Picking Apart Microsoft's Patent for Cloud Data Migration

Share it on Twitter  
Share it on Facebook  
Share it on Linked in  

So, I'm strolling along the Intertubes today when I stumble across this little item, which seems to have first appeared on InformationWeek but is spreading like kudzu along a southern highway: "Microsoft Seeks Patent For Cloud Data Migration."


Oh, reeeeeeeally?


Now, it's a patent and only a patent, but this could be big news for those worried about cloud-lock in, which certainly includes Microsoft. After all, Microsoft won't be arriving at the cloud ball until next year's release of Azure, which, as I've already noted, reflects Microsoft's support for interoperability. So arriving with a plan for helping others migrate their data to a new cloud is like arriving to the party with Lady Gaga-it's bound to turn some heads.


You can read the patent for yourself online, but The Register offered this summary of Microsoft's application:

The company describes an architecture that involves executor, detection, organizer, and summary components that will received and verify notices that a cloud service is to be terminated, find the relevant data and service, prioritize the data or service, and give a summary. According to the filing, Microsoft's technology will preserve the data, meta-data, or service that's being terminated, roll it back to a state that allows migration, find a similar service comparable to the one already use, and then provide the migration.

InformationWeek's Alexander Wolfe offered a more in-depth look at the Microsoft cloud migration patent request. One interesting aspect, he notes, is the specified rationale for the technology: Basically, Microsoft is saying the data would be migrated as "an auto fail-over data protection mechanism," and not, as one would expect, because you want to switch cloud vendors.


But come on, this is business-savvy Microsoft, so, as Wolfe points out, that whole switch-vendors thing is in the fine print, a bit farther down.


Cloud and SOA vendor Vordel is already nay-saying that this patent would address lock-in problems, "as identified by the European Network and Information Security Agency," The Register reports:

Vordel has instead pitched the case that only a broker - which it's in the process of rolling out - can really help customers avoid lock in. The company claimed brokers allow users to make a switch over at the interface level to a back up in the event of a failure.

There is a third possible motivation for the patent: At the bottom of the InformationWeek article, Wolfe notes that @swardley believes the patent suggests that when when Azure launches, Microsoft will launch with multiple ISPs, in addition to its own.


Whatever the driver, Wolfe points out Microsoft's proposed patent doesn't address the problem of transferring large amounts-think petabytes-of data over the Internet. How does he know this? Apparently, Sun Microsystems CEO and mathematician Jonathan Schwartz calculated it out, determining that most Internet connections are so slow, you'd be better off moving the data to tapes and shipping it-by boat.