeon 2.4 with a GB of ram. My kernel is not SMP even though the Zeon has hyperthreading. I also ran ab on the same machine as apache and tomcat. I was interested in seeing the WARP connector maintained, so I also include those numbers in my tests. I find that WARP works with both Apache 1.3.29 and 2.0.49. It does not however work with tomcat 5. I also found that it does not have the performance of the jk2 connector. This was told to me, but I didn't believe it until I tested it myself. I found that the warp connector was removed from tomcat 5, and I did not work on porting it. +--+ | Tomcat | Apache| Connector| Mean | | Version | Version | | Connection | +==+ | 4.1.30 | 1.3.29| webapp | 73.81 | | +---+ | | 2.0.49| webapp | 74.19 | | | +---+ | | | jk2(ajp) | 40.16 | | | +---+ | | | jk2(coyote) | 29.31 | | +---+ | | none | http(coyote) | 18.98 | +--+ | 5.0.19 | 2.0.49| jk2 | 27.35 | | +---+ | | none | http(coyote) | 18.76 | +--+ All my tests were using the HelloWorldExample servlet. I did not perform any static page tests. The example servlet changed URLs between 4.1 and 5.0 tomcat versions. I used the ab version from my 1.3.29 apache. I don't know if there were any changes between it and the one in 2.0.49. Tomcat 4.1.30 - ab -n 1 -c 30 http://www/examples/servlet/HelloWorldExample Tomcat 5.0.19 - ab -n 1 -c 30 http://www/servlets-examples/servlet/HelloWorldExample These numbers did not match my expectations, but then I am not familiar with the code in each module. It shows that using the Http Connector is faster than going through apache for the dynamic content. I would expect dynamic content to be the reverse, so consider what percentage of your pages/images are dynamic when choosing which to use. As a side note, I believe that since webapp works with 2.0.49, it does support threads. Because it doesn't have the performance of jk2, I won't be trying to get it out of deprication. I know that my tests don't compare directly to the other post, and it doesn't look like there is as much of a change in performance as the other tests showed between versions. The HelloWorldExample servlet was 1.2% faster using 5.0.19 compared to 4.1.30 with the HTTP Connector. It was 7.2% faster when comparing 5.0.19 and 4.1.30 using the jk2 Connector. jk2 4.1.30 100.0 % http 4.1.30 154.4 % jk2 5.0.19 107.2 % http 5.0.19 156.2 % -walter Apache 1.3.29/Tomcat 4.1.30/mod_webapp 1.20 Total transferred: 5442176 bytes HTML transferred: 3931572 bytes Requests per second:406.47 [#/sec] (mean) Time per request: 73.81 [ms] (mean) Time per request: 2.46 [ms] (mean, across all concurrent requests) Transfer rate: 221.21 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/mod_webapp 1.20 Total transferred: 5641692 bytes HTML transferred: 3931179 bytes Requests per second:404.38 [#/sec] (mean) Time per request: 74.19 [ms] (mean) Time per request: 2.47 [ms] (mean, across all concurrent requests) Transfer rate: 228.14 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/jk2 2.04(org.apache.ajp.tomcat4.Ajp13Connector) Total transferred: 5575013 bytes HTML transferred: 3933537 bytes Requests per second:746.99 [#/sec] (mean) Time per request: 40.16 [ms] (mean) Time per request: 1.34 [ms] (mean, across all concurrent requests) Transfer rate: 416.45 [Kbytes/sec] received Apache 2.0.49/Tomcat 4.1.30/jk2 2.04(org.apache.coyote.tomcat4.CoyoteConnector) Total transferred: 5910030 bytes HTML transferred: 4056885 bytes Requests per second:1023.44 [#/sec] (mean) Time per request: 29.31 [ms] (mean) Time per request: 0.98 [ms] (mean, across all concurrent requests) Transfer rate: 604.85 [Kbytes/sec] received Tomcat 4.1.30/Coyote Http Total transferred: 5541060 bytes HTML transferred: 4058100 bytes Requests per second:1581.03 [#/sec] (mean) Time
RE: tomcat5/tomcat4 comparison
I pray for that day, but not holding my breath. :) peter Shapira, Yoav [EMAIL PROTECTED] wrote: Hi, Great stuff, thanks for posting. Now if we could only get people to search the archives before posting questions ;) Yoav Shapira Millennium Research Informatics - Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time.
