Pender's Blog

Blog archive

Another Bad Day for the Cloud: Microsoft Office 365, Dynamics CRM Go Down

Darn it, Microsoft, you're going to ruin the whole thing. Well, you won't be alone -- after all, you're certainly not the only vendor to suffer an outage of cloud-based applications lately. But every time the sky falls and the cloud goes down, it hurts the whole cloud computing model.

This week's outage seems especially bad, not so much because of how long it lasted (although it was hours, apparently), but because of what actually went down. Office 365, Microsoft's long-awaited, almost-brand-new online productivity suite, was part of the crash. Office 365 is meant to be the successor to BPOS but only seemed to live up to the last three letters of its predecessor's name this week. Oh, and if you wanted to stay on BPOS and not move to Office 365, too bad. Office 365 is coming to BPOS users -- if Microsoft can keep it up and running.

The other piece of Microsoft's cloud that came down -- in an unrelated incident, apparently, although we can't decide if that's better or worse -- was Dynamics CRM Online. Now, granted, that might not be Microsoft's biggest hosted offering and it doesn't necessarily have the same impact as Office 365, but Dynamics CRM Online is one of the few facets of the (actually very good) Dynamics ERP and CRM offering that Microsoft actually seems to be actively trying to sell. While the ERP suites kind of gently rock along in Redmond, Microsoft is pushing Dynamics CRM big time.

And the hosted version, Dynamics CRM Online, is definitely one of the company's high-priority hosted products and is unquestionably the main attraction among hosted Dynamics applications. CRM is high-profile stuff by Dynamics' standards, and it looks bad when Microsoft can't manage to keep such a hard-sell hosted offering consistently available.

(As a side note, but not an unrelated one, a company called Nimbula had great timing in announcing this week its Nimbula Director 1.5 product, which lets administrators tie together geographically distributed clouds with a single set of permissions and policies. The infrastructure-as-a-service platform, as Nimbula calls it, is big with cloud service providers. Run by some ex-Amazon employees, Nimbula is already serving customers that want to be mini-Amazons.

"The interest has been heavy from service providers," Jay Judkowitz, director of product management at Nimbula, told RCPU last week. "They see what Amazon's doing and they say, 'We need to do that, too.' They're not saying we have to compete with Amazon. They want to do Amazon-like things but not compete with Amazon."

Of course, Amazon has had its share of problems with outages in the past, but what Nimbula is doing is pretty interesting nonetheless. End of digression.)

As is always the case, the anti-cloud hordes are hitting the blogs and message boards talking about how this latest hosted-software problem is just another reason why the cloud is a waste of time and money and doesn't really work. At this point, though, who's to say that they're wrong? OK, so the outages we read about might be drops in a bucket, maybe even a lake, compared to the hours and days and weeks in which cloud services run smoothly and work just fine. But drops make ripples nonetheless, and they make bigger ripples when they fall into high-profile cloud offerings. What these outages really do is show that cloud computing isn't ready to handle business-critical applications, and that in turn relegates the whole cloud model to fringe functionality within most organizations.

We at RCPU have always been supporters and proponents of the cloud. It promises cost savings to organizations and seems like a clean solution compared to the labyrinth of systems and applications that's often hiding inside corporate IT departments. We're starting to wonder why we like it so much, though. The cloud can be a real mess for partners who are trying to sell and make money off of it. It's a potential threat to IT professionals' jobs. Security and (obviously) reliability remain major concerns. And we can confirm that if your editor couldn't get to his e-mail or some other critical application for several hours during the workday, you'd hear the unprintable screaming coming all the way from Greater Boston, no matter where you are.

We in the punditsphere are always looking for the next big thing in technology (or the next big failure -- thank you, Windows Phone 7). As exciting as the industry can be, covering it every single day can actually get to be a grind. When a brand-new (actually old and repurposed, but that's another discussion) model like the cloud comes along, we tend to jump on it and hype it as much as we possibly can. Then, of course, we have to wait for it to fail a few times so that we can tear it down and talk about how it was nothing but a creation of hype, anyway, and really isn't all it's cracked up to be. Hey, it keeps people reading...we think.

And so we're in the tearing-down phase with the cloud -- although it's the service providers who are having the outages and providing those of us in the punditsphere with the fodder for destroying what we helped build. The truth about the cloud, of course -- as most partners and IT people already know -- is somewhere in the middle. Yes, it can be a cost saver. Yes, it's good for certain functions and applications. Yes, service-level agreements and choosing the right provider are massively important. And no, despite some of the success stories we've heard in recent years, it's probably not time to shift all of your business-critical applications to the cloud. It might never be. Maybe if we go years without an outage, if stories like this week's Microsoft blunders sort of fade away, the cloud will become the norm and nobody will think twice about outsourcing to it. But we're not there yet. Not even close.

How much do you trust the cloud computing model? What are you willing to outsource to the cloud? Tell your stories at [email protected] or leave a comment below.


Posted by Lee Pender on August 18, 2011