The Connected Web

Listen to my interview with Amitabh Srivastava, corporate VP of Windows Azure, Microsoft's cloud platform, which was unveiled last month at the giant Professional Developers Conference in Los Angeles. The interview is recorded as two podcasts (Part I discusses Azure for the enterprise). In Part II below, he explains Microsoft's roadmap for Azure and what developers can do with it today.

Listen to or download the 7:37 minute podcast below:

Download file

---Transcript---

PW: Right. So okay, so now, I mean this platform is based on the experience you got in running your web properties. Have you actually have you put any of Microsoft's properties onto Azure?

AS: Yes, remember we are using both our experiences on running the web properties, but also, we are using the experiences that we had from using from running enterprise software because we know what our enterprise customers want. And the reason we were building -- the way we are building this platform for is basically we're going to move all our Microsoft services on top of it.

Clearly, this is going to be a phased process. So we started on top of that. We've already gotten -- Live Mesh is already using Windows Azure. You know we have gotten the next version of Live Meeting is being built on top of Windows Azure. There are a bunch of properties, which are in the process at different stages of being moved to Windows Azure. Like I said, the ultimate goal is to move all our properties on top of Windows Azure.

Right. So you're gradually moving everything across and that really is actually eating your own Red Dog food isn't it?

That's right. I mean, if you look at it, the biggest service like inside Windows Azure, we have our storage service, for example. It is a very complicated storage service. It's all blobs and tables, and we have written that as a service. It follows all the rules of our service. It itself runs on Red Dog.

Right. Yes, I mean that's very interesting ...

Yes, we actually have been bootstrapping and using -- I'll give you an even more interesting one is that the entire development of Red Dog was done on Red Dog itself. So for example, all our developers that we have today they don't have their own machines or any clusters. We basically have a Red Dog cluster on which they keep on running and testing the services out. So the whole process that we have done is basically being really to keep testing Red Dog, keep testing Windows Azure as we developed it. I'm sorry I've been using the word Red Dog, but it's really Windows Azure.

Right, right, okay. But hence my pun about you eating your own Red Dog food. I mean what about customers, can customers start using Azure today?

Yes, they can go to www.azure.com. One thing they should do is they should hit the "Connect" button and get a token. Try to get -- put themselves on a list to get the token there. They should download the SDK. You can do an entire development on your desktop. I mean that's very, very -- we used the same tool for developing Windows Azure itself.

Right, and you've enabled that so that people can use their existing tools to develop to it and develop on the desktop and then kind of ...

Press a button and then --

-- push the results to the cloud, yeah.

Yes, exactly and then take it to the cloud. So you can download the SDK today and start experimenting with it and then get and start using the cloud there. I mean what we're really looking for is the feedback and that was --

But today, it's not a production platform from a customer's point of view?

Well it is real, I want to emphasize that. It is real because we have been running in it internally. We are running real services on it. Externally, we want to see the feedback. We want to really understand, are we really solving the key problems that we designed the systems for and we want to tune the system down once we get the feedback.

So right now, what we want to go do is like have a lot of people use it, learn from their experiences, and see what they say, and then tune the system, and then hopefully, depending upon how the feedback is, we will decide what the next phase should be. So our plan is to kind of like expose Windows Azure in a staged way.

Right. I mean that's interesting because the normal cloud provider kind of -- you know, if Google was doing this, they would kind of slap it out there, put a Beta sign on it and basically say kind of, well, start using it and let us know if you have any problems. It sounds as if you're taking a more phased approach to rolling it out.

That's right. And we want to -- and that's why we're not charging anything either. And what we want to do is because we're trying for multiple audiences and our goal is to go from hobbyist to enterprises. And we want to make it free, we want people to try it, we want to get the feedback, and we're working with different segments internally not just the internal customers, not even the ones there. We're also working with a whole bunch of ISVs and also some enterprises customers at different phases to really see, are we really solving the hard problems.

You see, our goal was not just to put a few services on the cloud -- that would have been an easier thing there. And our goal was to really say, can we really tackle the hard problems these customers are facing, the enterprise customers and things. And that's the one we want to really go, and understand, and see are we really solving them, and we're willing to tweak or change if we missed anything.

So finally, when are we going -- because I know products like, applications like SharePoint and Dynamic CRM were mentioned in the announcement as products that would run on Azure. When will customers actually be able to start using those services on Azure?

I think like I said, I mean, we are already in the process of moving, revamping them as fast as we can. And we started with Mesh. You'll see SQL Services on it. I mean we're just going to go one after another one and just move services on it. We're right now already -- we have these Windows Azure running in our data centers now, so you will see more and more services happening in the course of next year.

Right. I mean in a sense, it's the kind of classic announcement of a product which is there, but actually when people are not going to be able to use it in its full glory probably into next year or maybe the year after.

Well, see there are two approaches that were there, Phil. One was that we could actually wait and get the whole thing completely done and come out.

Right.

And the other approach was we call it like a tech preview. And we have enough functionalities in that can allow people to use it and do meaningful work with it.

Right.

Because we -- and really see to that -- we chose this approach which is to -- we didn't want to announce it before we had anything, but we have enough of it so that people can do useful things with it, that we want to get out and start getting the experiences in and then we will iterate really fast. We will keep on adding more features and see are we really solving the problems and depending upon how the feedback we come out, we can open it up more.

Okay. And the other thing that I think is great from my perspective as someone who writes about Cloud, and Software-as-a-Service, and all of these wonderful kind of web-based things is that Microsoft has really got behind the Cloud now hasn't it?

Oh yes, this is something that we've been doing these services for a long time and we were taking sort of like a thoughtful approach to really say how do we take an approach for it. And our thinking was to really go build the operating system for the Cloud and that's what Windows Azure is.

Right. Okay. Great. Well, thanks very much. I'm Phil Wainewright. I've been talking with Amitabh Srivastava who is Corporate Vice President responsible for Windows Azure, Microsoft's Cloud platform. Amitabh, thank you very much for joining us today.

Thank you very much Phil.