Preventative Care: Backup and Disaster Recovery

For Mohammed Ateeque, the thought of his company's systems going down conjures images of a clock ticking loudly and urgently.

The network administrator and his colleagues at Roanoke Trade Services, a firm specializing in supply chain–related insurance services based in Schaumburg, Ill., know that backing up data so that it can be restored during an outage is just a sliver of what's involved in being prepared to operate during and after a disaster. Equally important is how quickly data becomes available to employees and, even more critically, to customers.

"Speed is essential for us," Ateeque says. "Time of recovery translates directly into money. We can lose $25,000 a day in business if one of our small offices, like the one in Miami, goes down. A larger outage can cost the company hundreds of thousands of dollars in business and operational costs."

Roanoke Trade is one of a growing number of small to midsize businesses developing high-availability backup strategies because they recognize that swift recovery from a disaster can be critical to the success — maybe even the survival — of their companies.

Although plenty of effective technology exists, tight budgets and small staffs often complicate backup and disaster recovery planning for small businesses. IT managers say it's essential that the strategy fit the specific needs and resources of the individual business.

Enterprise Partners Push the Importance of Backup and Disaster Recovery Plans

Disaster recovery and high-availability backup became top priorities at Roanoke Trade in 2008 when the company was acquired by global insurance leader Münchener Rückversicherungs-Gesellschaft (Munich Re), and became an affiliate of Watkins Underwriter's at Lloyd's of London. The larger organization encouraged its new acquisition to expand its DR and backup.

As part of that push, Sudewa Wanigasinghe, the company's IT vice president, assembled a team — Network Administrator Vukasin Koprivica, Database Administrator Suganthi Inbaraj and Ateeque — to help him deploy a new backup strategy. Last July, they began implementing an EMC Celerra NS-120 unified storage system, which integrates EMC's Clariion CX4 storage networking for high availability. The NS-120 supports an Oracle 11g RAC cluster database, and the two technologies together form the core of the online backup plan for Roanoke Trade's centralized applications. Oracle's native Real Applications Clusters feature adds to both availability and redundancy pieces of the requirements, says Inbaraj.

"There are three basic reasons for the new strategy — availability, reliability and redundancy," says Wanigasinghe. "We've moved to a model where we and our customers need 24x7 access to apps."

The company uses Symantec Backup Exec software to support disk-to-disk-to-tape backup every 24 hours; but the NS-120 enables snapshots of the database, allowing for fast data recovery to a point in time within a half-hour of an outage.

Time to recovery was the main criterion for Roanoke Trade's choice of the EMC system, but other factors, such as compatibility with the company's Cisco Systems switching infrastructure and with its use of VMware virtualization tools, figured prominently as well, Ateeque says.

"Besides the speed, there's everything a small business needs to work with bigger businesses," he says. "There's also the reliability of knowing that a company like EMC is always going to be around."

Contemplating the Cloud

Just as Roanoke Trade is absorbing the priorities and practices of its much larger parent company, growth is nudging Freedman Seating in Chicago toward a reassessment of its systems infrastructure, including backup and DR processes, says IT Director Ross Rutherford. He has been researching technologies that will move the maker of vehicle seats away from tape backup to more scalable options offering high availability.

"As our company approaches $100 million in sales a year, we're moving from a small to a midtier architecture," Rutherford says. "One of my immediate goals is to formulate a new backup strategy and disaster recovery plan to match our new needs. For now, we're still in the tape world. We use Backup Exec, and I'm happy with its performance; but as we get bigger I'm looking for ways to streamline backup and recovery."

Complicating the planning process is Freedman Seating's reliance on Microsoft Exchange and SharePoint for messaging and collaboration. After conversations with many vendors and product reviews, Rutherford says he has been unable to find tools that will back up the information in these key applications and provide the recovery times and granularity that Freedman Seating needs. The company's increasing use of virtualization is another factor that requires consideration, he says, because many systems and tools that work on bare metal don't translate well to a virtual environment.

Rutherford's research is increasingly leading him toward a decision to outsource some of Freedman Seating's infrastructure and applications, including backup and disaster recovery, to the cloud.

"I've thought about going to collocation and still doing the backup myself, but there is too much development in cloud services to ignore," he says. "The decision is all about keeping speed and time to recovery up, but you have to think about how much control you're willing to give up. In the cloud, you're putting your livelihood in someone else's hands."

>>Visit our cloud computing resource center.

Data Center Restoration in 30 Minutes

At Smart Start in Irving, Texas, CIO Thomas Allison has implemented a collocation strategy that allows him to maintain control of the company's data while cutting the time for a full restoration of the data center from 24 hours to 30 minutes.

"We have been using disk-to-tape rather than tape alone for recovery speed reasons, and we'll probably continue to use tape for archiving because it's cheap," says Allison, whose company distributes and monitors ignition interlock systems to prevent drunk driving. "But in the past year, we've added true disaster recovery capability by using virtualization and a collocation facility."

Smart Start recently moved from Fibre Channel SANs to virtual SAN technology, which offers failover replication at a comparable cost. A virtual SAN in the Smart Start data center replicates all corporate data on three servers in the collocation facility, Allison says.

"We'd have to run a little slower if we had to failover, but the point is that we've got these capabilities that were once only within reach of big enterprises," he says. "With the collocation, we have two copies of every piece of data, one for the tape archive and one for high-availability disaster recovery from the other facility."

Going with What Works: Tape Backup

For Brian Mintel, IT manager at collection agency Frost-Arnett in Nashville, backup is too important to use anything but a conservative approach.

Using tape and CA ARCserve software, each of Frost-Arnett's four offices are fully backed up every night, with copies of the tapes placed in a fire safe in the main data center and in a bank safety deposit box. At a more granular level, ARCserve deploys backup agents that make it possible to recover individual files within minutes.

"I know we're not on the cutting edge of backup technology, but I'm a hands-on manager in a small company, and I want to know that something works, especially for backup," Mintel says. "ARCserve is scalable, manageable, and you get your data back quickly."

Mintel acknowledges that Frost-Arnett could improve backup procedures and shorten recovery times, but he cautions that SMBs have to protect their data within the limits of their staffs and budgets.

"I'm open to new solutions, but I'm not a risk-taker with my company and its business," he says. "I've had 100 percent availability whenever I've needed it with our current strategy, and that means a lot. If we don't have the backup data to get up and running, it can cost us tremendous downtime and revenue — that's our livelihood."

>>For more on tape automation, see how Crystal Springs Resort will be able to access all its most critical data during recovery.