RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Great to see the shootout results. Also interesting to see the Amazon RDS announcement (MySQL based) with possibility of using quadruple extra large EC2 instances: db.m2.4xlarge - 68 GB of RAM http://aws.typepad.com/aws/2009/10/introducing-rds-the-amazon-relational-dat abase-service-.html http://aws.typepad.com/aws/2009/10/two-new-ec2-instance-types-additional-mem ory.html Maybe next shootout the DB layer could look at Amazon RDS(mySQL) and PostgreSQL/PostGIS using a quadruple extra large instance 68Gb RAM. After reading Todd Hoff's blog I'd be curious to see if PostGIS could be configured to make use of large memory capacities and how it affects performance: http://highscalability.com/are-cloud-based-memory-architectures-next-big-thi ng Thanks Randy From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of Craig Miller Sent: Friday, October 23, 2009 8:38 PM To: 'OSGeo Discussions' Subject: RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results I agree wholeheartedly. It looks like the bottleneck was the database. I've been privy to some MapServer tests done by testing teams over several months and the result there was always deploying the data with long update cycles to the middle tier disks instead of using the database. Only then could the performance of the actual map servers be evaluated. Performance shootouts/testing take time to do correctly as each run teaches you more and more about how your deployment architecture affects the results. Craig Geospatial Software Engineer Spatial Minds, LLC http://spatialminds.com/ From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of antti roppola Sent: Friday, October 23, 2009 7:34 PM To: OSGeo Discussions Subject: Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results It was really interesting. The very close results suggests to me that the bottlenecks were external to the WMS and more related to external limitations like the ability to supply things like I/O. It would be interesting to have profiling data on where the response time was spent. For Mapserver it'd be a simple case of running Valgrinf and KCacheGrind: http://kcachegrind.sourceforge.net/cgi-bin/show.cgi/KcacheGrindIndex Case point. We had an in house app for crunching big raster and KCacheGrind showed us that an external library was the biggest bottleneck. A. On Sat, Oct 24, 2009 at 10:39 AM, Jeff McKenna jmcke...@gatewaygeomatics.com wrote: For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results
http://magazine.redhat.com/2007/12/12/tip-from-an-rhce-memory-storage-on-postgresql/ Doug Newcomb USFWS Raleigh, NC 919-856-4520 ext. 14 doug_newc...@fws.gov - The opinions I express are my own and are not representative of the official policy of the U.S.Fish and Wildlife Service or Dept. of the Interior. Life is too short for undocumented, proprietary data formats. Randy George rkgeo...@cadmaps .com To Sent by: 'OSGeo Discussions' discuss-boun...@l discuss@lists.osgeo.org ists.osgeo.org cc Subject 10/28/2009 01:42 RE: [OSGeo-Discuss] WMS Performance PMShootout presentation/results Please respond to rkgeo...@cadmaps. com; Please respond to OSGeo Discussions disc...@lists.os geo.org Great to see the shootout results. Also interesting to see the Amazon RDS announcement (MySQL based) with possibility of using quadruple extra large EC2 instances: db.m2.4xlarge - 68 GB of RAM http://aws.typepad.com/aws/2009/10/introducing-rds-the-amazon-relational-database-service-.html http://aws.typepad.com/aws/2009/10/two-new-ec2-instance-types-additional-memory.html Maybe next shootout the DB layer could look at Amazon RDS(mySQL) and PostgreSQL/PostGIS using a quadruple extra large instance 68Gb RAM. After reading Todd Hoff’s blog I’d be curious to see if PostGIS could be configured to make use of large memory capacities and how it affects performance: http://highscalability.com/are-cloud-based-memory-architectures-next-big-thing Thanks Randy From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of Craig Miller Sent: Friday, October 23, 2009 8:38 PM To: 'OSGeo Discussions' Subject: RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results I agree wholeheartedly. It looks like the bottleneck was the database. I’ve been privy to some MapServer tests done by testing teams over several months and the result there was always deploying the data with long update cycles to the middle tier disks instead of using the database. Only then could the performance of the actual map servers be evaluated. Performance shootouts/testing take time to do correctly as each run teaches you more and more about how your deployment architecture