Planning for availability and performance

When planning a Microsoft Provisioning System deployment, it is important to consider how to optimize the availability and performance of your servers.

Many variables affect the performance of Microsoft Provisioning System. Factors such as the capacity and speed of the server hardware on which Microsoft Provisioning System is deployed and the network infrastructure (server interconnections, the use of load balancers and switches, and physical layer network congestion) can affect performance as experienced by users. It is also important to consider the bandwidth of the network connection to ensure that the throughput of the system can be handled.

To ensure a baseline level of availability, use backup systems and ensure that spare parts are readily available. It is also important to provide fast drive access and a fault-tolerant hardware configuration. Providing multiple logical drives or drive arrays to partition the components improves performance. Depending on the level of reliability you need, you might want to use the different RAID technologies to determine the most reliable configuration, for example, RAID 0+1.

The following approaches can help optimize network performance:

In addition, depending on the size of your deployment and the projected number of users, you can implement load balancing, clustering, and failover to help reduce interruptions in service and ensure the high performance of your system. Besides improving performance, using these techniques also makes it easier to add additional servers when you need to scale up your deployment.

By implementing failover between clustered servers, you ensure that requests continue to be processed even when a server stops functioning. By combining load balancing with clustering, you gain the added benefit of dividing incoming requests between redundant servers. This improves performance by reducing the time it takes to process requests.

For specific information about how to set up clustering, load balancing, and failover for the servers in your deployment of Microsoft Provisioning System, see Implementing redundant servers.