Getting a handle on their respective disaster recovery plans was of big concern among many attendees at Storage Decisions 2003. According to T-Systems Vice President of Consulting Damian Walch, their concern is well placed.
Walch, who spoke to conference attendees on the topic of disaster recovery, shared a widening chasm between what internal customers have come to expect from IT in the event of disaster and what IT is prepared to deliver.
"If you go to users and ask them, 'When is your recovery time objective?', most of what I'm hearing is less than 6 hours," he said. "But, if you go to the IT group, they'll tell you it will probably take 72 hours or more [to recover]."
One reason Walch identified for this disconnect is the fact that collaboration between the business group and IT never happens. "IT never goes back to the business users and says, 'You want this recovery time objective? Here's what it's going to cost.'"
Walch found 56% of attendees were using internal resources (versus a commercial hot-site vendor) to implement their main alternate site strategy for disaster recovery and business continuity.
Why is IT so wary of using commercial vendors for this service? Walch cited a few reasons, such as:
- Most recovery vendors cannot keep up with the technology.
- Exhorbitant charges by commercial venodrs for such areas as technical support and floor space.
- Most companies won't declare disasters and choose to handle them internally, because in a lot of cases, they know they will likely be up and running pretty soon.
- Customers feel a lack of control and are not always guaranteed access to the hot site in the event of a disaster.
Walch indicated the most common causes of business interruption were power outages, followed by a hardware problem, then telecommunications failures.
During the session, he summarized situations in which either internal recovery from disaster or use of an external hot-site vendor would make the most sense. He also provided a starting point for implementing a disaster recovery strategy quickly on just a few of your most mission-criticial applications.
"Mirroring technologies are critical," he said. "If you have one to two mission-critical applications that you can separate out, you have to mirror that information and get it offsite."
He suggested starting small with disaster recovery planning to cover critical systems first, then evolving to encompass other areas later. "Take your number one or number 2 application and say, 'Our company is going to build recoverability in that application over the next six months.' As you move on, you can look at things like e-mail."
Presentation slides and links to other session proceedings are available here.
About the speaker: Damian Walch is the vice president of consulting at T-Systems.