Software Chief Gives IBM a Vertical Push

In an eWEEK interview, IBM Software Group executive Steve Mills weighs in on the company's vertical initiative.

Although the IBM Software Group generates billions of dollars of revenue for IBM each year, it is a nimble multibillion-dollar division. Responding to an evolution in the buying patterns of its customers, the Somers, N.Y., group late last year pushed into high gear an initiative to embrace vertical-industry customers through comprehensive middleware solutions.

Steve Mills, senior vice president and group executive with the division, met with eWEEK Editor in Chief Eric Lundquist, Executive Editor/ News Michael R. Zimmerman and Senior Editor Darryl K. Taft recently in IBMs Armonk, N.Y., headquarters to discuss the vertical push and its results.

/zimages/5/28571.gifFor more on IBMs vertical strategy, read "IBM Scales the Vertical Industry."

Why the vertical push?

Its perfectly logical in terms of what the customer is doing. The buying patterns and interest and behaviors of customers is something you always have to be watchful of. And what were trying to do is to ensure that the appeal of our technologies, which is a statement of what the technology is, combined with how we bring it to market, matches the way that customers are thinking about their business problems and how they go through their eval processes for choosing technology. And so the affinity of our middleware infrastructure technology to line-of-business needs and requirements—the industry-specific needs and requirements—is a critical appeal point.

In a logical sense, you dont want to leave yourself at the bottom, if you will, of the technology ladder, because thats invariably where the commoditization is occurring. You need to continue to move toward areas that businesses see greater value add, more connection to the business problems that they have today, and those that they see as critical long term. You know—strategic, transformational, needs requirements and issues. And weve been on this path now for some time. Our middleware strategy has been one of a focus on satisfying the need for biz process integration, you know, tie the value chain together, this whole notion of connecting the customer through your internal processes, through your supply chain. Real-time enterprise, synchronized commerce, straight-through processing. You hear lots of terms out there that different businesses in different industries use to describe the way they think about the future state of the way they want their IT environment to operate. And the way you want to support the constituencies that they have to support in their business.

And delivering on that is more than tossing a bunch of pipes, if you will, into the mix, and saying we have great low-level infrastructure and we can flow transactions. And you know, were good at that, and yes, we have that technology. Theyre looking to solve the problem not just from a raw transport layer notion of connectivity and interoperability but also, hows it touch on the way the business processes have to run? Are you mapped into the kinds of applications that are typically deployed within a particular industry? Do you understand what the process flows are? We talk about collaborations, and we think about the way in which a process step goes from A to B to C to D.

In a sense youre having combinations of human beings and applications collaborate with each other. Sometimes its people collaborating, sometimes its actually applications collaborating. You know, sharing data, passing an element of a transaction from one system to the next to the next. And so all that mapping of one thing to the next, and the incorporation of interfaces and standards that are unique to different industries. So you want to sell into companies with big back-end manufacturing supply chain environments. There are predefined standard interfaces for these things, to deal with their EDI environments or how they do replenishment. And the big players often describe their own interface structures. And most participate in ecosystems that are broader than their own company.

How fresh is this focus on the unique requirements of the vertical segments?

We started that in the late 90s. Our perspective was the value of the middleware rose as you did a better job of integrating with the way in which a business runs within an industry.

Certainly IBM has over the course of decades delivered various pieces of function that were designed to help banks, for example, deal with Fedwire or swiftwire service. Clearing system connections. Weve tied into traditional proprietary EDI environments. A lot of this has been small bits of code, custom crafting kinds of things. Our perspective was that with Internet standards sweeping the market that this was bringing about a profound renovation of old-style proprietary environments into newer more open standards driven environment and that we could actually establish market leadership by adhering to these standards. So in retail, its uccnet. With Wal-Mart, its their AS2 interfaces. You have things like HIPAA [Health Insurance Portability and Accountability Act] data structure compliance in the insurance health-care industry. You have different things going on in financial services related to BASEL II, or a traditional wire connection for money clearing, payment systems. There are lots of them. You go through every industry, and there are lots and lots of them.

I think we started to get peoples attention in this area as we rolled out the WebSphere Business Integrator portfolio of offerings. The acquisition of Crossworlds was obviously important to adding more functionality, more mappings and adapters. Holosophix for business process modeling. A lot of this requires tooling around it. How do I instantiate this process around some set of rules? Model the process. Test the process. Make sure things work as I expect from either a functional perspective or from an economic value perspective. And then monitor the activity as its occurring.

So is it fair to say the strategy has been evolving, but more rapidly over the last couple of years?

