As any IT pro knows, any amount of downtime is it too much, and transforming disaster recovery is a top priori...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
According to a recent study by Enterprise Strategy Group (ESG), outages cost companies anywhere from $5,000 to more than $1 million each hour their systems are down, said Terri McClure, ESG's senior analyst. And with more data being stored in the cloud, companies are more vulnerable than ever to outages.
At the same time, the amount of downtime companies can manage has decreased significantly -- from four hours in 2004 to only a few minutes in 2010, the survey found.
Managers need to understand the indirect costs of downtime.
senior analyst, ESG
ESG, which conducted the study in conjunction with Silver Peak Systems, surveyed 540 IT professionals. The consultancy found that business continuity (BC) and disaster recovery (DR) programs were among the most cited priorities for IT, with 20% stating that BC and DR were their most important initiatives.
"I think the biggest surprise of the survey was how valuable virtual machines are in this transition from a physical to a virtual world. Being able to deploy VMs and disaster recovery where you absolutely wouldn't have been able to do so before is a testament to how much virtualization has changed the industry," McClure said. "What really jumped out at me from the interviews is the difference VMs played. I don't go as far as saying 'meeting the impossible recovery point objective [RPO],' but it really was. They really couldn't meet their RPOs without the acceleration technology."
Even companies with robust DR have performance issues
Everett Dolgner, Silver Peak's director of storage and replication product management, used 2012's Hurricane Sandy as an example of how quickly downtime costs can mount up -- even for companies with robust BC and DR systems.
"If you look at the environment they have set up it's impressive," Dolgner said of the Manhattan-based companies that were battered by the storm. "They had bandwidth coming from two different wires. One goes over a tunnel, one goes over a bridge. And they had redundancy protection." But when Sandy tore through the city, all their efforts were not enough. "The winds and the rain went miles out from the eye of the storm. Even if it's only 100 miles, that is still a great distance from the synchronous replication. It takes down your data center and disaster recovery center at the same time. The longest period of downtime I heard was 12 to 13 minutes. In this situation, the tens of millions of dollars they spent on disaster recovery didn't protect them when the scale of the disaster was so high."
Loss of reputation is another cost, McClure said. "Managers need to understand the indirect costs of downtime. It's not just about lost transactions. Think about the advertising industry. It's about meeting deadlines. During the Super Bowl, you can't miss your deadline for your commercial to go live. If you miss your deadline, no one will want to work with you again. It isn't always just a one-time dollar amount; it's your reputation value."
Silver Peak used the ESG research to create a disaster recovery return-on-investment graphic to illustrate the impacts of downtime and the steps companies can take to diminish or eliminate its effects.
Cloud storage transforming disaster recovery
In addition to the potential impacts of natural disasters, the increasing amount of data stored in the cloud adds another element companies must consider when designing a DR plan. "What we've really seen over the last few years is people wanting to move to the Internet with cloud- based recovery or MPLS [Multiprotocol Label Switching] to recover data. But because they are shared mediums there will be packet loss, out of order [transmission] and contention. When you combine this with latency, things go horribly very quickly," Dolgner said.
"From a business perspective, you live and die by your recovery point objective. A simple way to explain it is how much data can you afford to lose? Set that objective for your application and you need to manage that RPO."