[Geoserver-users] Choosing a webmapping server: GeoServer?

GeoServer GeoServer User
1

Hi list,

I'm new to the list and subscribed to have your input about GeoServer and webmapping applications.

Honestly, I'm a little confused about webmapping apps. There's so many out there. Recently, we came to consider GeoServer as a possible avenue. Our front runner is still MapGuide Open Source, but the fact that it is harder to install on Debian Linux and that the Web Studio isn't fully featured is a significant limitation to us. MapServer was also considered. Please see this post and comments to learn more about our context: http://industry.slashgeo.org/article.pl?sid=07/03/07/183222

To your knowledge, do you think a java webmapping-based solution, including GeoServer, could be suitable? Would it be considered mature in terms of features? Performance is important to us - are these java apps up to par to serve maps and WM*? And finally, how would you justify choosing a java-based solution in comparison to, say, a MapGuide Open Source solution?

I understand all tools can to do job and the context and constraints are to be considered, but my hope is your comments will help us make the best choice for our organization.

Thanks a lot! :slight_smile:

Alex
--
Alexandre Leroux, M.Sc., Ing.
Environnement Canada / Environment Canada
Centre météorologique canadien / Canadian Meteorological Centre
Division de la réponse aux urgences environnementales /
Environmental Emergency Response Division
alexandre.leroux@anonymised.com

2

Well, obviously an answer on this list is probably going to be a biased response, so take it with a grain of salt.

I would definitely say GeoServer is quite mature in terms of features. It's been an open source project for over 5 years, which means that all the little fixes and improvements that make it easier have slowly gotten in. We've also got a great framework in place, which makes it much easier for us to add new features.

In terms of performance, right now we're a bit slower than MapServer, but in the next month or two we're going to have some improvements that should put us right on par in terms of raw performance. We also offer nice anti-aliasing, which slows down times a bit but makes things look a lot nicer. We've also been focusing on another angle of performance, which is doing tiled maps and caching them. We have some decent support for that, which leads to static maps performing really well, better than you could do in a fully dynamic server. And by the end of summer we should have a java based caching engine that will let you plug in to LRU and distributed caches.

If you're talking about performance just in terms of java vs. C, I wouldn't worry about it. It now comes down more to implementation than language, as the JVM is really good and gets faster every release. For example streaming GML data through WFS was 4.1 MB/s with Java 1.4, 8.8 MB/s with Java 1.5, and 10.7 MB/s on Java 1.6

We are doing a lot of performance testing these days, and I am confident that relatively soon we will be as fast if not faster than anything out there.

In terms of comparison with MapGuide open source, where GeoServer excels is implementation of open standards. We are fully compliant on WMS 1.1.1, WCS 1.0, and are the Reference Implementation for WFS 1.0 and 1.1, including WFS-T and locking. And we use the open SLD standard for all our styling stuff, and support all kinds of SLD WMS functions. Also all layers are 'standard by default' - you don't have to do any extra configuration to support the standard. GeoServer also just got really good KML support, and the next release will have GeoRSS output.

MapGuide excels at being a full stack mapping solution to quickly generate mapping sites. They have their own ajax viewer, generated by the studio apps. For GeoServer we recommend OpenLayers, which is its own project that has a ton of power. We include some demos of how to get started with it, but for designing mapping sites you do have to take a bit more initiative. But I think you also may get a bit more control, since its not auto-generated (though I could be wrong, MapGuide may have nice separation of concerns). The other thing lacking in GeoServer is a nice online styling tool - you have to make your SLD in another tool or by hand. I had thought that the Map Guide Web Studio was pretty full featured, where do you find it lacking? I fear GeoServer's web admin will be less featured, though I could be wrong, I haven't investigated MapGuide's in depth.

I hope that helps. I'd say the best thing is to just download the software and start playing with it. Good luck on figuring it all out, and let us know what you decide and why.

best regards,

Chris

Alexandre Leroux wrote:

Hi list,

I'm new to the list and subscribed to have your input about GeoServer and webmapping applications.

Honestly, I'm a little confused about webmapping apps. There's so many out there. Recently, we came to consider GeoServer as a possible avenue. Our front runner is still MapGuide Open Source, but the fact that it is harder to install on Debian Linux and that the Web Studio isn't fully featured is a significant limitation to us. MapServer was also considered. Please see this post and comments to learn more about our context: http://industry.slashgeo.org/article.pl?sid=07/03/07/183222

To your knowledge, do you think a java webmapping-based solution, including GeoServer, could be suitable? Would it be considered mature in terms of features? Performance is important to us - are these java apps up to par to serve maps and WM*? And finally, how would you justify choosing a java-based solution in comparison to, say, a MapGuide Open Source solution?