Re: tomcat5/tomcat4 comparison
Walter Truitt wrote: I know that my tests don't compare directly to the other post, and it doesn't look like there is as much of a change in performance as the other tests showed between versions. The HelloWorldExample servlet was 1.2% faster using 5.0.19 compared to 4.1.30 with the HTTP Connector. It was 7.2% faster when comparing 5.0.19 and 4.1.30 using the jk2 Connector. There's no difference because the HWE is a rather bad test (if you're out there to test throughtput): - it retrieves an i18n bundle on each request, and gets a String from it (profiling has shown it's not something trivial) - it uses a writer: this means most of the activity is converting the chars to bytes Since it doesn't exercise anything else (filters, request dispatcher, any API method), there's little difference between 4.1.x and 5.0.x (which now also use the same connectors, so there's no difference there either). If you want to test the output speed, you need a servlet which simply writes byte arrays (and does nothing else). You could also use keepalive when testing (unfortunately, ab doesn't allow a mix). Again, you'll see bigger differences. -- x Rémy Maucherat Developer Consultant JBoss Group (Europe) SàRL x - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: tomcat5/tomcat4 comparison
Just a quick compare using 'ab -n 1 -c 25' 4.1.30 Total transferred: 553 bytes HTML transferred: 405 bytes Requests per second:447.18 [#/sec] (mean) Time per request: 55.905 [ms] (mean) Time per request: 2.236 [ms] (mean, across all concurrent requests) Transfer rate: 241.48 [Kbytes/sec] received 5.0.20 Total transferred: 553 bytes HTML transferred: 405 bytes Requests per second:502.55 [#/sec] (mean) Time per request: 49.747 [ms] (mean) Time per request: 1.990 [ms] (mean, across all concurrent requests) Transfer rate: 271.38 [Kbytes/sec] received OS WINXP java version 1.4.2_03 Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_03-b02) Java HotSpot(TM) Client VM (build 1.4.2_03-b02, mixed mode) Test run with empty log after sturtup using 4x1000 requests to 'worm-up' on HelloWorld servlet, thus testing pure non-app performance (just servlet container and http connector). MT. -Original Message- From: Neil MacMillan [mailto:[EMAIL PROTECTED] Sent: Tuesday, March 30, 2004 7:44 PM To: [EMAIL PROTECTED] Subject: tomcat5/tomcat4 comparison Hi all, I'm trying to gather all the information I can about pros/cons of migrating 10-15 servers from tomcat4 to tomcat5. I have been unsuccessful in finding some good performance comparisons, and major feature differences between the two (major)versions. Can anyone provide some links/docs for this? also, is it possible to use JSP 2.0 in tomcat4.x? thanks neilm - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: tomcat5/tomcat4 comparison
Forgot the following: Apache 2.0.49 (405 bytes html file) Total transferred: 687 bytes HTML transferred: 405 bytes Requests per second:630.80 [#/sec] (mean) Time per request: 15.853 [ms] (mean) Time per request: 1.585 [ms] (mean, across all concurrent requests) Transfer rate: 423.14 [Kbytes/sec] received Apache2.0.49/mod_jk2.04/Tomcat5.0.20 (same html file, but served through TC and JK2) Total transferred: 666 bytes HTML transferred: 405 bytes Requests per second:401.84 [#/sec] (mean) Time per request: 24.886 [ms] (mean) Time per request: 2.489 [ms] (mean, across all concurrent requests) Transfer rate: 261.31 [Kbytes/sec] received Apache2.0.49/mod_jk2.04/Tomcat5.0.20 (HelloWorldExamle servlet) Total transferred: 591 bytes HTML transferred: 405 bytes Requests per second:330.10 [#/sec] (mean) Time per request: 75.734 [ms] (mean) Time per request: 3.029 [ms] (mean, across all concurrent requests) Transfer rate: 190.50 [Kbytes/sec] received So here are my results (would like to see some other platform results). TC41 100% JK2 78% TC50 112% Just a quick compare using 'ab -n 1 -c 25' 4.1.30 Total transferred: 553 bytes HTML transferred: 405 bytes Requests per second:447.18 [#/sec] (mean) Time per request: 55.905 [ms] (mean) Time per request: 2.236 [ms] (mean, across all concurrent requests) Transfer rate: 241.48 [Kbytes/sec] received 5.0.20 Total transferred: 553 bytes HTML transferred: 405 bytes Requests per second:502.55 [#/sec] (mean) Time per request: 49.747 [ms] (mean) Time per request: 1.990 [ms] (mean, across all concurrent requests) Transfer rate: 271.38 [Kbytes/sec] received OS WINXP java version 1.4.2_03 Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_03-b02) Java HotSpot(TM) Client VM (build 1.4.2_03-b02, mixed mode) Test run with empty log after sturtup using 4x1000 requests to 'worm-up' on HelloWorld servlet, thus testing pure non-app performance (just servlet container and http connector). MT. -Original Message- From: Neil MacMillan [mailto:[EMAIL PROTECTED] Sent: Tuesday, March 30, 2004 7:44 PM To: [EMAIL PROTECTED] Subject: tomcat5/tomcat4 comparison Hi all, I'm trying to gather all the information I can about pros/cons of migrating 10-15 servers from tomcat4 to tomcat5. I have been unsuccessful in finding some good performance comparisons, and major feature differences between the two (major)versions. Can anyone provide some links/docs for this? also, is it possible to use JSP 2.0 in tomcat4.x? thanks neilm - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: tomcat5/tomcat4 comparison
