Two Clicks Away from Disaster: IBM Addresses ERP Recovery Planning

Until recently, if a system went down due to natural or man-made disaster, the business could work around the failed system until it was up and running.But with the growth of integrated ERP systems that affect every corner of the business, a system failure could be devastating.

Until recently, if a system went down due to natural or man-made disaster, the business could work around the failed system until it was up and running, which usually took up to a week. But with the growth of integrated ERP systems that affect every corner of the business, a system failure could be devastating.

"The time a company can withstand a disaster and stay in business has shrunken dramatically," says Stephen Higgins, marketing director for IBM Business Recovery Services (BRS). "Your ERP application may be looking for a management or call center application. If one of those is down, you can't service your customers, or get to your suppliers to order your next set of resources. Today, your competitor is only two clicks away."

Within the next two years, most large organizations will require recovery of ERP systems in less than 24 hours, relates Donna Scott, VP of software infrastructure for GartnerGroup. That's because increased dependence on enterprise-wide systems has dramatically increased the vulnerability of the entire business.

To support this new model, "the hotsite vendors need to move from a platform focus to a business and application focus," she notes.

With this in mind, IBM Global Services recently announced a formal program for reviving ERP applications that run across multi-platform, networked data center environments. At this time, roughly 20% of IBM BRS's customer base is running SAP or other ERP systems on a Windows NT platform, says Stephen Higgins, marketing director for IBM BRS. "Until recently, NT servers were initially used as the development and test beds," he says.

"Folks would buy an SAP suite, a few application sets, and play with it on the NT servers. Then deploy it on Unix. Now, we're seeing a growing NT base for SAP implementation."

However, for many companies implementing ERP, disaster planning is still an afterthought, Higgins states. "Many customers that we've been working with are already in post-production," he says. "They come to us after the fact for recovery services. That's okay, we can do that, but the process is heightened." Ideally, business continuity should be a part of the application implementation process right from the get go, he says. With proper up-front planning, a typical SAP system can be restored in a matter of hours, he adds.

IBM BRS plans to deliver its new ERP continuity services at intervals over the coming year, with the first services for SAP's R/3 system available immediately. IBM is also planning to roll out services and support for other ERP suites, such as JDEdwards, later in the year, says IBM's Higgins. "First, we want to come out with a proof point so we can demonstrate that the solution is ready to go, with live customers behind the announcement. We're being very methodical about how we're approaching this marketplace, making sure we can develop the proper scripts and methodologies to help us support this very complicated and very important application set."

That's because ERP systems are extremely complex systems to get back up and running in a short time, Higgins explains. "It's no longer a situation where we can have a machine up and running after it breaks. You have to have the skills in place to bring up applications."

In a Window NT-based ERP environment, "you typically only run one application per node, GartnerGroup's Scott points out. But your critical application is order entry, not manufacturing, purchasing or human resources. You still have to recover the whole thing at once. You can't prioritize individual components, because it's one single database. The whole thing has to come back up."

The challenge becomes restoring or re-synchronizing a system spread across hundreds of Unix and Windows NT servers. "Simply prioritizing the recovery across all those is a big effort," she points out. "You want to make sure that some of your applications and data are going to have recovery time objectives of an hour, some will be a day, some will be two days, some will be three days."

One customer, Phillips Petroleum Company, has been working with IBM to develop its ERP continuity plan for its worldwide SAP implementation. "Our SAP-supported business processes are highly integrated and therefore interdependent," says Marshall McGraw, manager of IT architecture and strategy for Phillips Petroleum.

The company has been working with IBM BRS to develop a plan that will minimize the impact of disruptions to the SAP system.

IBM's new ERP offerings include consulting, outsourcing and application contiguity services. ERP continuity plans include business impact analysis, test and recovery procedures, WAN and LAN recovery topology, and ongoing test and recovery management.

IBM BRS has been shifting its focus to concentrate more on building business continuity processes into the systems development process up front, as opposed to simply providing hot-site services.