‘Hyperscale’ cloud providers (the big guns like AWS and Azure) have assumed a dominant position in the cloud market. But scale brings its own set of problems, as any obese person can attest.
Patient picture from Shutterstock
The distinction between “hyperscale” providers (who operate globally and across a broad range of markets) and more targeted cloud providers (either regional or aimed at a specific vertical) has been a recurring theme at the NetApp Insight conference in Las Vegas, which I’m attending this week. NetApp’s vision of a /”data fabric” which can link together multiple providers in a hybrid environment presumes that a hyperscale provider is likely to be in the mix.
“If the hyperscale players have a role to play and create value in your enterprise, then not only do we want to support that, we want to accelerate that,” NetApp CEO Tom Georgens said during the opening keynote. But that remains easier said than done.
One obvious challenge with that integration is lock-in: one you’ve tied data into one cloud platform, using it elsewhere can be fiddly. “A lot of clients have this vision that clouds are interchangeable,” Georgens said. “To some extent that’s true, and it’s particularly true of fungible technologies like servers, where there’s no history and no state. But storage is different.”
“Every cloud provider has this discussion with the enterprise: rewrite you apps so they work in my cloud,” said George Kurian, executive VP for product operations. “Once your data is on that cloud, there is now way to move it to another cloud.”
A second issue is that the kind of enterprise contracts which large companies rely on often aren’t always on offer from hyperscale providers, who rely heavily on standardisation — offering the same terms to all comers. Analyst Trevor Clarke pointed this out during the keynote:
Challenge with cloud integration or ‘data fabrics’ isn’t tech. It’s the commercial models and contract terms. Long way to go. #NTAPInsight
— Trevor Clarke (@trevclarke) October 28, 2014
To be clear, hyperscale providers are happy to negotiate individual terms if the deal is large enough. That’s arguably especially the case for Microsoft, which is used to negotiating enterprise contracts. However, ensuring that all your contracts are on broadly similar terms and time frames is likely to be a significant challenge.
Finally, it’s worth remembering that even the global providers don’t offer everything everywhere. As we noted when Azure launched its local data centres in Sydney and Melbourne earlier this week, not every Azure service is available in every Azure region.
Case in point: neither the Sydney or Melbourne regions offer Azure’s Site Recovery Feature. NetApp’s Clustered Data ONTAP 8.3 platform can tap into and integrate Site Recovery — but that’s not going to be an option in Australia just yet. Even hyperscale has its limits.
Disclosure: Angus Kidman travelled to Las Vegas as a guest of NetApp.
Comments
One response to “Hyperscale Cloud Hassles: Lock-In, Contracts And Availability”
uhh, you couldn’t think of a better and more appropriate or relevant cover photo than a fat guy? Then you make an attempt at a fat joke in the first section…? Hard to take anything else written in the article seriously after that… and (not that it’s really relevant, but) I’m a skinny guy…