Petabyte Prognostications EDS's futurist is preparing for the coming data flood with context-sensitive text-mining tools. Future Watch by Gary H. Anthes
			OCTOBER 25, 2004 (COMPUTERWORLD) - Jeff Wacker's job as Electronic Data Systems Corp.'s futurist is to develop companywide initiatives that will shape the future of EDS. He recently told Computerworld's Gary H. Anthes why mobile workers, unstructured information and communications infrastructures are worthy of special attention in the coming years. Is there an information explosion coming? Yes, and it's based on two major factors. First, there's all this sensor and RFID information that's starting to flow into corporations, and it will only accelerate; Wal-Mart is looking at 5TB to 7TB a day. The other factor is that unstructured information makes up 80% to 90% of the average corporation's information content. It's not in a form computers can readily use, such as e-mail. We are using context-sensitive text mining as a tool for structuring that content. When you do that, all that information becomes a corporate asset. How might that new asset be used? Most of the new information that allows you to predict the future is nontraditional corporate information -- what we call indicator information. You have transactions, which are past; you have events, which are current; and you have indicators, which are not traditionally used in business. You can put that information in a data mine and after the fact try to figure out what you should have done. Or you can feed it into a cause-and-effect model ... and use pattern-recognition technology. [That gives you] the ability to understand patterns of business activity that are going to repeat and say, "What do I want the outcome to be?" Can you give a couple of examples? Weather predictions for a warmer-than-usual winter in New England change the model's probability of selling x fruitcakes to y. That, coupled with a colder-than-normal forecast for the mid-Atlantic states, drives a directive to channel the fruitcake to those states. EDS futurist Jeff Wacker sees toys as an analogy for rapid change, since “kids turn toys at an astonishing rate.” This one is made from old IT components, demonstrating “that we create the future on the bones of the past—but different from the past.” Or an erroneous newspaper article reports that a certain food product has now been linked to heart attacks. In every city where that article has appeared, you can expect the demand to fall off precipitously, and therefore you need to "reverse-logistic" the existing supplies of the product to areas where the article has not run. What are some opportunities for IT in the mobile world? Mobile workers are the orphans of the information revolution, because IT has basically ensconced itself in a stationary environment. Mobile workers work asynchronously with their IT. The office worker works the problem with the computer, but the mobile workers for the most part get their information, go out and do their work, come back and report their information. The IT is not overlaid directly on top of their work. We have built a testbed called process mentor. It's the ability for you to deliver just-in-time information to the mobile worker. Information flows, unprompted in some cases, as you need it. Can you give an example? In the petroleum industry, 70% of workers are in the field. The average age of a field worker is 55, and 50% of them will retire in the next five years. That's a hell of a lot of expertise not easily replaced. So if I'm out at the wellhead and I ask, "How do I take this thing apart?" all of a sudden I have just-in-time training that tells me exactly how to do it. It says, "Here's Step 1 -- now do it. Now here's Step 2 -- do it," and so on. It's the opportunity to put IT into the business process rather than work around the business process. Where will we see sensors deployed? That airplane that's flying has sensors that tell me that a certain part is going to go out, so I can fly the part to the destination before the plane gets there. That's the kind of model that's going to be deployed throughout all of business. All the items coming into my supply chain, all the items going through manufacturing, all items going out to the customer and all the people in my organization are going to be sending me information. All that extra data must have huge implications for storage systems. When we start sending petabytes of information from one company to another, it's cheaper to have a petabyte storage cube put on an airplane and fly it to the other coast. That will be a constraint in the future, because petabytes of information don't work well on our communications infrastructure. What's the solution to that problem? The amount of information that can be stored on a hard drive doubles every year ... and fiber-optic capacity is doubling every nine months. While it will take some time, this will overcome the problem. Also, you can break the file into many pieces, transmit the pieces and put it all together on the other end. The problem here is the infamous last mile. Most companies do not have [networks] to allow them to send the data through multiple pipes from a single source and to a single destination. This is an infrastructure design problem that never contemplated needing many T3s to the same point of origination/delivery. Profile Name: Jeff Wacker Title: EDS fellow and futurist Company: Electronic Data Systems Corp. Background: Wacker is chief technology officer for EDS's Global Industry groups. He earned a bachelor's degree in computer science and an MBA from the University of Nebraska. He is a member of MIT's Society of Learning.