Q&A: Disaster recovery when your business sits on the San Andreas Fault
Graniterock, a century-old provider of construction materials in Watsonville, Calif., operates a quarry that happens to sit on the San Andreas Fault. The company has 20 locations from Oakland to the Monterey Bay, connecting to the data center via a wide area network. If any CIO ever needed a sound data back-up and disaster recovery plan, it's Graniterock's Steve Snodgrass. Graniterock started using a hosted solution from WTS Inc. for its JD Edwards system in the late 1990s, eventually receiving disaster recovery and business continuity service as part of the package.
When hosting provider Velocity Technology Solutions recently acquired WTS, it expanded its place in the disaster recovery business. When a catastrophe hits, Velocity recalls a customer's data and ships it to an alternate data center, where the customer's hardware sits "ever-ready." Operating systems and applications are in place for hardware that matches the customer's production configuration. Customer data is restored from tape, with a recovery time of 48 hours.
In an interview with FierceCIO, Snodgrass, who holds the title of CFO as well as CIO at Graniterock, discussed his approach to disaster recovery in an earthquake-prone area.
FierceCIO: How large does the prospect of an earthquake loom for you?
Steve Snodgrass: I experienced an earthquake two nights ago, but we haven't had anything extensive since 1989. We were definitely affected by the 1989 earthquake. It was centered about 10 miles from where I sit. There was significant damage throughout California.
FCIO: What other types of catastrophes do you worry about?
Snodgrass: It's really just quakes that keep me up at night. Fire is generally a localized problem. If an earthquake takes down a building where our corporate data center is, it's going to take down a lot more than that. The company has about 20 locations, and the data center is accessed by a wide area network.
FCIO: What prompted you to go with your current disaster recovery solution?
Snodgrass: We installed JD Edwards in 1998 and ran it internally for two to three years. I come at IT from the finance side, and I looked at the system and said, I don't think I have the redundancy of personnel we need. What happens if your primary support on your ERP system gets hit by a truck? We have only one and a half people dedicated to the ERP implementation. This was around the year 2000, and finding good people and retaining them was a problem.
I called WTS [now Velocity Technology Solutions], which had basically started a hosting business. People generally were not going to hosted solutions or the cloud back in 2001. We didn't call it that, but that's what it was. We swallowed hard and chose them as a hosting provider. Over time they've improved their infrastructure and kept their price to us fairly constant.
FCIO: Can you describe a specific time when you were really glad that you had this solution in place?
Snodgrass: When you're a mid-market company you cannot afford to have multiple AS400s. What we've seen happen over the years is that Oracle will release upgrades to JD Edwards so it will no longer run on the operating system. The hosting provider has multiple AS400s, and when this happens we can migrate over in an orderly fashion. Without a hosted solution, a company our size would have to go out and rent AS400s. But we may not even have the space for them.
FCIO: What are some of the other advantages of cloud-based solution over a non-cloud based solution?
Snodgrass: Part of the reason I still like the decision [to go with a hosted solution] is that their expertise and ability to run a data center far exceeds our ability to do so. I've been with this company for about 20 years, and the decision to go with WTS [now Velocity], when I revisit it, was probably one of the best decisions I've made in my career.
The reality is that we're a mid-market company. We have 600 employees, including nine on the IT team. We're a construction company--we're not an IT company. The economics are such that it would probably have cost us more to host a solution internally, and we would not have done as good a job.




Comments