Adjustments to the following components of your server can improve performance on servers whose primary role is to support user mailboxes:
In most customer configurations, the disk subsystem and memory are the first resources to become overloaded in a Microsoft Exchange Server computer. One way to improve disk performance is to put the information store log files onto their own physical disks, apart from other files in the system. Because the logs are always written sequentially, having dedicated disks means that the disk heads won't be used for other activities and will always be in the right place for writing the next file. This improves performance because seeking is the slowest thing a disk does.
In the case of the information store database files, which are randomly accessed, the best thing you can do is put the disks in a RAID array for maximum random access performance.
You also can improve performance by using good disk controllers. Caching controllers increases your disk's performance for reading and, if the controller has battery backups, improves write performance as well. If the controller has a write cache, it must have a battery backup. Problems with cache writes (for example, in the event of a power outage) cause the computer to lose data. The batteries are on the card to keep data on the card backed up. If the data is cached and there is a power outage, the data is saved. Make sure you don't put so many disks on one controller that the controller becomes a bottleneck.
It is important to look carefully at the reliability of your disks because users depend on them. It might be wise to mirror the log files and to use RAID 5 on the database.
Summary of Recommendations
Two other resources that can become overloaded are memory and network capacity. Insufficient memory in the system causes the system to thrash, that is, page excessively. Normally, you do not notice paging at all; so if the system is paging excessively, it is obvious. Fortunately, it is easy to add memory, which corrects the problem. You can also fix thrashing by removing the load on the server.
If there is not enough network capacity (bandwidth) between the users and the server, your efforts to provide a responsive server will be unsuccessful. The network becomes your bottleneck.
When you are using a local area network (LAN), you have the option of replacing a 10 MB Ethernet with a 100 MB Ethernet network.
When you are using a wide area network (WAN), it is important to remember two things:
Summary of Recommendations
After the disks and memory, the resource most likely to become overloaded is the CPU.
If the server supports only users and does not run connectors and other processes, there are limits to how much work you can offload by adding CPUs. Therefore, use the fastest possible CPU. However, if the server performs other processing functions, there is a definite benefit in adding more processors. For example, it might be useful to add processors if your server supports connectors or if it doubles as a messaging hub, a domain controller, or a Microsoft SQL Server database
To make the best use of your processors, use a large L2 cache. This keeps the processor from waiting while data is retrieved from main memory. Also, in a multiprocessor server, the cache keeps load off the system bus, which improves your overall throughput.
Summary of Recommendations