I understand all tools can to do job and the context and constraints are to be considered, but my hope is your comments will help us make the best choice for our organization.

Thanks a lot! :slight_smile:

Alex

--
Chris Holmes
The Open Planning Project
http://topp.openplans.org

3

Alexandre Leroux ha scritto:

Hi list,

To your knowledge, do you think a java webmapping-based solution, including GeoServer, could be suitable? Would it be considered mature in terms of features? Performance is important to us - are these java apps up to par to serve maps and WM*?

Just adding my 2 cents on the performance side of the equation.
If you need to serve shapefiles, we're still a bit lagging behind MapServer, but if you will store your data in Oracle or Postgis, we have
users that did test both reporting that Geoserver is faster than MapServer (if you search the archives, you'll find a user reporting
he chose Geoserver because it was 40% faster than MapServer when
connected against Oracle).

Doing your own tests is easy, why don't you grab the Geoserver 1.5.1 release, fiddle with sample data, and then try connecting to whatever
is your actual data source? :slight_smile:

Oh, and let us know how things are going.
Cheers
Andrea

4

Hi list,

Just a quick followup on my initial inquiry. So far, it seems my colleagues and I have elected to go on with GeoServer instead of other open source webmapping solutions. We're still in the process of learning how everything works and exploring, but the prototypes we easily built were convincing that GeoServer can suit very well our needs.

For the curious ones, we plan to use GeoServer in conjunction with PostGIS, TileCache and OpenLayers. We also used uDig to generate SLD symbology.

I have been positively surprised at the speed to which questions were answered and reported bugs attended to. I hope our small team will be able to help improve GeoServer and related apps in the future. Amongst the reasons behind the choice of GeoServer was how easy it is to install (we had several issues installing MGOS on Debian).

Regards,

Alex
--
Alexandre Leroux, M.Sc., Ing.
Environnement Canada / Environment Canada
Centre météorologique canadien / Canadian Meteorological Centre
Division de la réponse aux urgences environnementales /
Environmental Emergency Response Division
alexandre.leroux@anonymised.com

Chris Holmes wrote:

Well, obviously an answer on this list is probably going to be a biased response, so take it with a grain of salt.

I would definitely say GeoServer is quite mature in terms of features. It's been an open source project for over 5 years, which means that all the little fixes and improvements that make it easier have slowly gotten in. We've also got a great framework in place, which makes it much easier for us to add new features.

In terms of performance, right now we're a bit slower than MapServer, but in the next month or two we're going to have some improvements that should put us right on par in terms of raw performance. We also offer nice anti-aliasing, which slows down times a bit but makes things look a lot nicer. We've also been focusing on another angle of performance, which is doing tiled maps and caching them. We have some decent support for that, which leads to static maps performing really well, better than you could do in a fully dynamic server. And by the end of summer we should have a java based caching engine that will let you plug in to LRU and distributed caches.

If you're talking about performance just in terms of java vs. C, I wouldn't worry about it. It now comes down more to implementation than language, as the JVM is really good and gets faster every release. For example streaming GML data through WFS was 4.1 MB/s with Java 1.4, 8.8 MB/s with Java 1.5, and 10.7 MB/s on Java 1.6

We are doing a lot of performance testing these days, and I am confident that relatively soon we will be as fast if not faster than anything out there.

In terms of comparison with MapGuide open source, where GeoServer excels is implementation of open standards. We are fully compliant on WMS 1.1.1, WCS 1.0, and are the Reference Implementation for WFS 1.0 and 1.1, including WFS-T and locking. And we use the open SLD standard for all our styling stuff, and support all kinds of SLD WMS functions. Also all layers are 'standard by default' - you don't have to do any extra configuration to support the standard. GeoServer also just got really good KML support, and the next release will have GeoRSS output.

MapGuide excels at being a full stack mapping solution to quickly generate mapping sites. They have their own ajax viewer, generated by the studio apps. For GeoServer we recommend OpenLayers, which is its own project that has a ton of power. We include some demos of how to get started with it, but for designing mapping sites you do have to take a bit more initiative. But I think you also may get a bit more control, since its not auto-generated (though I could be wrong, MapGuide may have nice separation of concerns). The other thing lacking in GeoServer is a nice online styling tool - you have to make your SLD in another tool or by hand. I had thought that the Map Guide Web Studio was pretty full featured, where do you find it lacking? I fear GeoServer's web admin will be less featured, though I could be wrong, I haven't investigated MapGuide's in depth.

