GIS Cloud Computing

ArcGIS Server Cloud

In response to the recent announcement at ESRI’s Federal User Conference that ArcGIS Server will soon be available for an hourly license fee on Cloud infrastructure, we would like to publicly share some of our experiences with working with ArcGIS Server, and other GIS Server software, “on the Cloud.” 

We have been testing GIS Server applications (ArcGIS Server, MapServer, GeoServer, and more) on several different Cloud environments, including Amazon EC2, GoGrid, and Skygone Cloud since 2008.  We have had much success configuring, scaling, and squeezing the maximum performance out of, not just single Cloud GIS Server instances (i.e. Amazon Machine Images (AMI’s)), but also complex multi-tier environments for large, enterprise GIS operations.  

We have learned many things and are excited for the future of GIS on Cloud Infrastructure (we wouldn’t be a Cloud infrastructure and GIS Services provider ourselves if we didn’t!), but obviously the word “Cloud” can be an intimidating for people that simply want to get a GIS server up and running.  Our experience is that the amazing tools, flexibility, and business opportunities that the Cloud brings far outweigh the study time needed to understand how to operate and leverage Cloud infrastructure.

What We've Learned

One of the major lessons learned from our experiences with GIS applications in the Cloud is that the Cloud makes simply throwing more computing resources (CPU, RAM, etc.) at application performance issues so simple, that it is easy to ignore the system design methodologies of more traditional IT implementations… most of which should still be at least considered during a Cloud implementation too.  Of course, deep down, this is probably what most Cloud computing companies want, because the result of such easily deployable resources is higher customer billings.  However, the amazing throughput and user loads that Cloud infrastructure can help GIS application achieve should not be confused with an increase in application performance just because the application is deployed on a Cloud. 

Unfortunately, the performance issues that can hamper GIS Server applications are not automatically solved by throwing compute resources at the issue.  Without a proper IT system design to support your GIS Server applications, whether it be in the Cloud or in a traditional dedicated server environment, your applications can suffer performance issues and your Cloud computing billings will be larger than needed as you try to compensate. 

How We Solve It

At Skygone, in addition to professional services from our highly-experienced GIS System Designers, we are helping our Cloud customers ensure proper GIS system architecture design by offering the unique ability to create and instantly-deploy, fully operational, multi-tier (web server, application, database server, etc.) infrastructures and not just single tier Cloud instances (like an Amazon Machine Image (AMI)).  See Skygone Infrastructure Templates for more information.

Why is the ability to create instantly-scalable, multi-tier infrastructure templates so important?  Multi-tier, pre-networked infrastructure helps customers deploy, clone, and scale their entire GIS operations as a single object instead of many single Cloud instances, lowering maintenance time and errors due to human interaction. 

Managing Multiple Clouds

However, we are realistic to the fact that Skygone Cloud will not be a perfect fit for every application or customer location; therefore, soon we will be releasing our Multi-Cloud control panel that will enable our customers to deploy their applications and data on many different Cloud infrastructures, including their own Private Clouds that they have implemented on-premise behind their firewall, from a single object oriented control panel.  Very exciting!