When your business lives in a virtual world and your company is open 24/7, managing a business continuity plan is vitally important. Because the stakes get very high very quickly in the instantaneous world of the Internet, a continuity plan can actually save a business.

Experts in e-business continuity issues say IT plays an increasingly important role in today’s business success, and IT managers should have a plan that addresses the kind of risk their electronic operations face.

The plan also should contain one or more of these components:

  • Geographic load balancing and a design for no single point-of-failure
  • Internet traffic redirection and/or call center automatic fail-over

High stakes shouldn’t include a gamble
“Unlike five years ago, today IT does not take a back seat to the business process; it often is the business process,” said Donna Scott, vice president and research director at Gartner Group, Inc., in her program notes from the recent GartnerGroup Symposium/ITxpo 2000.

She observed that recovery from a business interruption must be quicker than at any time in the past, or else e-businesses (or the e-business components of brick-and-mortar companies) will suffer unwanted press coverage, loss of current and future revenue, and a loss of customer and supplier confidence.

As an example, Scott pointed to the now-famous eBay outage that lasted 22 hours in June 1999, resulting in a $3 to $5 million loss for the company. But, she said, eBay had few serious competitors in the online auction market at the time, and the damage to its competitiveness was limited because customers had few alternatives.

According to IT auditor Keith Young, “ … from a managerial perspective, [IT professionals need] to focus on where’s the risk and how do we address it.” Young recently wrote an article about e-business continuity that was posted on ITAudit.org , a Web site sponsored by The Institute of Internal Auditors. “When you look at the risk, in terms of continuity, you need to look at logical risks as well as physical risks.”

He defines the two types of risks this way:

  • Logical risks: Disruptions that occur as a result of events such as component failure, denial of service attack, or other security failure.
  • Physical risks: Disruptions that occur as a result of events such as server failure, power failure, or building damage.

Control your fail-safe
Logical risks are the most difficult to predict and to anticipate. System components never fail when they are not in use. An attack on your site’s security is impossible to predict. Physical risks can include fires, flooding, or any physical threat to your equipment and site.

To ensure that your continuity plan addresses both logical and physical risks, it should include:

  • A geographic load-balancing scheme.
  • A duplication of Web and application servers.
  • A synchronized or backed-up database.

From the Internet client at the top of this graphic, a pattern of duplication begins with load balancing that routes the client to the appropriate server, which in turn accesses services and data from a backed-up database.

“From an infrastructure perspective, to achieve the highest levels of availability the enterprise must ensure no single point of failure,” Scott wrote. To do this, every component should have a backup to help eliminate unplanned downtimes, like a component failure, or for planned downtimes, like system maintenance or upgrading.

Along with duplication of the Web servers spread between different geographic regions, that kind of distribution also prevents the system being compromised by a natural disaster or other geographical anomalies, such as a blackout of a state electrical grid.

Scott also emphasized that any system failure should be transparent to the client users; otherwise, their confidence in your site may be undermined.

Redirecting traffic
At the very least, a continuity plan should contain protections that include Internet traffic redirection and call center automatic fail-over, Young said.

Of the two strategies, having more than one Internet traffic pattern to your virtual business is the easiest to envision, and one that many businesses already incorporate into their intranet. If the primary connection fails, traffic is automatically routed to an alternate ISP.

While traffic redirection can be invisible to the customer, sometimes it helps reduce customer frustrations with some transactions if the alternative sends them to a real person with whom they can talk.

“Even if your system is down, if you have something like LivePerson , which is a technology where you can click on a button and get a live person, your continuity goes all the way from your architectural design to the browser,” Young said.
You may feel like you work 24 hours a day, seven days a week, but your network and Internet clients probably really do. How fail-safe is your presence on the Internet? How many things would have to go wrong before your company would disappear from the virtual world? Got a weak link in your system? Is duplicating everything impractical, unrealistic, or cost-prohibitive? Post a note below and let your peers know what you think or send us a note via e-mail.