I hope that helps. I'd say the best thing is to just download the software and start playing with it. Good luck on figuring it all out, and let us know what you decide and why.

best regards,

Chris

Alexandre Leroux wrote:

Hi list,

I'm new to the list and subscribed to have your input about GeoServer and webmapping applications.

Honestly, I'm a little confused about webmapping apps. There's so many out there. Recently, we came to consider GeoServer as a possible avenue. Our front runner is still MapGuide Open Source, but the fact that it is harder to install on Debian Linux and that the Web Studio isn't fully featured is a significant limitation to us. MapServer was also considered. Please see this post and comments to learn more about our context: http://industry.slashgeo.org/article.pl?sid=07/03/07/183222

To your knowledge, do you think a java webmapping-based solution, including GeoServer, could be suitable? Would it be considered mature in terms of features? Performance is important to us - are these java apps up to par to serve maps and WM*? And finally, how would you justify choosing a java-based solution in comparison to, say, a MapGuide Open Source solution?

I understand all tools can to do job and the context and constraints are to be considered, but my hope is your comments will help us make the best choice for our organization.

Thanks a lot! :slight_smile:

Alex

5

Alexandre Leroux wrote:

Hi list,

Just a quick followup on my initial inquiry. So far, it seems my colleagues and I have elected to go on with GeoServer instead of other open source webmapping solutions. We're still in the process of learning how everything works and exploring, but the prototypes we easily built were convincing that GeoServer can suit very well our needs.

For the curious ones, we plan to use GeoServer in conjunction with PostGIS, TileCache and OpenLayers. We also used uDig to generate SLD symbology.

I have been positively surprised at the speed to which questions were answered and reported bugs attended to. I hope our small team will be able to help improve GeoServer and related apps in the future. Amongst the reasons behind the choice of GeoServer was how easy it is to install (we had several issues installing MGOS on Debian).

Awesome! Welcome to the community, though obviously you've already joined with some nice bug reports :slight_smile: And speaking of debian, if anyone knows anything about how to package for debian I would love to start to make GeoServer available under apt-get. I recently installed ubuntu, and GeoServer worked out of the box with the java you can apt-get. I'd love to make GeoServer easier to install on linux (or at least debian) than it is on Windows.

best regards,

Chris

Regards,

Alex

6

Alex,

I think that is a strong stack you've put together. I'm using the same stack working with the Portland Transit Authority to put up slippy maps for their bus routes, light rail, etc. The website is due to go live in the next month or two. (I'll post URLs to this list when it goes live.)

[shameless plug]
It's also the same stack I talk about in my latest book, GIS for Web Developers. <http://www.pragmaticprogrammer.com/titles/sdgis/&gt; It's available now in PDF, dead tree edition will be hitting the shelves in Q4. There is a chapter dedicated to each of the technologies you list.
[\shameless plug]

Not only are the various technologies strong on their own, but I have to echo your sentiment about the vitality and speed of the corresponding mailing lists. The communities surrounding PostGIS, OpenLayers, GeoServer, et al are quite active and always helpful.

Best of luck,
Scott Davis
scott@anonymised.com

On Jul 12, 2007, at 1:30 PM, Alexandre Leroux wrote:

Hi list,

Just a quick followup on my initial inquiry. So far, it seems my
colleagues and I have elected to go on with GeoServer instead of other
open source webmapping solutions. We're still in the process of learning
how everything works and exploring, but the prototypes we easily built
were convincing that GeoServer can suit very well our needs.

For the curious ones, we plan to use GeoServer in conjunction with
PostGIS, TileCache and OpenLayers. We also used uDig to generate SLD
symbology.

I have been positively surprised at the speed to which questions were
answered and reported bugs attended to. I hope our small team will be
able to help improve GeoServer and related apps in the future. Amongst
the reasons behind the choice of GeoServer was how easy it is to install
(we had several issues installing MGOS on Debian).

Regards,

Alex
--
Alexandre Leroux, M.Sc., Ing.
Environnement Canada / Environment Canada
Centre météorologique canadien / Canadian Meteorological Centre
Division de la réponse aux urgences environnementales /
Environmental Emergency Response Division
alexandre.leroux@anonymised.com

Chris Holmes wrote:

Well, obviously an answer on this list is probably going to be a biased
response, so take it with a grain of salt.