Oh, theres always an acceleration as you have more and more pieces you can do more and more things. The nature of technology is always, while its hard to get the base in place, the establishment of the deepest technologies down at the bottom are often a big challenge to get a solid baseline. And then from there, as you begin more pieces, you begin to discover you can adapt to different needs and requirements faster than you ever thought. You win more customer engagements, you discover more things, you build more stuff, you look at more acquisitions.

This is our view; we believe the world keeps moving and we have to keep moving. We cant be static and rest on our laurels. And so we have a vision of continuing to widen out our portfolio. And there has been an acceleration. And the acceleration has been brought about by the wide variety of technologies we now have to meet these needs. The number of engagements we do.

And frankly the fact that the customers are very focused on this as a key aspect of the way they think about transforming their business, improving their ratios. Whether you use our on-demand term, or others descriptions of real-time enterprises, these ideas do resonate with businesses. And its not hard for a c-level executive to say, yeah, velocity is really important. If I could process things faster with fewer steps, fewer handoffs, fewer interruptions. If I had more knowledge of what was happening in my business as things were actually being processed and executing, if I could relate external factors to internal systems factors, I would probably be able to improve my efficiency. Id have better ratios. If I had better ratios, Id go to Wall Street, and Id drive for higher share price.

We used to think of the horizontal, broad-based platforms as the turf of IBM. Vertical activities were always the part of IBMs partners.

Oh, for sure. I mean theres a big gap between the middleware and the applications. Lets not confuse these things. Were not delivering the application logic, the logic thats required to do your general ledger, to do your actual production planning, supply chain planning. All the things you do from an application perspective, were not delivering that technology. Were partnering with others. But as the customers continue to push on integration as a core element of any application decision, you find these things having to dovetail with each other much, much more than they did in the past.

How are your Web services initiatives resonating with your vertical customers so far?

Web services is not a mandatory requirement for integration. What Web services are all about—we love these industry euphemisms—Web services, whatever that means. So, XML is a nice thing. Its a 40-plus-year-old idea. So, how do I separate content from structure and tag information so that I can use the tagging definitional mechanism as a common way to define everything? So its very much a lingua franca kind of notion. And so [its] a good thing to do. ...

It does not in any way define the quality of service. It doesnt define performance [and] doesnt define the robustness and the recoverability, and there are many other aspects you have to look at. But if life gets easier because I have common interface structures brought about by Web services standards, then thats a good thing.

[But] its not the universal solution to every problem.

Not right now.

Nor will it ever be. It will never be the universal solution to every problem because no tech ever is. Every time something new comes along, this is an industry that loves the next new thing. We go crazy, wow, this is great, itll transform the world. And everything will be different, everything will be simple and wont this be wonderful? Right? But the hype, its like the Gartner hype curve thing … here you have something thats very useful, and then you over-hype it. And then what happens is it drops into the valley of despair. The hype can never be fulfilled. No thing in the world can ever be that good.

Its an unfair to place on any technology. And so the expectation is unfulfilled [and] you read all kinds of articles about the disappointment of this and it didnt do what it was supposed to do. And I read an article that said that applications were going to appear spontaneously in the universe all made up of little Web services components and things would magically come together. And the world would be a wonderful place. Its candyland. I mean, how ridiculous.

The whole notion is to keep improving on what it is you have. And these things are additive, not subtractive. It takes a long time for things to literally be subtractive from the infrastructure out there. Its mostly an additive world, and new things come along that are a help to getting your job done a little faster. And standard interfaces and the Web services notion are part of a longer march of standards. I think today most businesses would say that versus a decade ago, they are much better off as far as standards. There are standards for connectivity, standards for displaying information, standards for exchanging formats and so on—profoundly better than it was 10 years ago.

Now, youd say, given the way it could be, we have a long way to go. And thats true. We do have a long way to go.

But its a bit like living in New York City or Boston. These cities have been around a long time. And not everything works perfectly. There are potholes in the road, and stuff breaks. But thats the nature of infrastructure. When its brand spanking new, its usually brand spanking new. And then it ages. And you have to go in and renovate, and patch and re-splice and add on. Sometimes you get embroiled in the Big Dig.

Look, Web services inherits from obviously HTML, SGML, GML—these ideas of tags and scripting approaches have been around for a very long time.

/zimages/5/28571.gifCheck out eWEEK.coms Enterprise Applications Center for the latest news, reviews and analysis about productivity and business solutions.


Be sure to add our enterprise applications news feed to your RSS newsreader or My Yahoo page