On Tue, Mar 30, 2004 at 09:44:23AM -0800, Neil MacMillan wrote: : Hi all, I'm trying to gather all the information I can about pros/cons of : migrating 10-15 servers from tomcat4 to tomcat5. I have been unsuccessful in : finding some good performance comparisons, and major feature differences : between the two (major)versions. Can anyone provide some links/docs for : this? Jokes aside, have you checked the release docs? I don't have the exact URL but they're on the website. just a few TC5 new features/improvements that come to mind: - support for servlet spec 2.4 and JSP spec 2.0 - clustering - improved manager app (IIRC) Unless you see some total whiz-bang feature that you've been craving, then it's a matter of deciding whether you want to upgrade just for the heck of it, or to be running the latest version. (-and that's not necessarily a bad thing.) -QM -- software -- http://www.brandxdev.net tech news -- http://www.RoarNetworX.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: tomcat5/tomcat4 comparison
Basically, the whiz-bang feature I want is JSP2.0 I have looked over whats on the site as they do show about 10 differences in containers, so the clustering, GC improvments, some others, I've read. I just need to justify the move of a handful of servers to tomcat 5. neilm -Original Message- From: QM [mailto:[EMAIL PROTECTED] Sent: Tue 3/30/2004 11:55 AM To: Tomcat Users List Cc: Subject:Re: tomcat5/tomcat4 comparison On Tue, Mar 30, 2004 at 09:44:23AM -0800, Neil MacMillan wrote: : Hi all, I'm trying to gather all the information I can about pros/cons of : migrating 10-15 servers from tomcat4 to tomcat5. I have been unsuccessful in : finding some good performance comparisons, and major feature differences : between the two (major)versions. Can anyone provide some links/docs for : this? Jokes aside, have you checked the release docs? I don't have the exact URL but they're on the website. just a few TC5 new features/improvements that come to mind: - support for servlet spec 2.4 and JSP spec 2.0 - clustering - improved manager app (IIRC) Unless you see some total whiz-bang feature that you've been craving, then it's a matter of deciding whether you want to upgrade just for the heck of it, or to be running the latest version. (-and that's not necessarily a bad thing.) -QM -- software -- http://www.brandxdev.net tech news -- http://www.RoarNetworX.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: tomcat5/tomcat4 comparison
At 12:22 PM 3/30/2004, you wrote: Basically, the whiz-bang feature I want is JSP2.0 I have looked over whats on the site as they do show about 10 differences in containers, so the clustering, GC improvments, some others, I've read. I just need to justify the move of a handful of servers to tomcat 5. Up there amongst the best reasons is availability of support. As with all OSS, you benefit by aligning yourself to the main trunk of development from a bug fix and *especially* forum support point-of-view. Many (most, it seems) of the regular contributors to this list have migrated to Tomcat5, so there's less of an audience anxious to work on Tomcat4 problems. justin -Original Message- From: QM [mailto:[EMAIL PROTECTED] Sent: Tue 3/30/2004 11:55 AM To: Tomcat Users List Cc: Subject:Re: tomcat5/tomcat4 comparison On Tue, Mar 30, 2004 at 09:44:23AM -0800, Neil MacMillan wrote: : Hi all, I'm trying to gather all the information I can about pros/cons of : migrating 10-15 servers from tomcat4 to tomcat5. I have been unsuccessful in : finding some good performance comparisons, and major feature differences : between the two (major)versions. Can anyone provide some links/docs for : this? Jokes aside, have you checked the release docs? I don't have the exact URL but they're on the website. just a few TC5 new features/improvements that come to mind: - support for servlet spec 2.4 and JSP spec 2.0 - clustering - improved manager app (IIRC) Unless you see some total whiz-bang feature that you've been craving, then it's a matter of deciding whether you want to upgrade just for the heck of it, or to be running the latest version. (-and that's not necessarily a bad thing.) -QM - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: tomcat5/tomcat4 comparison
Howdy, Up there amongst the best reasons is availability of support. As with all OSS, you benefit by aligning yourself to the main trunk of development from a bug fix and *especially* forum support point-of-view. Many (most, it seems) of the regular contributors to this list have migrated to Tomcat5, so there's less of an audience anxious to work on Tomcat4 problems. Yup, well-said. JSP 2.0 is a tomcat5 feature: if you require it, you require tomcat5, end of story ;) Yoav Shapira This e-mail, including any attachments, is a confidential business communication, and may contain information that is confidential, proprietary and/or privileged. This e-mail is intended only for the individual(s) to whom it is addressed, and may not be saved, copied, printed, disclosed or used by anyone else. If you are not the(an) intended recipient, please immediately delete this e-mail from your computer system and notify the sender. Thank you. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
