Maximizing availability and performance of your enterprise Microsoft Exchange environment with an advanced network-based solution.

People expect Microsoft Exchange to always be up and running. Staff must dependably send and receive email and wireless messages, refer to records and attachments, and access calendars and contact information. Continuity of service is paramount--any data unavailability can cause serious damage to your enterprise's operations and bottom line.

Microsoft Exchange is specifically designed to deliver high-volume collaboration capabilities and fast transaction rates. Yet because Exchange is so widely used, its storage can be especially challenging to manage effectively.

As message volumes and file attachments consume gigabytes of disk space, Exchange administrators are left struggling with this onslaught of data. Backup times increase relentlessly, compounded by regulatory compliance, where email messages must often be stored for lengthy periods of time. Recovery must always be fast and accurate, even after disaster.

How to better handle this situation? Today's IT professionals have a wealth of solutions at their fingertips to help implement best practices for optimizing Exchange environments. The key is to make the right choice--a comprehensive solution that delivers the entire gamut of storage services is needed so that you are not left with niche solutions, adding more work and frustration instead of alleviating it.

In today's 24X7 business environment, to maximize availability and business productivity, a practical storage solution for messaging/collaboration data needs to deliver advanced storage services that protect data and improve its accessibility.

Services should include:

* Disk mirroring for data redundancy 

* Multipathing for protection against network node failures

* Remote data replication for disaster recovery

* Point-in-time snapshots for rapid, granular, non-disruptive recovery of individual mailboxes and entire data stores

* Impactless backup that can be completed within the backup window and doesn't burden Exchange servers.

Simplify, Centralize Storage Infrastructure and Management

Exchange administrators cite capacity management as their biggest problem. Their Exchange environment can host hundreds or thousands of active mailboxes, multiple servers, and terabytes of constantly growing email storage capacity.

Because enterprises typically grow in a piecemeal fashion, the result is often disconnected and/or duplicated infrastructure (e.g., surplus Exchange servers and storage arrays); unnecessarily complex processes; and excessive administrative overhead.

Another problem for administrators is Exchange servers running out of disk space. There's constant demand for more capacity to store emails/attachments--and to store them longer to comply with regulations such SEC 17a-3 and SEC 17a-4--which can lead to last minute purchasing of disk, manual reconfiguration of physical disk resources, and wasted storage capacity due to overprovisioning.

A comprehensive solution should offer a capacity-on-demand storage service that automatically prevents out-of-space conditions by monitoring disk space consumption and providing proactive, just-in-time capacity provisioning, as well as using virtualization to leverage and repurpose existing storage and know-how. In this storage model, capacity management is easier because all disk resources are joined into a "storage pool."

Exchange administrators can dynamically carve and provision capacity from this storage pool on an as-needed basis, with just a few mouse clicks at an easy-to-use, centralized console. The underlying disk interfaces are hidden, and all storage provisioning and storage services for an unlimited number of heterogeneous application and file servers can be controlled from this console.

Achieve High Availability 

Strong business continuity tools can enhance a typical Exchange environment. Even if one or more niche data protection solutions have been implemented, chances are they are only providing a portion of the necessary protection. The result is inevitably inadequate protection from disk-, cabinet-, and network-level failures that sabotage data availability.

To protect against disk failure, look for a solution with synchronous and asynchronous mirroring capabilities to create redundant data sets. This way, a disk containing Exchange data can be mirrored to a second disk, which may reside on the same or on a different storage array of a different vendor/type/interface, providing a layer of cabinet redundancy over and above the RAID redundancy at the disk drive level. The disk subsystems themselves can be located in different locations to protect against a localized disaster.

Network-level failures can be averted by deploying failover and multipathing services that provide Exchange servers with multiple paths to storage, with server traffic intelligently rerouted to an available path for business continuity.

Disaster Recovery That Works

It is crucial to choose a storage infrastructure solution that delivers a remote replication service providing automated off-site data protection. Administrators should be able to specify a variety of policies to control the replication process, giving them a granular and flexible mechanism for keeping an extra set of data off-site for rapid recovery.

Choose a solution enabling data replication over any existing MAN or WAN network infrastructure without the need for Fibre Channel-to-IP converter boxes. The solution should provide support for tiered storage on both the production and DR side so that the source and target storage hardware need not be the same, allowing for the use of cost-effective disk at the DR center. Data consistency is of utmost importance. What good is a replica if you can't recover from it? A solution with replication utilizing snapshot images should provide data with point-in-time integrity, working in conjunction with an Exchange-specific snapshot agent, which coordinates the replication sequence with the Exchange server and with the operating system. Just prior to replication, the snapshot agent is notified, making use of Exchange APIs so that snapshots are taken with highest possible degree of data consistency, meaning that all transactions are complete and in order. This ensures that the replicated data is usable.

Ideally, the replication service should perform an initial full replication, with subsequent replication sending only changed data (deltas), while also allowing for a means of flexible scheduling so that replication can occur during off-peak hours.

Really Rapid Restore

Even with mirroring and replication, data loss can still occur from soft errors, where data is destroyed, even while the hardware is still up and running, from a virus, a malicious attacker, human error, or a corrupted message store. If soft errors persist undetected, they can develop into a rolling disaster, which becomes impossible to recover from, as the data damage gets progressively worse.

Hardware protection schemes like mirroring do not protect against soft errors. An inadvertent file deletion means the file is gone from both sides of the mirror. Once the data is lost, you can't go back in time to get it. Or can you?

Yes. An advanced storage solution offers continuous disk journaling using point-in-time delta snapshots of a data disk that can be scheduled at specific time intervals or high water marks of new data changes. When needed, the data can be "rolled back" to an earlier point in time with a few clicks in the management console. If the Exchange data store is corrupted or deleted, it can be recovered entirely from disk in a matter of minutes.

If recovering a single mailbox, the solution should provide delta-based snapshots mountable as an independent readable/writable drive, without having to roll back the entire volume. While the live volume is still running, the Exchange administrator can "read" an earlier version of the database, assign it to a standby server, and retrieve a deleted mailbox or mail messages. This feature alleviates "brick-level" Exchange backups to tape, a process both time consuming and taxing to Exchange server performance.

Better Backup

Shrinking backup windows with larger backup requirements are making traditional methods of tape backup less realistic. Yet, for many enterprises, using tape to back up Exchange data stores will not go away anytime soon. There are solutions that can make it far less arduous. These zero-impact backup acceleration and consolidation solutions ease tape backup operations by taking the workload off the Exchange servers. This means data moves to tape fast, at full drive speeds, as much as 100 gigabytes per hour per LTO tape drive.

With zero-impact backup, tape drives/libraries continue to be connected to your existing dedicated backup server--leveraging the environment you already know--except that file or block-level backups are now dramatically accelerated and can occur at any time without impacting Exchange, other applications, or end-users. The backup works in conjunction with a delta-snapshot service (which should fully integrate with the snapshot agent for Exchange, so that tape backups will have data consistency), enabling the backup software to back up point-in-time snapshots of data directly from SAN storage, offloading backup processing from application servers and the LAN, and eliminating the backup window.

If you want to combine the best of both worlds, replicate data off-site, and deploy your new backup model at the off-site location, so that in essence your tapes are already out of the building before you even start the backup process.

A network storage solution for Exchange is available. Today, collaborative applications--particularly messaging--are the critical force driving business, all across the enterprise. With an advanced network storage solution for Exchange, you can optimize your Exchange investment to achieve the performance you want and deserve--today.

Irving Moy is director of product management at FalconStor Software, Inc., Melville, NY

www.falconstor.com