A few months ago I got one of those semi-spam e-mails from a market research company acting on behalf of a major auto manufacturer (not named, but it was clearly GM) asking me to participate in a survey about some aspect of engine technology. It turned out that they wanted to know if the use of overhead valve technology, instead of overhead camshafts, would influence my decision on buying a car. Duh? What do I care? I buy a car as a means of transportation, with the critical selection criteria being safety, ease of operation, value and style. I could care less if the engine uses OHV, OHC or rubber bands, so long as it works as advertised, whenever I need it. What I'm interested in is a safe, reliable, cost-effective transportation platformand if it's fun to use and looks good, so much the better.
One of the most pronouncedbut not necessarily best understoodshifts in the current information technology landscape is the steady evolution away from a focus on product and toward the idea of a "platform." For most of the history of corporate IT, CIOs and their staffs have been asked to acquire, assemble, deploy and support a growing portfolio of products of various typeseverything from mainframes, servers, PCs and storage systems to multiple layers of software and attendant management utilities. Fifty years of rapid innovation have given CIOs plenty to choose from, and the slower, but steady, evolution of standards has promoted improved interoperability and eased integration efforts. Even so, the results of generation after generation of product evolution in hardware and software, and the integration architectures each generation has spawned, aren't exactly pretty. Complexity and fragility are common companions in the resulting business automation infrastructure.
There have been periodic and partially successful attempts to fix this. The major technology vendors have tended to work to make their products either "bundled"the vendor does the integration prior to purchaseor better aligned so that assembling the components after you buy them is straightforward. "Configurators" make sure that all the right components are ordered, and that the specification is (at least approximately) appropriate to the need. Components come together to form subsystems; subsystems combine to form systems; systems link to form "solutions." And solutions integrate to form a cohesive business automation footprint. What's not to like?
The good news is that we are moving steadily toward a focus on what's possible from a business sense, and away from a technocratic "speeds and feeds" mentality that focuses on the (often impressive) technical characteristics of the things we buy. The bad news is that many vendors still push speeds and feeds (which they are comfortable with) ahead of business capability and potential (which are often foreign to their sales forces and marketing departments, among others).
Before the hate mail pours in, let me state that this relentless technocratic voice manifests itself in different degrees in different sectors of IT. It's most common among smaller vendors whose founders are excellent technologists and justifiably proud of what they have built. But it manifests itself in larger vendors, too. Just look at the fascinating shifts in message mix that have taken place in technology advertising over the last few years, as exemplified by sustained campaigns from Microsoft Corp. ("Your potential. Our passion" supplementing Windows and Office product ads), SAP ("The best-run businesses run SAP"), IBM Corp. ("e-business" and "On Demand" evolving into the very clever "Help Desk" series) and Hewlett-Packard Co. ("Invent"). IBM, in particular, has become somewhat confused in its mix of hardware/software focus and business capability focus as it tries to grow away from its hardware and software roots.
And then there is arguably the most successful and sustained technology brand exercise of the past 50 years"Intel Inside." That's ironic in many ways, because Intel Corp., as much as anyone in the technology industry, has been at the forefront of the steady shift away from "components" and toward platformsfrom the bottom up. Microprocessors grew to subsystems, to systems on a board or to systems on a chip, thus allowing Intel to capture more and more of the total value and, not accidentally, to accelerate the take-up of new technologies. Even though Intel doesn't make every kind of subsystem and system, it does develop and publish comprehensive reference designs and promote standards, both of which allow its OEMs and channel partners to get sophisticated products to market faster.
Now let's take a spin through the world of _the CIO and the businesses they serve. CIOs of the largest companies have become extraordinarily pragmatic over the past few years. Their core mission is to make the technology they oversee disappear from the thoughts of their users (and their management teams, boards of directors and shareholders) as fast as possibleto create what I call "invisible infrastructure" that "just works," anywhere, all the time. CIOs in smaller companies, who are generally better served by the mainstream product mindset of their vendors, aren't there yet, but signs are they are heading that way, too. The root cause driving this shift is, in my view, the tenuous correlation between IT investment and business value. And for consumers, the shift is even more markedand best typified by the (so far farcical) attempts to promote a ubiquitous digital home infrastructure.
Vendors undertaking this evolutionary shift face a number of hurdles. First and foremost, they must reconstruct their messages to the market. They have to figure out a whole new message strategyand economic modelif they are to maintain traction within a market that just doesn't care about the internals, only about what the internals make possible. They have to turn their message aroundtell the story from the customer's viewpointby creating recognizable, relevant scenarios that their offerings can satisfy. There are some early and interesting _examples of this (UPS's "What Can Brown Do For You?" comes to mind), but most attempts are less than encouraging.
Second, and almost as difficult, vendors have to figure out what will differentiate their offerings from everyone else's in a platform world. What will be the key buying differentiators when speeds and feeds don't matter? Reputation, reliability and quality, interoperability, adherence to open standardswe can already see new differentiating characteristics and potential brand values opening up.
And third, there can probably only be a few successful platforms if these new differentiating characteristics take hold. Just as the global auto industry has a glut of generally undifferentiated brands, IT has a glut of undifferentiated "product families" at all levels of the business automation stack. We can see that both J2EE and .NET have platform characteristics at the lower levels of the stack, but what's going to emerge above them? SAP, with NetWeaver, clearly has ambitions in this direction, as does Microsoft with MBS. Will everyone else have to pick a platform to be part of? Or will it be possible to actually become a platform in your own right if you see the user's need clearly enough (as RIM is trying to do with BlackBerry, and Palm arguably has failed to do with the PDA)?
However this comes out, the shift to platform looks inevitableand necessary if we are ever to get to "it just works" in business and at home. The train is leaving the stationif you're not on the platform, you are in danger of missing it.
John Parkinson is chief technologist for the Americas at Capgemini. His next column will appear in August.