Covering your back up - data storage on Exchange

In a world that talks via email, there are no time boundaries and windows for planned downtime have all but disappeared. Email has become a business critical application that the whole organisation depends upon. If the email system stops working then everyone from the CEO to the receptionist knows about it.

Without a plan to avoid downtime, or at least recover quickly when it does occur, the IT department can face a lynch mob. The users are not aware of the size of the challenge faced by the IT team, nor would they care if they were. All they see is that email has stopped working and that is impacting their working day, and in most cases the business as a whole.

'Over a third of corporates are losing critical e-mail correspondence due to inflating volumes and a lack of sufficient back-up and support, according to a survey by Bloor Research. In a statement accompanying the research, senior analyst Robin Bloor said 'Given the central role e-mail now plays in business, it is remarkable that many companies do not implement the necessary level of support and back-up.'' Microscope 22/10/02

This is just part of the problem. If Exchange were just about sending short messages in text format then life would be a little easier. However, the sheer volume of attachments that are being sent and received within emails has meant that Exchange storage has become a major issue. Day in and day out we are sending and receiving jpeg's, mpeg's, PowerPoint presentations, Excel files, Word documents and many more attachments. This has resulted in Exchange being one of the fastest growing consumers of storage capacity.

Rules of engagement
For Exchange administrators to survive and escape the wrath of unforgiving email users, there are three rules to adhere to:


• Ensure that the Exchange Server is kept online


• Ensure that the data is protected


• Ensure that lost data can be recovered quickly, right down to individual emails

In order to meet these rules it is important to architect a solution that will protect the entire IT environment from the storage hardware and its data, across the network infrastructure, through the server hardware and right up to the application level, in a cost-effective and efficient manner.

Downtime avoidance
The best way to ensure quick recovery from an outage (planned or otherwise) is to avoid it in the first place. Aside from sound administrative practices, putting the correct technology in place can overcome many of the causes of unplanned downtime.

Using a logical volume management tool such as VERITAS Volume Manager will avoid running out of disk space and protect against disk failure. It will bring redundancy capabilities such as mirroring, striping and other RAID levels to any disk or combination of disks, regardless of vendor. It will allow more storage to be plugged in without the need to stop or reboot Exchange.

If the storage requirements of the application change then the administrator can dynamically reconfigure or allocate storage as the need dictates. Creating virtual pools of storage in a Storage Area Network (SAN) configuration allows much greater storage utilisation to be achieved.

Faster recovery can be achieved through allowing additional copies of the data or point in time snapshots to be made. These same snapshots, achieved through VERITAS Volume Manager's FlashSnap option, can also be used for off host backups and processing. Using this technology will eliminate a very large amount of day-to-day planned and unplanned downtime.

Greater application availability still can be achieved through clustering of the application and the server or servers. If the application fails, or the server itself fails, another machine in the cluster can take over.

VERITAS Cluster Server will allow clusters of anything from 2 to 32 nodes. Aren't we going to 64? Isn't Microsoft Cluster shipped with every Windows 2000 Advanced Server and Windows 2000 Datacenter Server? In which case what does VERITAS Cluster Server add over Microsoft Cluster? The VERITAS approach to clustering means that you don't have to have one standby server for every production server and allows high availability to be achieved with less hardware. Also it gives that added protection of cascading failover in the case of more than one server within the cluster failing.

If there is a need to protect against an entire site failure, but still have the need to keep data loss and downtime to the absolute minimum there are a couple of options. One is to simply ensure that you have copies of backup data on a secondary site, but this could mean a time consuming recovery. Not only that, but the data recovered will only be as current as the most recent backup.

The alternative is to use replication technology to ensure that all Exchange data is replicated at a secondary site. Dependent on what level of protection and recoverability is needed there is a choice of VERITAS Storage Replicator or VERITAS Volume Replicator.

For the most comprehensive solution integrating Volume Replicator with VERITAS Global Cluster Manager will allow the centralised management of multiple Exchange clusters and the ability to automate the migration and failover of clusters across geographically dispersed sites.

Recovery is everything
No matter what level of technology you deploy to avoid downtime in the first place, there is no 100% guarantee. This is why it is imperative that the basis for any Exchange data protection and storage strategy is a sound backup architecture. An effective approach to backup can ensure that all the data is protected, but more importantly, that it can be recovered. And not just an entire server, but right down to individual mailbox and even down to the recovery of an individual email.

There are a number of factors to consider when looking at backing up. Two of the most important ones are the size of your backup window and the amount of data that you have to backup. These two factors are inextricably linked but there are technologies under the data protection umbrella to enhance the core backup offerings and address these issues.

If you are just concentrating on backing up Windows server on a workgroup or departmental level then VERITAS Backup Exec for Window Servers will offer a very effective backup and recovery solution. If backing up Exchange is part of a larger heterogeneous enterprise backup and recovery strategy then VERITAS NetBackup may provide the correct solution.

Whichever solution is the most appropriate there are a number of ways to enhance the core product. There are agents to allow backup of Exchange and other applications online (including open files). VERITAS' Storage Migrator product allows old files to be migrated away from that application server. The same can be applied to email attachments. This means that there is less data on the server so it can be backed up and recovered faster.

Other options help build in centralised management to your backup and recovery environment, speed up and automate the recovery of downed server and sharing of hardware when backing up over a SAN.

So when it comes to making sure that your back is covered VERITAS has the answers. Whether it is avoiding Exchange downtime or making sure the Exchange data is protected and recoverable then speak to VERITAS.