Solution provider takeaway: Solution providers can improve their customers' Exchange data protection by looking beyond snapshots to CDP and near-CDP products.
Most customers cannot live without Exchange -- it's increasingly critical to business communications -- but when's the last time you looked at how you're helping customers protect and recover their Exchange environments?
If your customers can't recover their Exchange environment within minutes and with minimal downtime or data loss, they don't have a state-of-the-art solution. The silver lining, of course, is that this presents an Exchange data protection opportunity for you, and that opportunity can be found in
In some organizations, email systems are more important than a dial tone. (Have you ever reached someone's voicemail and been instructed to send an email rather than leave a message?) Beyond email and attachments to messages, Exchange, like other collaboration systems, also contains calendar and contact information. In some organizations, email is used as the file server and the Outlook client is the CRM tool. Exchange stores are growing rapidly, and, due to increasing legal regulations, it's becoming critical to retain the information in those stores.
Rapid full recovery of the entire environment is vital to your customers, but so is recovery of single messages or mailboxes, and many customers want to incorporate some form of email retention and discovery into their Exchange data protection process.
Ironically, despite its importance, Exchange is typically protected only once per day by a backup application with an Exchange module that may be able to do a brick-level backup. Considering the thousands upon thousands of messages these systems handle per hour, this simply is not granular enough. There should be multiple protection instances throughout the day; the technology to do this already exists.
The problem with snapshots
Shared storage with snapshot capability can get your customer partway to where they should be in their Exchange data protection process, but it leaves gaps. Most storage suppliers can integrate their snapshot technology directly into Exchange, allowing them to put Exchange in a backup state so a valid snapshot can be taken. Then in the event of an Exchange database corruption or virus corruption, the environment can be rolled back to the latest snapshot fairly quickly. This snapshot can also be mounted to an alternate server for backups or even to a standby Exchange server for message extraction.
The integration of the snapshot to Exchange is a nice step forward, but it leaves much to be desired. The biggest problem with snapshots is that if the volume fails in a catastrophic fashion, all the snapshots and your Exchange history go with it. This means you need to recover from a backup disk or tape. Even with disk, all that data has to be transported across the wire.
And in some cases, snapshots take a toll on the performance of the area controller. Each snapshot has to be tracked and managed by the storage controller, burdening it with extra workload. Snapshots are also expensive. The space allocated to snapshots consumes the same or a similar area of disk as the primary -- always on the same manufacturer's expensive hardware. It's like storing your backup data on primary storage. As a result, most customers keep only a finite number of snapshots available; rarely do they retain more than a week's worth. This certainly makes snapshots unviable for most forms of discovery requests.
Beyond performance and cost issues, there's the issue of inadequate protection: Exchange snapshots only protect the environment if there is an application data failure. Because there are a limited number of snapshots being taken -- every few hours is rare, once or twice per day is more likely -- a lot of messages have to be reread into the Exchange store, a time-consuming step in the recovery process. In addition, message extraction is challenging. Message extraction via a standby server requires that an Exchange server and license sit idle much of the time, only being used for extraction. Most customers cannot or will not do this. And if a customer is willing to dedicate a standby server for extraction, it also has to commit a fair amount of primary disk resources to the retention of this data.
A very small number of snapshot vendors, NetApp for example, have made it possible to actually peer into the snapshot and perform message extraction. The challenge is that, even with this enhanced functionality, it is difficult to justify long-term retention of Exchange snapshots. The cost of the storage and the dependency on the original are too great to be used for a long-term storage repository.
While snapshots are ideal for recovery from data corruption, customers need a higher level of recoverability.
Continuous or near-continuous data protection
To fill this void, companies like Asempra, Inmage and SyncSort have created continuous or near-continuous data protection solutions. These products essentially capture data more frequently than snapshots to a separate storage controller. This alleviates two shortcomings of snapshots: the cost of disk and the performance penalty for retaining many snapshots. Also, these standalone copies are independent of the original data set; if the primary storage controller fails, the data and all its historical snapshots are still available.
In addition, because the targets are separate storage areas, the Exchange server can mount these standalone images directly. So in the event of a data corruption, instead of having to go through the time-consuming recovery process, you can merely redirect the Exchange server to mount the image on the secondary target. Because of the frequency of protection, only a limited number of transaction logs need to be replayed.
As storage suppliers have done with snapshots, most CDP or near-CDP vendors have incorporated Exchange integration into their products to ensure that the Exchange environment is in the correct state, but as with snapshots, these systems have limited tools for extraction of specific messages. SyncSort provides the ability to search and extract specific mailboxes or messages from the secondary Exchange copy. Message- or mailbox-level extraction, coupled with inexpensive long-term storage on secondary hardware, makes a viable solution for handling service discovery requests.
SyncSort takes this a step further by integrating the near-CDP function into the backup process. Most CDP or near-CDP solutions involve a process that's external to the backup application. This adds an extra burden on your customers because the secondary copy has to be manually backed up, separate jobs need to be created and separate GUIs need to be monitored. With integration to the backup process, on the other hand, older images can be moved to tape as the secondary disk target begins to run out of capacity. The combination of near-CDP functionality, message-level extraction and integration into the backup process make this a formidable combination and a big problem solver for resellers.
One final point: Most of the companies that can provide enhanced Exchange data protection have excellent channel programs and, with the exception of NetApp, have limited direct sales teams for the channel to compete with.
About the author
George Crump is president and founder of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. With 25 years of experience designing storage solutions for data centers across the United States, he has seen the birth of such technologies as RAID, NAS and SAN. Prior to founding Storage Switzerland, George was chief technology officer at one of the nation's largest storage integrators, where he was in charge of technology testing, integration and product selection. Find Storage Switzerland's disclosure statement here.
This was first published in November 2008