affects the results. Craig Geospatial Software Engineer Spatial Minds, LLC From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of antti roppola Sent: Friday, October 23, 2009 7:34 PM To: OSGeo Discussions Subject: Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results It was really interesting. The very close results suggests to me that the bottlenecks were external to the WMS and more related to external limitations like the ability to supply things like I/O. It would be interesting to have profiling data on where the response time was spent. For Mapserver it'd be a simple case of running Valgrinf and KCacheGrind: http://kcachegrind.sourceforge.net/cgi-bin/show.cgi/KcacheGrindIndex Case point. We had an in house app for crunching big raster and KCacheGrind showed us that an external library was the biggest bottleneck. A. On Sat, Oct 24, 2009 at 10:39 AM, Jeff McKenna jmcke...@gatewaygeomatics.com wrote: For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
On Sat, Oct 24, 2009 at 1:50 PM, Andrea Aime aa...@opengeo.org wrote: During the benchmark I had very little time to run profilers, but the few times I've tried in GeoServer the time seemed to be splitted quite equally between data fetching, actual drawing, and output image encoding... which is kind of the worst thing you can get out of a profile run, since it does not point to any culprit. OTOH, to me this says that CPU time is mostly being spent where it ought to be (drawing maps). :o) Antti ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Hi, Thanks for publishing this test for people who couldn't make it to Sydney. I would be interested in getting the Fcgi configuration of Mapserver in Apache. Any chance to get that ? Regards Guillaume Le samedi 24 octobre 2009 à 10:39 +1100, Jeff McKenna a écrit : For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
We used a static FCGI specification FastCgiServer /var/www/cgi-bin/mapserv.fcgi -processes 8 -initial_env LD_LIBRARY_PATH=/usr/local/lib:/opt/build/instant_client_11_1/lib:/opt/build /arcsde_c_sdk/lib Mike -- Michael Smith RSGIS Center US Army Corps of Engineers w: (603) 646-4765 On 10/24/09 6:36 PM, Guillaume Sueur no-re...@neogeo-online.net wrote: Hi, Thanks for publishing this test for people who couldn't make it to Sydney. I would be interested in getting the Fcgi configuration of Mapserver in Apache. Any chance to get that ? Regards Guillaume Le samedi 24 octobre 2009 à 10:39 +1100, Jeff McKenna a écrit : For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Le samedi 24 octobre 2009 à 21:53 +1100, Smith, Michael ERDC-USACE-NH a écrit : We used a static FCGI specification FastCgiServer /var/www/cgi-bin/mapserv.fcgi -processes 8 -initial_env LD_LIBRARY_PATH=/usr/local/lib:/opt/build/instant_client_11_1/lib:/opt/build /arcsde_c_sdk/lib Mike So no particular tweaking. thanks for this piece of info. regards, Guillaume ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Smith, Michael ERDC-USACE-NH wrote: We used a static FCGI specification FastCgiServer /var/www/cgi-bin/mapserv.fcgi -processes 8 -initial_env LD_LIBRARY_PATH=/usr/local/lib:/opt/build/instant_client_11_1/lib:/opt/build /arcsde_c_sdk/lib Mike, I added this info into the shared slides (on the last slide). -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Anyone know why Mapguide OS wasn't represented? I'd be happy to set one up if there is a conference in my area (Seattle/Vancouver BC area). Craig -Original Message- From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of Jeff McKenna Sent: Friday, October 23, 2009 4:39 PM To: osgeo Subject: [OSGeo-Discuss] WMS Performance Shootout presentation/results For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Unfortunately, no one stepped up to represent mapguide. Cheers Paul On 2009-10-23, at 9:28 PM, Craig Miller wrote: Anyone know why Mapguide OS wasn't represented? I'd be happy to set one up if there is a conference in my area (Seattle/Vancouver BC area). Craig -Original Message- From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of Jeff McKenna Sent: Friday, October 23, 2009 4:39 PM To: osgeo Subject: [OSGeo-Discuss] WMS Performance Shootout presentation/ results For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss __ Paul Spencer Chief Technology Officer DM Solutions Group Inc http://research.dmsolutions.ca/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
