In the wake of recent regional disasters, a number of businesses are creating more room between their main offices and their secondary data centers. But as they ponder moving their data further away, enterprise decision makers are running into new problems and higher costs, forcing more creative solutions.
Following the September 11 terrorist attacks two years ago and this summer's East Coast blackouts, regional disasters have become a greater concern for the average enterprise. Disaster planners can no longer feel content that a data center across town, or even across the state, will not fail along with the one at headquarters, said Mike Ferrell, a risk manager with Roseland, N.J.-based payroll processing and computer services firm Automatic Data Processing Inc.
At the same time, merely backing up data is not enough to ensure that a business will not go dark during an emergency. Some businesses are enabling entire applications and systems to run from remote sites. In the event of a disaster, businesses need to keep their core applications running, said Chris Nichols, vice president of strategy for Dallas-based data communications and consulting firm MDM Group Inc., which works with many of New York's financial services companies. Simply having database records does not help if a business can't use the data, Nichols said.
But that kind of redundancy is more expensive and requires more resources.
Locating a data center hundred or thousands of miles away from a business' headquarters poses a number of problems. If a company chooses to operate the data center on its own, then it must be able to staff it during an emergency.
Many New York financial services companies have located their data centers just across the river in New Jersey. But during recent disasters, they found that they couldn't even get their staff across the Hudson River, Ferrell said. Plus, he said, even when physical barriers aren't an issue, staffing remote data centers full time or even just during emergencies adds to a company's costs.
For those considering an outsourcer, cost is often prohibitive there as well. Nichols said that most large companies have already outsourced their data centers to service providers, like Electronic Data Systems Corp. But many midsized firms simply can't afford to do so, he said.
Some enterprises are also concerned about the reliability of outsourced data centers. In the event of a widespread problem, like August's blackout, all of a data center's customers may need to run their systems from the facility at the same time.
"It is unlikely that outsourcers design for 100% availability for 100% of their customers," said William Terrill, an analyst with Midvale, Utah-based research firm the Burton Group.
A company should determine which functions are absolutely crucial to its survival during a disaster, and which ones it can live without for a few days, said Terrill. For example, a bank may want to ensure that its ATMs still work or that it can continue to process checks, while online banking or employee e-mail might not be necessary to keep the business going.
Not every business needs to be able to run its operations from a remote site. Terrill said a company can keep costs down by being selective about which systems need redundancy.
Thomas Weisel Partners LLC, a merchant bank, faced just that issues a few years ago. Because the firm is based in San Francisco, earthquakes are a natural concern. At the height of the economic boom, Thomas Weisel looked at outsourcing its data center but found even then that the cost was prohibitive for the small, 500-employee firm.
Instead, the company decided to use its three distributed offices to provide itself with redundancy, said Beth Cannon, the company's chief technology officer. The company has offices in New York and Boston, as well as San Francisco. It decided to use San Francisco and New York as redundant data centers.
Each of the company's departments determined which applications and business functions were critical in order to ensure that the firm would be available during an emergency. It also regularly ships backup data tapes between the two offices.
With this approach, the company has redundancy on opposite coasts at sites that are already staffed. The process has not incurred significant additional cost, and it enables Thomas Weisel to access crucial systems if one of its offices goes down.
More companies are likely to face these problems in the near future, Terrill said. Enterprises are growing more concerned about their ability to function in a disaster, and regulatory bodies are becoming more concerned with disaster recovery as well.
FOR MORE INFORMATION:
Get expert advice on redundancy and load-sharing configurations.
Learn about the value of redundant network links.