To keep this document limited to a defined scope, all information provided is based upon a ‘typical’ deployment of Dekho. In reality, no two Dekho implementations (except your Test and Production) look alike. However, it will assist in providing practical information on the support capacity (e.g. number of supported users) for scalability of a ‘typical Dekho deployment’.
Figure 1 - A 'Typical' deployment for a small to medium scope deployments.
In an organisation requiring a small to medium scope deployment of Dekho and ArcGIS for Server, a typical deployment may resemble a solution stack similar to figure 1 and contain the following tiers:
• ArcGIS Server Site only contains one GIS Server.
o Dekho and DASM components installed on an Apache Tomcat on the same machine as ArcGIS for Server 10.1
o A physical 64bit Windows Server 2008 R2 Operating System.
o 8GB RAM and a 4 core mid-range CPU (~2.8 GHz)
o Configuration Store and Server directories remain on same machine, but if deployment is to scale to multiple GIS Server machines, these would be on a separate UNC accessible share.
• ArcGIS Web Adapter installed within IIS on a separate web server tier.
• Dekho Database on a separate tier in either Oracle or SQL Server.
It is important to note that this ‘typical deployment’ is not a reference for best practice, as specific deployments vary based on business and user needs of Dekho and ArcGIS for Server. Please refer to the Infrastructure Performance Considerations and System Design Strategies wiki for assistance in designing a Server infrastructure that would meet your business requirements.
The next 5 sections of the document describe the 4 key factors which could impact performance in a ‘typical’ deployment scenario. Those key factors are:
• Authoring Dekho Maps
• Configuring the Dekho Database
• Designing Queries
• Bookmark Management
• Configuring Tomcat Memory
The document will also touch on how variations to this deployment can impact performance.