As a VAR, DR seemed like an ideal solution since it included multiple products and was relatively complex so it would drag a decent amount of PS. But it was always a tough sell. Customers were usually clueless about what a good disaster recovery solution entailed, but getting them to pay anything to resolve this problem was very difficult. You could say they “didn’t know what they didn’t know,” namely, that they had a problem, and therefore were less apt to spend anything on it.

One way to sell DR is to get users to focus on the negatives, the risks they’re running, the cost of downtime, etc. This is the approach everyone takes after a hurricane, like Sandy or Katrina, and was particularly popular after the Sept 11 terrorist attacks. But people have strong denial skills and just aren’t compelled by potential pain to themselves when it’s exemplified by the misfortune of others. Part of the reason may be that the solution has more than a little pain itself.

This is another example of the pain-of-change equation. If it’s more disruptive, expensive, etc., to do nothing than it is to fix a problem, people do nothing. Testing a traditional DR system can be disruptive and expensive as it often requires after-hours work by a number of people at the primary and remote locations and maybe some application downtime as well. It would follow, then, that making DR testing quick and easy is a good way to lower that pain-of-change delta and get people interested in a  disaster recovery solution.

Hybrid cloud DR systems allow customers to back up their application servers to the cloud as VM images and then restart those virtual machines on host servers in the cloud. These systems have the added benefit of making DR testing almost trivial. Users can start these virtual servers in the cloud with a couple of mouse clicks. This can significantly lower the pain of running a DR solution and potentially make it an attractive topic to bring up with customers. 

Follow me on Twitter: EricSSwiss

Humans are reactive, not proactive, except in terms of the next event. After each of these disasters, there was certainly a heightened awareness and some action taken by companies, but it’s safe to say that fewer companies took the lessons of disaster preparedness to heart and actually implemented credible DR plans.

Part of the reason is that credible disaster recovery planning has historically been expensive and complex. Starting with offsite vaulting of backup tapes and evolving through disk backup, deduplication and WAN-optimized replication, up until recently DR solutions have remained beyond the means of most SMBs. Now, however, technology may have come to the rescue. The cloud and widespread server virtualization have created a real DR solution that most companies can afford.

Hybrid cloud backup combines cloud backup with an onsite appliance that takes backups from local servers and then facilitates their transfer to the cloud. When this appliance is designed to take backed-up virtual machine images and even perform physical-to-virtual conversion when backing up standalone servers, a real recovery capability is born. Downtime is reduced to almost zero since these VM images can then be restarted on the backup device. When the appliance synchronizes itself with the cloud, where a compute platform is available from which to start and run those VMs, it becomes a real DR solution.

This “hybrid cloud DR” is being offered by many of the same companies that provided hybrid cloud backup and is becoming a part of more and more storage appliance offerings, as an optional service. This means good things for SMBs since it promises to bring costs down further, although hybrid cloud DR is already in a separate class from “traditional” DR solutions, from a cost and complexity perspective. It’s also good news for VARs, especially those that can offer the cloud service as well.

Follow me on Twitter: EricSSwiss

And while natural disasters underscore the random nature of catastrophes, they don’t make everyone feel vulnerable. Here in Colorado, CIOs don’t fear tsunamis or earthquakes like those experienced last week by the people of Japan, since we don’t live on any fault lines. But the problem that befell Google was not caused by external events but by a bug in storage software. 

Apparently, corruption and data loss was propagated throughout the storage infrastructure with real-time replication — one could say, a different kind of wave. There are two messages that come out of this story that VARs can use to their advantage. First, if this can happen to Google, it can happen to anyone, anywhere. Second, simply relying on snapshots, replication and other complex forms of data protection that don’t actually make a copy of data (an off-site copy, if you’re concerned about big waves) can leave you vulnerable.

For VARs, the first thing to do is compile a good “elevator speech” to describe what happened. As is the case in technology, every point made requires some education. This topic, data protection and the vulnerability of abstract storage processes, is especially compelling for midmarket organizations since many people don’t really understand all that’s going on under the covers of their SAN. Of course, the reason to bring up problems is because you have an answer and in this case it’s a broad-based one. Users need to know what they’re doing when they set up critical processes like data protection. If they don’t know, they should get in touch with their “trusted advisor.”

Abstract processes like snapshots and change-based replication rely on a base data set being intact. In this case, the solution could involve something as simple as a periodic copy made to an independent storage system (even to tape). Think of it as the gap in the line of dominoes; the problem stops here. Another area of concern is the simple reliance on getting data back and not on getting applications up and running. Applications are getting more complex and often involve more than one server — like a database back end to a Web front end, for example. The importance of “application aware availability” is a key message, one that touches several systems and one that VARs are uniquely qualified to deliver. In the final analysis, disasters can help VARs generate the right kinds of discussions and focus attention on their potential value.

Follow me on Twitter: EricSSwiss