Email and messaging systems are the lifeblood of business, so when a company’s Exchange server goes down, it can have devastating effects on business productivity. Even losing a single email can have dire consequences. Protecting the Microsoft Exchange environment is a business necessity. Given the mission critical nature of Exchange, I have focused lately on writing about Microsoft Exchange high-availability and data recovery for consistent databases.
In consideration of the various challenges associated with protecting the Microsoft Exchange environment, I first wrote about LLR and then talked to Utah State University about some design decisions and operational specifics. The buying behavior is clear; customers want better email availability. Microsoft does a great job of delivering good tools and solutions for Microsoft Exchange availability, but it is an operating platform and not a full-fledged data protection solution. As an operating platform it set’s the stage for applications designed to address business users and operational efficiencies.
Microsoft Exchange has many operational elements that require management, reporting and testing. The primary function of Exchange, since the introduction of Microsoft Exchange 2000, is an information storage system. During the initial design of Exchange, many companies consider the option to cluster systems. If you are thinking about clustering for highly available data, then you should also consider Continuous Data Protection (CDP) solutions such as InMage’s Scout. CDP can provide high-availability using the same techniques employed by Microsoft’s latest Service Pack for Microsoft Exchange Server 2007 (SP1).
Standby Continuous Replication (SCR) was introduced in SP1 and uses the same log shipping functionality in Microsoft Exchange Server 2007 Local Continuous Replication. SCR introduces new terms like sources and targets. On the surface it creates feelings of joy and sense that you can spend entire evenings and weekends with family and friends. Moreover, SCR is available on the Standard Edition of Microsoft Exchange 2007 SP1.
SCR is primarily designed for whole server recoveries. SCR wouldn’t give you the option to recover individual stores within a storage group. What an administrator really needs is the ability to provide disaster recovery and data recovery, in a single application and administrative console. Moreover, SCR does not change configuration options for LLR. Since, LLR is still enabled; you can suffer up to six megabytes of losses in a storage group, and never know about it.
Daniel Muller of Utah State University said that “Clustering, LLR, LCR, etc wasn’t even an option for us.” He went on to say “Having a single system to administer, rather than a collection of different systems made sense to us.” Daniel is talking about total data availability, not just failover and failback. Utah State University ultimately selected InMage’s Scout CDP solution for Exchange offering a centralized management console to protect day-to-day data as well as disaster recovery failover to an entirely different site.
Scalability which is critical in high-availability is achievable with host-offloaded CDP solutions such as InMage’s Scout. By offloading the CDP processing to an out-of-band appliance, a CDP solution can easily support databases reaching 100’s gigabytes in size, with thousands of users. For example, if you wanted to manage a limit on your databases of about 100 gigabytes and mailbox sizes of 2 gigabytes, you could have up to 50 users per store. That’s a bit extreme for the average user at the average company, perhaps 250 megabyte mailbox giving you up to 400 users per store. In either case, CDP software is proven in those environments.
Regardless of what options Microsoft delivers for high-availability, third party CDP software for Microsoft Exchange will maintain an edge in supporting a combined solution for data recovery and disaster recovery. Give InMage’s Scout CDP solution a test run, it only takes a few minutes to setup in testing and a few days to get a product environment up and running!