I would definitely say GeoServer is quite mature in terms of features.
It's been an open source project for over 5 years, which means that all
the little fixes and improvements that make it easier have slowly gotten
in. We've also got a great framework in place, which makes it much
easier for us to add new features.

In terms of performance, right now we're a bit slower than MapServer,
but in the next month or two we're going to have some improvements that
should put us right on par in terms of raw performance. We also offer
nice anti-aliasing, which slows down times a bit but makes things look a
lot nicer. We've also been focusing on another angle of performance,
which is doing tiled maps and caching them. We have some decent support
for that, which leads to static maps performing really well, better than
you could do in a fully dynamic server. And by the end of summer we
should have a java based caching engine that will let you plug in to LRU
and distributed caches.

If you're talking about performance just in terms of java vs. C, I
wouldn't worry about it. It now comes down more to implementation than
language, as the JVM is really good and gets faster every release. For
example streaming GML data through WFS was 4.1 MB/s with Java 1.4, 8.8
MB/s with Java 1.5, and 10.7 MB/s on Java 1.6

We are doing a lot of performance testing these days, and I am confident
that relatively soon we will be as fast if not faster than anything out
there.

In terms of comparison with MapGuide open source, where GeoServer excels
is implementation of open standards. We are fully compliant on WMS
1.1.1, WCS 1.0, and are the Reference Implementation for WFS 1.0 and
1.1, including WFS-T and locking. And we use the open SLD standard for
all our styling stuff, and support all kinds of SLD WMS functions. Also
all layers are 'standard by default' - you don't have to do any extra
configuration to support the standard. GeoServer also just got really
good KML support, and the next release will have GeoRSS output.

MapGuide excels at being a full stack mapping solution to quickly
generate mapping sites. They have their own ajax viewer, generated by
the studio apps. For GeoServer we recommend OpenLayers, which is its
own project that has a ton of power. We include some demos of how to
get started with it, but for designing mapping sites you do have to take
a bit more initiative. But I think you also may get a bit more control,
since its not auto-generated (though I could be wrong, MapGuide may have
nice separation of concerns). The other thing lacking in GeoServer is a
nice online styling tool - you have to make your SLD in another tool or
by hand. I had thought that the Map Guide Web Studio was pretty full
featured, where do you find it lacking? I fear GeoServer's web admin
will be less featured, though I could be wrong, I haven't investigated
MapGuide's in depth.

I hope that helps. I'd say the best thing is to just download the
software and start playing with it. Good luck on figuring it all out,
and let us know what you decide and why.

best regards,

Chris

Alexandre Leroux wrote:

Hi list,

I'm new to the list and subscribed to have your input about GeoServer
and webmapping applications.

Honestly, I'm a little confused about webmapping apps. There's so many
out there. Recently, we came to consider GeoServer as a possible
avenue. Our front runner is still MapGuide Open Source, but the fact
that it is harder to install on Debian Linux and that the Web Studio
isn't fully featured is a significant limitation to us. MapServer was
also considered. Please see this post and comments to learn more about
our context: http://industry.slashgeo.org/article.pl?sid=07/03/07/183222

To your knowledge, do you think a java webmapping-based solution,
including GeoServer, could be suitable? Would it be considered mature
in terms of features? Performance is important to us - are these java
apps up to par to serve maps and WM*? And finally, how would you
justify choosing a java-based solution in comparison to, say, a
MapGuide Open Source solution?

I understand all tools can to do job and the context and constraints
are to be considered, but my hope is your comments will help us make
the best choice for our organization.

Thanks a lot! :slight_smile:

Alex

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

7

Scott Davis ha scritto:

Alex,

I think that is a strong stack you've put together. I'm using the same stack working with the Portland Transit Authority to put up slippy maps for their bus routes, light rail, etc. The website is due to go live in the next month or two. (I'll post URLs to this list when it goes live.)

Good, always nice to have some references to point people to, especially
sites that are going to be used a lot (aka heavy production sites).

[shameless plug]
It's also the same stack I talk about in my latest book, GIS for Web Developers. <http://www.pragmaticprogrammer.com/titles/sdgis/&gt; It's available now in PDF, dead tree edition will be hitting the shelves in Q4. There is a chapter dedicated to each of the technologies you list.
[\shameless plug]

Woot, another book talking about Geoserver, this is great :slight_smile:
Cheers
Andrea