informa
/
Commentary

Will IBM Go 'All In' On Cloud Computing?

IBM has spent more than a decade and tens of billions of dollars buying software. Don't bet on it delivering its smartest technologies on demand any time soon.
IBM's private-cloud offerings include infrastructure software, like WebSphere and System Director VM Control, as well as lots of hardware options, including BladeCenter, Power System, System z mainframe and network attached storage options. It's all about building out flexible capacity you can serve up on demand from your own data center.

This is what the cloud purists call the false cloud; you're not tapping into somebody else's compute capacity on an as-needed basis. You can have IBM run the technology for you in an IBM data centers, but you're still buying servers and infrastructure software.

IBM also is pushing to get companies to move software applications into hosted deployments, but this is mostly about software from other vendors, including SAP and Oracle apps that are routinely integrated and deployed by IBM Global Business Services.

When it comes to IBM's software beyond infrastructure, there are targeted SaaS-delivered tools including LotusLive (email and calendaring), Blueworks (process modeling), Coremetrics (Web analytics), Cast Iron (integration), and Tivoli and Rational IT components. But what you don't hear much about is how the high-end analytics and optimization technologies are going to come to life in the cloud.

I've heard snippets from IBM, here and there, about big-data-analysis applications in the cloud, like genomics research. And I've reported on text-analysis services handled outsource-style from IBM GBS service centers in India.

But what's the big-picture strategy for delivering analytics capabilities from the Cognos, iLog, and SPSS portfolios via the cloud? I haven't seen anything on that front from IBM. Meanwhile, analytics rival SAS is pressing ahead with a number of subscription-style offerings (albeit in a hosted model, without options for rapid scaling or anything shorter than annual contracts). And SAP BusinessObjects has been running its BI OnDemand service for several years. IBM also has integration and consulting services rivals that are gearing up to deliver analytic services in the cloud, as described by Greg Todd of Accenture.

IBM says it foresees $16 billion in revenue coming from business analytics and optimization by 2015 -- more than double the company's prediction for cloud-computing revenue. But I have to wonder if that estimate is based on analytics and optimization being delivered primarily through on-premises deployments, or whether IBM foresees its "smarter" solutions moving heavily into cloud delivery modes? I suspect it's counting on a long window of continued on-premises deployments (just as SAP and Oracle are for their enterprise applications).

Sam Palmisano himself told us in 2009 that the two coming waves in technology would be business analytics and cloud computing. But he didn't have much to say about the prospect of those two waves crashing into each other. When analytics and cloud collide, all that intellectual property currently packaged up and sold as software will need to more flexibly and easily accessible as services.

In some areas the collision has already happened. Marketers, for example, are used to SaaS-based Web analytics and analysis tools; IBM's Unica and Coremetrics units were already moving into the cloud when they were acquired, so there they'll stay.

The bottom line in all this is money. If customers can get a better deal by moving to cloud-based services, they will. But incumbent vendors aren't going to put it out there until they have to. It took Salesforce.com to drag on-premises CRM vendors into the cloud, and it took Google to drag Microsoft into cloud-delivered email and collaboration. The downside for on-premises vendors is that when the competition heads into the cloud, profits tend to suffer. When Wall Street analysts see software turning into services, they assume it'll mean lower margins.

As mentioned, there are a few options for analytics in the cloud, but no single vendor has taking the market by storm -- yet. No doubt IBM will be right there in the fray when the competition heats up. For now, I wouldn't expect any all-in declarations from the top of IBM.