Guaranteeing Exchange availability has always been a little tricky, particularly with Exchange's single-point-of-failure-based clustering scenario. There are a huge number of companies in the market that sell products designed solely to help you increase the availability of your Exchange infrastructure. In a previous tip, I mentioned that Exchange 2007 will significantly improve on the ability for Exchange to be configured on a reliable cluster because the single point of failure problem has been eliminated, with data being replicated between two clustered nodes instead of sharing a common storage medium.
Exchange 2007 will also provide another compelling high availability feature that should make the system more manageable and easier to recover in the event of a failure: Local Continuous Replication. Local Continuous Replication (LCR) is similar to the Clustered Continuous Replication feature, but doesn't require that the Exchange server be configured in a cluster to work. Instead, LCR creates a replica of your Exchange data base on local storage and keeps it automatically up-to-date. In the event of a failure in your primary database, you may be able to immediately switch over to this backup database and continue operating.
Is it a perfect solution and will it drive the third-party data protection companies out of business? No. But, LCR is another layer of defense that may come in handy when you need to recover from minor disaster.