Gallery

Climbing on the Cloud

Posted on 8 Dec 2008 at 16:47

Jon Honeyball tries not to get too cynical about azure, finds the best of British in a high-end hi-fi kit, and breaks the iris scanner at Heathrow.

We went to Los Angeles to be stunned, and to be fair I did come away mostly impressed. It's now clear what Microsoft's strategy is going to be for the foreseeable future - rich client apps, lots of internet bandwidth, and a huge Cloud infrastructure of datacenters around the world that will act as the backend to the whole affair. This may sound brave in principle, and it surely is requiring a frankly mind-boggling level of investment by Microsoft in hardware and manpower for those datacenters.

The initial keynote laid out most of these plans, and then it was left up to the comic genius of Don Box to take us through the real-world coding implementation of the ideas. Of course, he made everything look easy - a few lines of code scattered throughout a simple project, and shazam, your app now runs in the Cloud. Of course, this was all too good to be entirely true.

Let's start with that name: to a car aficionado like me "Azure" is of course the name of a particularly swish Bentley convertible, but I realise that to those of a poetic disposition it's the colour of the sky. Unfortunately, though, it's the colour of a cloudless sky (but then "Fluffy White" wouldn't have quite the required ring). Clearly, I'm too stupid to appreciate this marketing message. But back to the development story - the reality is that you'll have to write your application to take advantage of the Cloud, and this isn't just a matter of ticking some checkboxes at compile stage, or running a wizard over your existing code. In reality, you'll need to mould your application to the needs of a Cloud-based client/server system. Okay, so the basics are in ASP.NET so this work shouldn't be too difficult, but there will still be a lot of it required.

Does this matter? In some ways no, because this was a conference for the professional development community and they were looking for a big picture, a route forward. Each attendee left with a USB hard disk filled with install images of Windows 7, Server 2008 R2 beta, Visual Studio in three versions (2008 Trial, 2008 SP1 and 2010), Windows Live Beta, Virtual Server, Silverlight tools - the list goes on. And, as if that weren't enough, members of the Press were given a Dell laptop with Windows 7 preinstalled to provide an instant play and a known reference platform for the beta program. Before anyone screams "bribe", please note that each laptop has a Microsoft hardware inventory number tattooed on it, that Microsoft knows the number of the unit I have, and it will be wanting it back.

At this point, I should be drooling at the potential benefits offered by these tools for moving your datacenter capabilities into the Microsoft Cloud: you'll have multiple instances of your code and data geographically distributed; load balancing across the Cloud; disaster recovery in the a blink of an eye; in fact, all you need for development and deployment. And yet...

Let me be clear that Microsoft is going to deliver on this vision. It has no choice, because its rivals such as Google and Amazon are delivering today in this application space, and you can even get Windows-hosted services from Amazon right now. One could almost say Microsoft is lagging behind, which may betrue, but it's also building a compelling, total end-to-end integration story that others will find hard to match. And yet I stillcannot shake off some cynicism.

There's no doubting that the development community has been badly stung by Vista. Microsoft bravely announced to everyone that there would be "the three pillars of Vista" - XAML-based 3D user interfaces, distributed web services and, lastly, the WinFS object-oriented filesystem. The reality has been that Microsoft under-delivered on the 3D graphics front, the distributed web services were a yawn for anyone outside the most extreme HTML hairdressing circles, and WinFS was simplydumped from the product during the beta. So those development houses that architected solutions using these promised features ended up with their fingers burnt. Wander around any software emporium and you'll find there's almost nothing on sale that's Vista-only: everything still runs on XP, too, which means from the development community's point of view there was little benefit in Vista, but rather a heap of pain.