Consultant Toigo issues caveats on storage virtualization for disaster recovery

Disaster recovery seminar speaker Jon Toigo warns that server and storage virtualization can lower visibility into the IT infrastructure; admins say compliance drives more DR awareness.

NEWTON, Mass. -- Although enterprise data storage administrators have pointed to server and storage virtualization as tools to make disaster recovery (DR) more affordable in the economic downturn, consultant Jon Toigo warns that virtualization can obscure visibility into the underlying infrastructure and possibly complicate recovery.

The CEO of Toigo Partners International spoke to users from the New England area attending the Storage Decisions Disaster Recovery Planning Seminar Tuesday and advised them to think carefully about the role of virtualization in disaster recovery plans.

"Some of the foibles we have in IT are being built right into products by joining proprietary software to proprietary storage controllers, locking the customer in," he said. "You should manage the software separately, in two different layers, or you risk obfuscating visibility into the underlying environment."

More on storage virtualization and DR planning
Top server virtualization myths in disaster recovery and business continuity

SteelEye launches SteelEye Protection Suite for Windows Server

DataCore Software debuts Advanced Site Recovery for physical and virtual disaster recovery
Toigo said some storage virtualization software avoids lock-in, such as the gateway version of DataCore's SANmelody and SANsymphony or FalconStor's Network Storage Server (NSS). However, FalconStor this week also introduced an integrated version of NSS, called the HC series, saying users want to simplify management.

Toigo also said server virtualization, credited by some for making disaster recovery more affordable for smaller businesses, isn't necessarily all it's cracked up to be. "Why is [server virtualization] being presented as God's only solution to everything that ails IT?" he said. "A significant technical issue none of the hypervisors have managed to resolve is the case of an 'illegal' call to the storage layer that the hypervisor doesn't understand. That can be a challenge to stability."

Toigo pointed out that implementing storage and server virtualization together can produce too many variables and potential points of incompatibility. Visibility into the VMware file system is an issue that has caused friction with storage vendors in the past.

"Virtualization doesn't solve the underlying problems of the infrastructure -- it just masks them," Toigo concluded.

However, he said newer storage systems like Xiotech Inc.'s ISE have the right idea when it comes to the use of a Web Services interface. "SOAP, XML and REST can glue together pieces of different applications and allow the application and infrastructure to interact. It's being built into the way data is created today," he said.

Toigo also praised new replication and failover products that take application dependencies into account, calling products such as Neverfail's self-titled products, EMC Corp.'s RepliStor, Double-Take Software Inc. and CA Inc.'s XOsoft "DR wrappers." He said DR "aggregators" that report on the whole environment, such as Continuity Software's RecoverGuard, are also on the rise and recommended for keeping up with DR change management.

Other industry experts view virtualization differently. "We have one client whose [VMware] ESX farm has better uptime than his [IBM] pSeries farm," Burton Group analyst Chris Wolf said. "Regarding stability … many enterprises trust x86 server virtualization for production workloads. Its track record speaks for itself."

Wolf added, "I could see how someone would say [virtualization] masks problems. However, if you look at virtualization from a practical sense, it significantly simplifies many operational challenges. I've worked with numerous clients who had nothing but superficial DR testing for several years prior to x86 virtualization. Many of my clients have purposely virtualized for the DR benefits alone."

Regulatory compliance driving advanced disaster recovery

Attendees at the DR Planning Seminar said regulatory compliance and increased attention from upper management is spurring more advanced DR plans.

Don Pierce, information systems manager for Harvest Co-op Markets in Cambridge, Mass., said regulations relating to credit card information require financial and point of sale data to be sent off-site. The organization is currently looking to move its disaster recovery plan beyond USB drives, though the down economy and a non-profit's budget make that a challenge.

"We might be able to start some kind of 'DR co-op' with other co-op organizations in the country," he said.

Rocco LaMonica, systems administrator for Milford, Mass.-based Waters Corporation, a maker of scientific instruments, said that his firm is regulated by the FDA because it makes equipment that is sold to drug companies. FDA compliance requires replication of data between two separate locations. A CA ARCserve customer, LaMonica said he is investigating CA's XOsoft replication software, as well as a software as a service (SaaS) offering from Microsoft Corp. and Iron Mountain. For budgetary reasons, "it would be advantageous to use same product we're already using for data backup, but we've also tested [Microsoft Data Protection Manager] DPM," he said.

Compliance has gained upper management's attention on the DR front, LaMonica said. "Without IT, the company doesn't run," he said.

Neil Parikh, DR specialist for Cambridge, Mass.-based biotech firm Genzyme Corporation, recently joined his company from the financial service sector, and said "The concept of disaster recovery is new territory" for Genzyme. However, Parikh said he was brought in specifically to develop the company's DR plan with a more systematic approach.

Compliance is raising awareness, but Parikh said IT still needs to be a better advocate for DR planning. "What people [in business management] don't realize are the day to day events that occur [in the data center]—IT tends to manage it and it's not very visible to the outside," he said. "We need to take it to senior management and say, 'here's what the potential cost would've been if this wasn't mitigated, here's why we were able to intercept it.' It's an educational process IT needs to be better at."

Dig Deeper on Disaster recovery planning - management

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.