X
Business

Ozzie responds: Is Microsoft Azure just 'Hailstorm' revisited?

At the Professional Developer Conference in Los Angeles, I've heard a few long-time Microsoft watchers wondering aloud whether Microsoft's newly unveiled "Azure" isn't simply Microsoft taking another run at "Hailstorm." I had a chance to ask Ray Ozzie, Microsoft's Chief Software Architect, that very question at the PDC.
Written by Mary Jo Foley, Senior Contributing Editor

At the Professional Developer Conference (PDC) in Los Angeles, I've heard a few long-time Microsoft watchers wondering aloud whether Microsoft's newly unveiled "Azure" isn't simply Microsoft taking another run at "Hailstorm."

I had a chance to ask Ray Ozzie, Microsoft's Chief Software Architect, that very question this week.

First, a quick refresher: For those who weren't following the Microsoft juggernaut back in the late 1990s, Hailstorm was Microsoft's first pass at a .Net services platform. "HailStorm" technologies will enable a new world of computing where all the applications, devices and services in an individual's life can work together, on their behalf and under their control," explained Microsoft in a 2001 press release. (Sounds eerily like Live Mesh/Live Services, doesn't it?)

Microsoft ended up killing off Hailstorm before it ever really launched. One of the main reasons was privacy: Microsoft customers were nervous about trusting Microsoft with hosting their data. And the idea of an on-premise, customer-managed Hailstorm cloud was not fleshed out.

Isn't Azure -- Microsoft's new cloud platform, of which Live Services are one key component -- just Hailstorm Take 2? And if it's not, how is it really different, I asked Ozzie.

Interestingly, back in 2001, Ozzie was working at Groove Networks, which was one of Microsoft's Hailstorm partners/customers.

Microsoft "did not understand identity federation at that point in time," Ozzie said. "I think we (Microsoft) were thinking we could bring it all there. There was no code in the cloud. It was a service that projected services outward but it didn't -- you didn't upload your stuff. It was just -- it was about 'my stuff.'

"In that era (Microsoft) believed in pure centralization in 'true integrated storage,' the middle. My DNA tells me -- and you can see this in my past and what I've done before -- that it's a world of decentralization and that you synchronize the truth amongst many things, both in the center and at the edge on the devices where you need them. So, that's very architectural different.

"One of the other really big things is (Microsoft) had no clue as to the business model in that era.  We actually understand the business model now."

With Azure and Mesh, Microsoft has moved beyond these limitations, Ozzie said. But there are still a few problems that plagued Hailstorm that are still unsolved now -- by not just Microsoft but the rest of the industry, too --  Ozzie conceded.

"It's amazing that at this point in time, as compared to that long ago, (that) we still don't have that one nailed from a privacy and ownership perspective. That was what so many people complained about. But right now you've got Open Social and Facebook Connect, and both of them want to still create walled gardens, open walled gardens, whatever that is, but that they are lending you your information back and withdrawing it within 24 hours or whatever.

"I think we need to get past that, and what we're trying to do with Mesh and the terms of use. We're trying to get to a point where you literally do own your data, we bring the personal of the personal computer to the cloud where it's your stuff, and if you do something with someone, it better be a symmetrical synch relationship where you're giving them rights, they're giving you rights, because I just don't see how it works. We can't create a walled garden; it's just not going to work."

Readers: Do you agree wtih Ozzie's assessment? From what you've heard so far about Live Mesh and Azure, do you think Microsoft has solved the issues that killed Hailstorm before it ever got off the ground?

Editorial standards