For these servers, consider the following recommendations:
-
User persistent (session-based) load balancing to distribute the workload among servers. These servers can be scaled up and out.
-
Use load balancing for high availability.
-
A 64-bit Application Server can be used to scale up.
-
These servers are typically CPU-bound, which means that the operating speed is determined by the processor. Because the server is limited by CPU processing, CPU speed should be considered when choosing system hardware.
-
If multiple worker processes are running on the same machine, allow 4GB of free memory for each worker process to prevent bottlenecks.
-
Consider multiple network ports for redundancy and perimeter network deployments.
-
The Gateway Caching Server should have a large and fast cache location to increase efficiency.
-
When splitting up the server tiers, provide more resources for the Application Server because it performs most of the business logic and communication to the database and Disk Groups.
-
The Application Server, reverse proxy, and Web Server have no special storage requirements.