Mountains Of Data Bring Recovery Issues
July 7, 2014 Dan Burger
If IBM pushed any harder on promoting “The Cloud,” it would rain for 40 days and 40 nights. Investments have been made in technologies such as PowerVM, SmartCloud management products, and Live Partition Mobility. And you can’t undervalue the investments from the managed service providers (MSPs) and the independent software vendors (ISVs).
“We see a revenue shift from the traditional buy/sell business to ‘the cloud’ and managed services,” says Jim Kandrac, who makes his living selling cloud-based backup and recovery services and remote hardware for disaster recovery and high availability. Kandrac’s business model relies a combination of IBM Power Systems and Intel servers to handle the managed service business he refers to as backup as a service (BaaS) and infrastructure as a service (IaaS). His company, United Computer Group (UCG), is equipped to handle online backups of databases from 10 GB to 100 TB.
From June 2013 to June 2014, Kandrac says UCG net monthly recurring revenue (MRR) from backup as a service and remote hardware recovery has increased more than 33 percent. He says UCG is on track to double MRR every three years.
This doesn’t prove MSPs are getting more disaster recovery business than any other kind of cloud-based business, but it is another piece of evidence that BaaS is lifting the cloud. MSPs consider disaster recovery as the gateway to the software as a service and the lucrative services business.
“We’ve been doing backup and recovery for eight years,” Kandrac says. “There has always been a situation where the IBM i people were doing their thing and the AIX and Windows people were doing what they do. Now we are getting calls from mostly IBM i people who have had some bad experiences on the Intel side. Things didn’t work as advertised. Too much pipe, too much data, problems compressing data, and trouble connecting to the remote location are some of the problems we’ve heard from the Intel side. Our business has improved because of this.”
The challenge on the Intel side has not been on with archiving, but on restoring because companies that used to have gigabytes of data are now breaking into the terabyte range. Several terabytes of data (what Kandrac estimates is reality for 85 percent of his customers) are unable to restore in a timely matter unless they move to a high availability environment, which for some companies is a great idea, but for a lot of companies is overkill. Kandrac says hybrid solutions using solid state disk will be the way this problem is solved of getting 24-hour restores (what most UCG customers require) without the cost (almost double) of HA. It’s happening now, but it’s not commonplace. As companies increase from 4 TB to 8 TB to 12 TB of data, at the current capability levels of rebuilding data, the 24-hour recovery is unattainable.
“The trend I see is that the Intel guys are talking to the i guys because the Intel guys tried something that didn’t work,” he says.
That’s opening the door for companies like Kandrac’s that can do both.