• Home
• Articles Archive
• Features List

• Media Pack
• Contact Us
• Subscribe

MUSCULAR THINKING…

From STORAGE Magazine Vol 9, issue 5- September 2009

AN INADEQUATE DISASTER RECOVERY STRATEGY CAN MEAN LONG DOWNTIMES, LOSS OF DATA, LOSS OF INCOME AND DETRIMENTAL IMPACT ON THE CORPORATE BRAND. ALL IN ALL, SOMETHING TO BE AVOIDED AT ALL COSTS, SAYS EDITOR BRIAN WALL

In many ways, from a storage point of view, disaster recover (DR) is a similar concept to business continuity (BC): both strategies aim to protect your data. But whereas BC should take into account how a business operates, DR is mainly concerned with storing and protecting raw data.

The data recovery methods you choose should depend on two factors, argues Andrew Brewerton, EMEA technical director at BakBone Software: the recovery point objective (RPO) - ie, how much data the organisation can afford to lose or recreate, and the recovery time objective (RTO), which is how long you have to recover the data before its absence causes business continuity problems. "If the RPO and RTO were seconds away from the point of failure, rather than hours, the DR strategy would demand that data changes are tracked as they happen and that data would be restored almost immediately," he points out.

"Today, only disk-based schemes can meet these demand. But this does not mean that other approaches to DR are not worthy of consideration or deployment when you have more flexibility. The purpose of your disaster recovery strategy, and of the hardware and software you choose to implement it, is to restore data to a known state after an incident; however, when you restore data, you should also be restoring functionality and it is this concept that differentiates DR from backup and BC." More than this, though, he adds, developing a DR strategy is also about getting your data in order before you start backing it up. "Creating an ordered approach to your data before it is backed up is more efficient, will save time, money and disk space, and ultimately will speed up data restores, so you can react more quickly to an incident."

Mike Scott, head of technology operations at Rightmove, the UK's largest property portal and a user of BakBone technology, knows all about effective DR. "BakBone NetVault: Replicator is a key component of our disaster recovery plan," he states, "delivering the convenience and peace of mind that comes from knowing that critical business and customer data can be easily accessed, in the event of a disaster." By contrast, Ian Carter, IT administrator at the University of Kent, tells of his former woes when restoring data: 'When a user requested a file restore, we would end up searching through multiple tapes to find the specific data. It was a real challenge to get that data back to the user quickly. We have a clear mandate to make sure all data can be restored with the shortest possible delay; by using NetVault's Virtual Disk Library, we can make sure we offer that service." Key to a solid DR strategy is its reliability and Carter agrees. 'We test NetVault on a regular basis. This is absolutely mandatory, if we are to be confident that we can get the university back up and running, in the event of a major catastrophe."

CONTINUITY IMPERATIVE

Crucially, business continuity, which is core to disaster recovery, demands the access and recall of data to support the restore or recovery of corporate assets. "The key to disaster recovery is the preservation and the access to information that resides within the data set, anywhere within the company," says Nathan Collins, technical director, CommVault. "The best DR scenario will encapsulate reliable data recovery with maximum cost efficiency and minimal specialised training, in order to facilitate the recovery."

The overriding value of a single recovery management platform is in its ability to deliver data protection, archiving, contentbased search, replication, resource management and reporting. The latest systems distinguish themselves, he states, by providing robust and scalable tools for tracking recovery options for thousands of systems. Recovering data from hardware vendor snapshots, tape and disk backups and, increasingly, leading third-party replication products and backup appliances, such as virtual tape library (VTL) and deduplication systems. "By using a system built around a single server and self protecting and recovering index system, only a small percentage of the index needs to be protected for disaster recovery," explains Collins. "In a DR scenario, the central server is quickly installed and re-populated with the protected index, and then the recovery of all other systems can begin immediately. The required index components can be automatically recovered as needed, in order to bring back the production systems to the required point-in-time. This significantly reduces the time to recover from DR by removing the complexity. "Without the ability to seamlessly integrate new concepts into useful data management frameworks, IT users will be forced to accept lower levels of data availability and IT departments will need to adopt more complex solutions, with loosely integrated point products. Such an approach requires complex specialist operating systems that are often outside the normal skill set of the IT organisation, inherently more difficult and certainly more time-consuming to recover in a disaster. Time is an acutely important resource when a company is faced with a disaster recovery scenario." The consequences of an inadequate strategy can be long downtimes, loss of data, loss of income and detrimental impact on the corporate brand. So there is tremendous advantage to be had by embracing a single, unified approach to backing up and managing data in all of its flavours. As Collins points out: "It not only helps meet those stringent recovery time objectives (RTOs) and recovery point objectives (RPOs) that are critical in a disaster recovery scenario, but, by seamlessly delivering simple, singular control capability, any of the IT team can preserve a company's data and recover the business from anywhere on the network, and all in record time.

RECOVERY CHALLENGES

Relying on a traditional 'end of day' back-up strategy can cause a number of potential problems for SMEs. If a disaster strikes five minutes before the next backup is due to be made (or 23 hours and 55 minutes since the previous one) all the new and/or amended data created that day will be lost. "Furthermore, it could take days to recover the previous day's data - especially if a company is relying on tape, which can be defective," says Kevin Moreau, general manager, Acronis UK & Ireland. "The other issue for companies to consider is this kind of backup strategy would only recover the file data and couldn't restore applications or systems settings. Any redeployment would therefore be a labour-intensive and timeconsuming procedure, requiring IT managers to find all original software disks and reset all system preferences by hand. These issues can cause massive business disruption and potential problems servicing customers and suppliers."

There is an alternative. Imaging software has the capability to create a master backup of all a server's contents on to a second unit (either physical or virtual) and then back up incremental changes to a company's data during the day. This means that, in the event of a disaster, the recovered data won't be more than a few minutes' old and there is no risk of media degradation. In addition, imaging software backs up all settings, applications and systems files - allowing the entire contents of a server to be recovered quickly and efficiently.

"Virtualisation technology is now within reach of SMEs through a number of simpleto- manage and cost effective packages," adds Moreau. "In its simplest form, virtualisation technology can be used to create a partition on one single server to form two distinct virtual machines. In this example, one partition of a server would run all the applications as it did before, but the other partition would be used to store a mirror image of the first machine, which could be automatically backed up by imaging software.

"Both new physical machines would run independently and the virtual partition would protect each half of the server from any viruses, data corruption or loss that occurred on the other. This means that SMEs can implement an extremely resilient disaster recovery strategy within their budget, without having to purchase additional servers. A simple reboot from the hidden partition would restore the entire server back to a known, good working condition in a matter of minutes." It's worth noting that some of the most advanced imaging software can allow SMEs to recover data to dissimilar servers - enhancing their ability to recover quickly and with fewer constraints. It also allows SMEs to consider business continuity in its positive sense, enabling them to run planned data migrations, rather than responding reactively to a failed or fullcapacity server.

"Today's imaging software and virtual operating systems can, therefore, offer SMEs a fast, agile and cost-effective data recovery and business continuity solution," Moreau concludes, "and one that's definitely worth investigating before your next physical server fails or simply needs upgrading."