Selecting the right hardware and its subsequent configuration in an Enterprise GIS implementation can have a significant impact on the performance and scalability of your solutions. For performance and scalabilty guidance concerning specific infrastructure components see the links to the right.
Performance Factors
There are several common performance indicators that need to be considered across the technology architecture including:
| Category | Impact* | Considerations |
|---|---|---|
| CPU | High | See SPEC.org for CPU speed comparisons |
| WAN Uplink | High | It is important to ensure sufficient bandwidth. Typically ArcGIS Server maps network requirements vary from 2 to 8 Mb/transaction. For high volume requests a T1 (1.544 Mbps) WAN uplink is frequently not be sufficient. |
| Application Virtualization |
High | Virtualization at the application level is common for centralizing datasets with remote users with minimal performance degradation for the end user. |
| Hardware Virtualization | Low | Virtualization at the application level is common for centralizing datasets with remote users with minimal performance degradation for the end user. Virtualized hardware solutions can have a significant impact as system I/O increases. |
| LAN Network | Medium | Watch for internal chatter between components. Typically 100Mbps LAN bandwidth is sufficient for clients and 1Gbps for server network connectivity. |
| Disk Speed | Medium | Typically ArcGIS systems are CPU bound. However system with high volume requests written to a single output or job directory might be I/O bound. You may need to use faster disks, Ram disk, or Solid state disk. |
| Shared Storage | Low | Impact of NAS and concerns with regards to network throughput. |
| Load Balancer | Medium | Load balancing allows ease of scaling out the web tier. |
| Web Application Server | Medium | Difference web application server platforms can provide additional management, monitoring and scalability options. |
| Hardware Virtualization | Medium | Virtualized hardware solutions can have a significant impact as system I/O increases. |
| Operating System | Medium | Typically a software component will run faster in the environment it was natively build in. |
| Data Source | Medium |
Relational database and file system based storage attributes vary significantly. |
*Where High implies a high impact, positive or negative, on overall performance
Additional Support
ESRI Professional Services offers a variety of services in these technical areas. Please contact ESRI Professional Services.