What were the odds we'd hear all about analytics at this week's IBM Information on Demand (IOD) conference in Las Vegas? I wish I could have placed that bet, as it seems analytics is all IBM is talking about these days. (I agree with Neil Raden's "whatever that means" comment in this blog, which suggests that the term is ill defined and over used.) "Analytics" was in the very name of two out of four new products announced. Here's my quick take:
"These applications are model-driven, maintained and 100-percent supported and upgradeable," says Mychelle Mollot,IBM's vice president worldwide marketing, analytics and performance management. "You customize by making changes to the model, and those changes are tracked and stored as part of the description of the application. When you upgrade, all of your changes and customizations are captured and recreated."
As I reported in this article, at least one analyst sees ongoing maintenance and support as a differentiator (with Oracle and SAS being leaders in this area). So here is IBM doing what it takes to be a supplier of "full-fledged" applications, as Mollot describes them. She adds that IBM's apps run against multiple ERP backbones, so they are "not locked into an SAP or Oracle environment," though I suspect the other guys let you bring in data from wherever, too. The new sales analytics, supply chain analytics and talent management analytics apps are already generally available.
IBM Cognos Content Analytics software. As described in this story, this is about analyzing documents, e-mail messages, blog posts, forum comments and, yes, even Twitter tweets to make sense of the latent meaning in the masses of content. The core tooling, to be released by the end of this year, will be used by customers to build text applications. Mollot says IBM is using the same technology to develop specific solutions such as customer sentiment analysis. There also will be cloud-based delivery options (something Attensity already offers).
The two non-analytics related announcements at IOD both relate to the InfoSphere family:
InfoSphere Streams. Released a few months ago, this stream processing capability is just emerging and I received sketchy details. Apparently it has been upgraded to support the use of PMML (predictive model markup language) in scoring streaming data. Sounds like you can make use of what you've already developed to score data in flight, but I'm unclear on the mechanics or examples of streaming analysis being targeted.
InfoSphere Master Information Hub software. There are plenty of prebuilt MDM offerings for customers, products and so on out there, but this new software helps you develop and govern your own, customized data sets. If you want a master data set for events, for example, this helps you create a custom domain with governance and stewardship supported by the master hub.
I'm surprised we're not hearing about a solid-state disk based data warehouse appliance from IBM at IOD. I'm guessing IBM is keeping its powder dry for now and planning a response at a later date. Now that Oracle and Teradata have made their plans public, IBM can go back to the lab and find ways to one-up the competition in the next release.What were the odds we'd hear all about analytics at this week's IBM Information on Demand (IOD) conference in Las Vegas? I wish I could have placed that bet! "Analytics" was in the very name of two out of four new products announced at IOD. Here's my quick take: