Scalability describes the ease with which you can add or remove components from a system while maintaining system reliability. As your audience grows, you may need to add servers so that the increased demand does not overload your system. Alternatively, you may want to break up a large server system into several smaller, more specialized ones. In either case, you will have to address software and hardware scalability as separate issues.
Windows Media Services is scalable by design to support a wide range of deployments—from small Internet radio stations that have hundreds of connection requests to large-scale streaming media Web sites that generate millions of requests. The Windows Media Services snap-in allows you to administer both groups of servers and publishing points as well as single servers and publishing points.
In the context of Windows Media Services, scalability refers primarily to the addition or removal of individual servers from a system. Adding servers to a system that has been overwhelmed by an increase in connections or content can dramatically improve its performance. The number of servers required in a system is based upon the bit rate of the content, the content type, and the number of concurrent client connections.
When using multiple servers, it is important to use some form of load balancing to prevent any one server from becoming overloaded. The servers should also closely match in performance and capacity to ensure that your load balancing method is as effective as possible. The number of servers that can be combined in one Windows Media Services system is unlimited.