It was really interesting. The very close results suggests to me that the bottlenecks were external to the WMS and more related to external limitations like the ability to supply things like I/O. It would be interesting to have profiling data on where the response time was spent. For Mapserver it'd be a simple case of running Valgrinf and KCacheGrind: http://kcachegrind.sourceforge.net/cgi-bin/show.cgi/KcacheGrindIndex Case point. We had an in house app for crunching big raster and KCacheGrind showed us that an external library was the biggest bottleneck. A. On Sat, Oct 24, 2009 at 10:39 AM, Jeff McKenna jmcke...@gatewaygeomatics.com wrote: For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
RE: [OSGeo-Discuss] WMS Performance Shootout presentation/results
I agree wholeheartedly. It looks like the bottleneck was the database. I've been privy to some MapServer tests done by testing teams over several months and the result there was always deploying the data with long update cycles to the middle tier disks instead of using the database. Only then could the performance of the actual map servers be evaluated. Performance shootouts/testing take time to do correctly as each run teaches you more and more about how your deployment architecture affects the results. Craig Geospatial Software Engineer Spatial Minds, LLC http://spatialminds.com/ From: discuss-boun...@lists.osgeo.org [mailto:discuss-boun...@lists.osgeo.org] On Behalf Of antti roppola Sent: Friday, October 23, 2009 7:34 PM To: OSGeo Discussions Subject: Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results It was really interesting. The very close results suggests to me that the bottlenecks were external to the WMS and more related to external limitations like the ability to supply things like I/O. It would be interesting to have profiling data on where the response time was spent. For Mapserver it'd be a simple case of running Valgrinf and KCacheGrind: http://kcachegrind.sourceforge.net/cgi-bin/show.cgi/KcacheGrindIndex Case point. We had an in house app for crunching big raster and KCacheGrind showed us that an external library was the biggest bottleneck. A. On Sat, Oct 24, 2009 at 10:39 AM, Jeff McKenna jmcke...@gatewaygeomatics.com wrote: For those that did not make it to Sydney, here is the WMS Performance Shootout presentation with results (GeoServer vs MapServer): http://www.slideshare.net/gatewaygeomatics.com/wms-performance-shootout MapServer: power users who manage MapServer sites with high loads/map draws should take note of the results of MapServer CGI vs MapServer FastCGI, even in the case of Shapefiles and Rasters (yes, quite surprising). All: a lot of credit should go to Andrea Aime from GeoServer who worked very hard in bringing the MapServer team up to speed to learn the testing process. It was a great experience and we're already looking forward to next year. -jeff -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ mapserver-users mailing list mapserver-us...@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/mapserver-users -- Jeff McKenna FOSS4G Consulting and Training Services http://www.gatewaygeomatics.com/ ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] WMS Performance Shootout presentation/results
Craig Miller ha scritto: I agree wholeheartedly. It looks like the bottleneck was the database. I’ve been privy to some MapServer tests done by testing teams over several months and the result there was always deploying the data with long update cycles to the middle tier disks instead of using the database. Only then could the performance of the actual map servers be evaluated. Performance shootouts/testing take time to do correctly as each run teaches you more and more about how your deployment architecture affects the results. Actually running the benchmarks it looked the CPU was the bottleneck, we had 25% cpu load in the single test (so 100% of the one CPU used by that test, out of the 4 available) and during the higher loads the CPU usage was jumping between 80% and 100% (so almost all 4 cpu maxed out). The database server was hardly using more than 25-30% of its CPU and the network shouldn't have played the bottleneck either (past experience suggests it takes a tilecache to actually turn a 100Mbit line into a bottlenek), thought we don't have numbers for that During the benchmark I had very little time to run profilers, but the few times I've tried in GeoServer the time seemed to be splitted quite equally between data fetching, actual drawing, and output image encoding... which is kind of the worst thing you can get out of a profile run, since it does not point to any culprit. Getting a profile out of MapServer is surely going to be interesting, the codebase is smaller and there is no garbage collection going on so the results should be clearer. Looking forward to see those results :-) Cheers Andrea ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss