Gallery

All together at VMware

Posted on 28 Apr 2009 at 12:34

Jon Honeyball returns from VMware Europe 2009 optimistic about the big picture for vmware, and tries not to laugh at other people's desktops.

The other parts of the VMware plan were just as compelling. It has a vision of full application virtualisation and delivery to the desktop, employing technologies such as VDI, and these can be delivered either from local VM servers or out of a private Cloud or a public datacenter Cloud. Mix and match what you need, provision the applications and run them on demand, then blow away those applications whenever you want a clean start. Or cache them onto laptops so they can be disconnected from the network, and keep these updated over public Wi-Fi/3G connections as necessary, ensuring that all user data gets back into a central store.

I was very impressed by the vision that Maritz is pulling together at VMware. In the past, I've feared that there were too many separate strands in the company's operations, and that no-one had a clear vision of a whole company strategy. The arrival of Maritz has exceeded my hopes for such a coherent vision. I've written here several times about how Microsoft went into such a flap when it didn't have all its VM plans in place, but how at last February's VMware conference the Microsoft crew were cocky and confident because they had finally worked out their plans. Well, today they're looking just a little bit nervous again, and with real justification methinks.

Ma.gnolia

Now for some tales of woe. I have to be incredibly careful whenever I gaze upon the steaming wrecks of other people's desktop computers or business networks, because it's far too easy to laugh out loud which is neither fair nor right, and perhaps most importantly is very bad for repeat business. Nevertheless, if a user has a machine fail for a third time that's just bad luck (or misuse), but if they've lost data again in this third implosion then it really is their fault.

A particularly painful example of this comes in the form of Ma.gnolia, a service that stored bookmarks for tens of thousands of users. A lot of people relied upon it for their bookmark storage rather than backing up this information locally, so it was a kind of SaaS or "software as a service", an early sample of a Cloud computing technology if you prefer.

So imagine the level of hurt each of those users must have felt when Ma.gnolia's servers not only collapsed, but also irretrievably lost all their users' data. And those users probably didn't feel any better after watching the video of the two enthusiastic young entrepreneurs behind the service: to view it yourself, head to http://ma.gnolia.com.

There are many squirm-worthy moments, but chief among this must be when they were asked: "Did you do any backup testing?" "Nope" came their answer. Their choice of hardware (Apple Mac minis) was a little baroque shall we say, but there's no doubt it could have worked just fine if handled properly. Baroque is probably being unfair actually, because some datacenters have found that Mac minis make ideal rack servers (and isn't there a push by Microsoft now to use low-power Atom processors in datacenter hardware?).

But this episode brings in wider questions concerning SaaS. If you make SaaS a key part of your business, have you fully assessed the SLA and the recovery tools available? I'm not suggesting that the likes of Amazon, Google and Microsoft treat data recovery with the same approach as Ma.gnolia, but ask yourself some tough questions: just how much care do you think Google et al are going to take of your data?