information on demand blogs

discussions on information and data integration, information as a service, content management, data services, and business information services.

information server >

Innovation requires speed

September 06, 2006

I was reading the Thomas L. Friedman book entitled 'The World is Flat' this past week while flying around the known universe. There is a chapter outlining the growth and bust of the dot com era which was focused around technical expansion of the fiber optic channels enabling our networks. This innovation enabled the technical landscape to shape and change quickly with information flowing seamlessly delivering deep collaboration and global growth. Even with the bust of the dot com era it was an innovation that helped push us where we are today in our uber connected anti-timezone innovation years.

This overall construct struck a nerve with me as this hits close to home. As a member of the Product Team with in IBM I help to shape the delivery of the IBM Information Server I see this theme throughout the product and its strategy. Deeper collaboration as a concept is not written into just a chat window, an email posting or a speed dial button. Instead it is fundamental and basic as an architecture. Like the fiber optic networks delivering answers as a human we require and desire speedy and accurate information. The world will continue to grow in size and in data and we must grow with it or work more quickly. With the upcoming quad CPU machines, fiber channel networks and disk becoming trivial we can see where the more more more world will keep taking us! Innovation is in the architecture and our architecture must support collaborative, re-usable, simple and excessively performant requirements.

Foot in my mouth?

August 25, 2006

Well, just as I made a statement to my colleague earlier this week that Web 2.0's impact on mainstream enterprise integration was still two years away, some evidence arises in Jim Moore's blog that indicates I may be underestimating the trend. Jim points out that NMG and RSS labs (presumably at Harvard) are working on RSS-based enterprise integration. He also references a blog post by CIO Magazine's Christopher Lindquist that further challenges my theory.

Okay, okay - this wouldn't be the first time I was wrong. I am all for this acceleration if it happens, since I believe that RSS can be an important integration technology. I do feel that semantic-driven integration has real promise in making this real. Of course, the first step is metadata-driven integration, which is where we have initially focused with IBM Information Server. I am very interested in the convergence of conventional metadata (from the information integration, BPM, and content management domains) with tag metadata (either manual tags, or mining-inferred tags) and semantics. We'll have to keep an eye on how this market evolves. In the meantime, I am preparing myself to be proven wrong.

Too much information ... or too little?

August 22, 2006

Our big problem is information overload, right? We get bombarded with so much information in every part of our lives that it becomes impossible to deal with it all.

Well, at a recent meeting of our Customer Advisory Board, we heard of an opposite trend. Many of our customers were struggling with the pressure to capture more and more “atomic” data in their warehouse, so that any business question could be asked. Not just aggregated sales results, but each sales order; not just each order, but each line item; and so on. The reason? In the fast-changing, competitive world, the business users needed to be able to ask questions they hadn’t thought of before, so the data warehouse had to be ready for anything.

I think information overload is a symptom, not a problem. The real problem is not being able to get to the important, trusted, complete information quickly. I’m not sure we’d ever feel like we had too much information if we could answer any question we wanted right away!

Information In Motion

August 16, 2006

Doesn’t it seem like a lot of information is dead on arrival? A customer service request is logged, handled and then closed. Or sales data is collected, reported, but not carefully analyzed. Or a strategic plan is created and then… put on the shelf.

So we’re missing a lot of opportunities - not using demand signals to drive supply chain, not tailoring our marketing and sales effectively, not achieving the process efficiencies we could be, not capitalizing on new opportunities, etc.

Why is that? A lot of reasons – information overload, disconnected systems, organizational silos, incomplete or poor quality information, not enough creative thinking, disconnects between what’s provided and what’s needed, etc.

I’m intrigued by the idea of setting information “in motion.” To me, this means getting the information off the shelf and into the business. Providing trusted information right when and where it’s needed, rather than making it a separate task to go and look up relevant information. I think the opportunity this presents is to actually leverage information more continuously and more effectively – an exciting concept.

Metadata everywhere

A couple of years ago, I remember hearing one of the analysts proclaim that this was the "Year of Metadata". I'm not sure I agreed at that point, but I sure am hearing a lot more metadata talk these days than I ever heard before. After two full-day briefings over the past two days with customers on the West coast, I am finding more and more requirements for metadata in every conversation. So I suppose I would agree that 2006 is probably the year of Metadata (or perhaps 2007) - particularly when you take the tagging phenomenon into account.

In the past two days, I have heard some interesting requirements that, while I've heard them before, are relatively new territory for metadata. One that I thought was particularly interesting was a requirement to rationalize and provide a shared ontology across structured metadata and unstructured metadata. This particular request was interesting since it was coming from a large media company. We've been implementing systems and seeing increasing demand around providing structure to unstructured information - essentially domain tagging based on text analytics, but taking pre-built meta-tags from content repositories and relating them to metadata elements in the structured world is not a common requirement.

I guess where this takes us is the merging of meta-tags and other metadata - to provide a super-view across all types of information, and allow this information to be associated together and organized appropriately. We are doing a lot of research into social tagging, and linking tags to metadata, so this is not much of a departure. I actually don't think it is too far away before we'll see crawlers that can take audio files or video footage and create tags based on the content. It brings up a bit of that big-brother creepiness, but it is also a very important tool for risk mitigation and root cause analysis. So perhaps this is actually just the start of the Decade of Metadata...

ibm information on demand