The Army Mapper system provides web hosting and data editing capabilities for the Army's Installation Geospatial Information & Services (IGI&S) program. It is an enterprise model, so the data and the publication tier all sit central with editors around the globe able to connect to and edit the data via Citrix. All data is viewable in a CAC-restricted web session.
We recently deployed the updated, more secure, Army Mapper portal, which is the front end for all Army Mapper functions. When we first began working with the client, all of the servers were running Windows 2003. Now all but the ArcGIS Server and database tiers have been virtualized and are all running Windows server 2008, so we've gone through a big upgrade.
- Replaced the JBoss code with .NET server side backend. The gain here is increased security and application development flexibility.
- Implemented a repeatable build process for all code.
- Solidified a testing & deployment policy. Our job was to implement the testing and deployment process policy and help the other stakeholders to adhere to it. The end result of that policy is a more stable deployment process, increased chances of successful deployments, and ensuring that deploying code for one application won't break the other applications that run on the same system.
- Prepared the Army Mapper enterprise system for a more cost-effective migration to the cloud
- Reduced the number of commercial off the shelf software components
- Implemented testing software for more in-depth user acceptance testing and testing management - end result: enabling direct feedback from users
- Resurrected the failover (Continuity of Operations or 'COOP') infrastructure to better ensure required operational 'up-times'
- Converted aging physical servers with virtual equivalents
- Implemented Splunk Enterprise to capture and analyze thousands of system's events for rapid monitoring and identification of potential issues
In the end, the client benefits from reduced complexity and reduction in costs. Tying back to cloud deployment - if we reduce complexity and, through the use of Splunk, have a better understanding of the system performance and we don't need to over purchase, then it's a greater value to the client. When they do migrate to the cloud, they don't have to go system-by-system. We have roles identified and the system requirements that are needed to meet those intended roles, all while reducing system overhead. So the system overhead goes down and the cost goes down when you migrate.