The main subjects of this chapter are as follows:
Determining availability requirements
Implementing fault tolerance and redundancy within your environment
Implementing redundancy for Hub Transport servers, Client Access servers, Unified Messaging, Mailbox servers, and Edge Transport servers
Implementing message hygiene
The term high availability can mean many things. The availability of a system is usually measured in increments of time, and a system is typically considered to be highly available if it is accessible at least 99 percent of the time. However, even a system that is available 99 percent of the time may not be sufficient for some situations. That is why defining availability requirements in a Service Level Agreement (SLA) is so important.
An SLA establishes the required availability of a system and how availability will be measured. (For example, would an Exchange server that is available to half of your users considered available or unavailable?) When you design an Exchange Server 2007 system, you must determine your availability requirements from the SLA and base your fault tolerance and redundancy plan upon that information.
Availability is sometimes discussed in terms the number of 9s offered. If someone says their system offers "five 9ss" of availability, ...