RE: [OSGeo-Discuss] Re: Comparision between MapServer/OpenLayers and ESRI ArcIMS
I think that it's generally less fear of the unknown or job security than it is the cost of adding complexity to what is often an already over-extended support load. In many cases it just makes sense to spend $1000 for a server OS that doesn't require additional training, is easy to get qualified techs for, and "just works" with the existing systems. It doesn't matter how easy Linux is; it's one more thing to keep track of and one more thing to go wrong. If you want to "win" the open source battle at small organisations that don't already have OS operating system tendencies, focus on the application level where you can make a strong business case on a feature-by-feature level, and with additional arguments about truly open data being more sustainable and less risky. Personally I think that an "open source or bust" attitude is not very pragmatic. "Sell" open source software where it is the best tool for the job, but pick your battles. Jason -Original Message- From: Alex Mandel Sent: Friday, May 29, 2009 4:25 PM To: OSGeo Discussions Subject: Re: [OSGeo-Discuss] Re: Comparision between MapServer/OpenLayers and ESRI ArcIMS That would be fear of the unknown(non gui) and job security at work. Wouldn't want someone else in the org who knows more about running servers. Maybe you can get them to throw a bone to demo something on a virtual machine hosted elsewhere(Amazon) just to show how easy it is. Welcome to the land of small to medium government agencies, etc. The best thing here is showing examples from equivalent groups, of which there are plenty online now. ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Re: Comparision between MapServer/OpenLayers and ESRI ArcIMS
Julia Harrell wrote: > >> Have you considered whether surplused hardware from within your organization >> could be used for some of this? >> >> It may not suffice for a windows server, and i did not see if you already >> have linux boxes. >> > > Surplus would work for a dev/test box, if/when you can manage to get your > hands on one. But really, by the time anything gets surplused around here, it > is completely worn out and dang near worthless. Unfortunately, the 'IT > Overlords' also have some bizarre squirrelly aversion to Linux and refuse to > allow it on any production GIS servers. That would be fear of the unknown(non gui) and job security at work. Wouldn't want someone else in the org who knows more about running servers. Maybe you can get them to throw a bone to demo something on a virtual machine hosted elsewhere(Amazon) just to show how easy it is. Welcome to the land of small to medium government agencies, etc. The best thing here is showing examples from equivalent groups, of which there are plenty online now. Alex ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
[OSGeo-Discuss] Re: Comparision between MapServer/OpenLayers and ESRI ArcIMS
> >Have you considered whether surplused hardware from within your organization >could be used for some of this? > >It may not suffice for a windows server, and i did not see if you already have >linux boxes. > Surplus would work for a dev/test box, if/when you can manage to get your hands on one. But really, by the time anything gets surplused around here, it is completely worn out and dang near worthless. Unfortunately, the 'IT Overlords' also have some bizarre squirrelly aversion to Linux and refuse to allow it on any production GIS servers. ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
[OSGeo-Discuss] Re: Comparision between MapServer/OpenLayers and ESRI ArcIMS
Hi Bill In addition to the excellent advice others have already given you, be sure to consider the extra hardware infrastructure that will be necessary to support this much 3D data, on-the-fly geoprocessing, and large numbers of concurrent users. If your client chooses an ESRI solution, it will take a lot more than they probably counted on purchasing to get a system that will perform adequately. ArcGIS Server is a real resource hog. Check out Dave Peter's new "Building a GIS" book. Not that I am a fan of the AGS software at all, but Dave is a really smart guy and he's put a huge amount of work into designing and fine tuning his "Capacity Planning Tool", which is now available. One used to have to go take the ESRI System Design Class in order to get a copy of the tool. I took it a couple years ago from Dave, and can attest to his thoroughness in building the CPT. You can download updated versions of the CPT from the website now. http://gis.esri.com/esripress/display/index.cfm?fuseaction=display&websiteID=141&moduleID=42 The CPT is pretty complex and does take some training to use properly, so unless you're going to take the system design class, you probably ought to get the book. It would be great if, one day, a similar tool was available for Mapserver/Geoserver/PostGIS systems (hint, hint listserv experts). Unfortunately, I'm just not knowledgeable enough about the inner guts of the OpenGIS softwares yet to be able to convert all the formulas in Dave's spreadsheet to something more appropriate for an OpenGIS system. I was a little surprised to see just how *much* server hardware and network bandwidth it would *really* take to implement a usable ESRI server-based GIS system for my organization, which has offices all over the state and crappy bandwidth in a lot of the more remote places, not to mention the uncountable anonymous public users. Right now we've got AGS running on a single server, and it is dog slow and needs a lot of babying. When I told my "IT overlords" that I'd need at least 2 more load balanced GIS processing servers and a separate production GIS database server to speed things up to an acceptable level, well, they just laughed at me. Mapserver, PostGIS and Geoserver all run "OK, but not blazingly fast" on a single, older shared GIS server for me. AGS gets the other, newer GIS server all to itself - and is still a complete pig. Sure, the OpenGIS software would be a lot faster if they each had their own dedicated box, but in this budget climate, extra funding for more hardware is just not possible... Hope this helps, Julia Harrell ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss