And Then There Were Two

Deuces are wild in the technology business. Today, there are two emerging Internet software platforms: Java and .NET. And, legacy platforms aside, there are two OSs competing for development of new applications: Windows and UNIX. Years ago, when UNIX tried consolidating, there existed for a while two general flavors of UNIX.

Now, add to that two flavors of Linux. In an announcement that was over six months in the offing, four distributors of LinuxCaldera, Conectiva, SuSE, and Turbolinuxunveiled UnitedLinux, an initiative that will standardize the basic kernel based on features approved by the Free Standards Group, a nonprofit organization of open source Linux developers. For now, UnitedLinux will focus solely on server, rather than client-side Linux.

In practical terms, each UnitedLinux vendor will issue a standard kernel CD, plus separate CDs for their own tools. Theoretically, customers could mix and match Linux vendor-specific tools to any compliant kernel distribution.

For the gang of four founders, converging kernel development efforts make sense since there is little money to be made selling something thatís already available for free download. The real business in Linux isn’t Linux itself, but the support, applications, and appliances that surround it.

But a major fly in the ointment is that group founders, which represent the dominant players in Europe, South America, and the Pacific Rim, left out the big name in North America: Red Hat. Whatever the rationale for omission, UnitedLinux doesn’t need the distraction. Just look at WS-I, whose viable work defining web services interoperability criteria is being obscured over the silly soap opera of whether Sun will be allowed in.

In so doing, UnitedLinux members could unwittingly confirm Red Hat’s dominance by isolating them. Which brings us back to those technology deuces. When it comes to Linux, there will be Red Hat and everybody else.

Open Enigmas

The term “open” has become so politically charged in the technology world that no vendor dares omitting it from their collaterals. Defined as technology complying with public standards, open standards have taken over virtually every facet of technology, from software development and deployment to process modeling, data management, network protocols, and data rendering.

Open standards are the reason that information technology has become more valuable to enterprises computing. Although hardware and software do more things today, the real benefit is that standards have eliminated needless legwork, like reinventing communications protocols or markup languages. Standards made the Internet, arguably the most valuable corporate computing innovation of the past decade, possible.

But standards aren’t ends in themselves. For instance, if you’re like most IBM WebSphere users, you’re probably using an older version that isn’t J2EE- compliant. To become compliant, you would probably have to make extensive changes to the innards of applications that wouldn’t necessarily add functionality. Maybe in the long run J2EE would ease tasks like enterprise integration, but in the short run, if you’re not doing an integration project, could you honestly make the ROI case to migrate?

The flip side of the argument is whether it’s cost effective to NOT implement standards. The challenge is that Web and mobile apps lack the transaction performance or protections of internal apps, areas that standards don’t currently address. Not surprisingly, we’ve seen vendors using proprietary technologies in different ways for attacking different parts of the problem.

* Akamai uses proprietary technology to reroute incoming requests for popular consumer sites like 1-800-Flowers to satellite servers for faster performance.

* Kanemea, adds its own connector technologies so web transaction applications gain transaction processing features like offline processing and rollback.

* SavaJe superimposes its own OS to run Java applications better on mobile devices.

* FineGround networks confines its proprietary technologies to algorithms that figure how and when to send IP messages the standard way over the Internet to reduce bottlenecks. It avoids doing anything proprietary to client or server.

Although these technologies are apples and oranges, some parallels still apply. Akamai is the exception that proves the rule because its customers are large B2C household names that cannot avoid losing market share, period. But most corporate deployments can’t afford the expense of deploying nonstandard technologies outside the firewall because of all the maintenance headaches. It’s a basic lesson that even Microsoft with Active X, had to learn the hard way. Instead, we like FineGround’s approach better, because it doesn’t require anything weird on the outside.

We’re just surprised that other vendors still haven’t caught on.

Yeast Rises

Now that a Delaware court has cleared the way for HP and Compaq to combine, we looked back at the track record of technology industry mergers. If the meltdown of WorldCom or the proxy fights at CA are any indication, prospects don’t look so hot for HP and Compaq.

But is organic growth a better alternative to M&A? It’s tempting to hold up Microsoft and IBM (both of which have grown with relatively few acquisitions) as shining examples, but a closer look reveals that reality isn’t always so black and white. Let’s glance at a few examples.

* IBM. OK, post-Enron, its numbers might look a bit aggressive. Still, it enters this decade in much better shape than the last. Was that due to the strength of e-servers, consolidation of global consulting services, or the acquisitions of Lotus and Tivoli? Little mystery here, servers and services met the need of large enterprises for strategic technology paths and investment protection — something to which competitors like HDS could only tactically respond. Lotus and Tivoli proved mere footnotes.

* Sun and Oracle, which largely emphasized organic growth, avoided the excesses that pushed CA and WorldCom to the ropes. While internal growth preserved focus, it kept their cultures too inward-looking, problems that we believe are fixable.

* CA, whose growth by acquisition was purely revenue-driven, succeeded only as long as the overall economy grew. The company has numerous problems — lost opportunities from acquisitions among them. For instance, the application development and data warehouse businesses of Sterling and Platinum largely evaporated because CA didn’t know those markets.

* HP and Compaq. Here’s an acquisition that has the real vision thing. The only problem is the vision itself. Dominance in PCs, printers, and accompanying maintenance services won’t deliver the margins or solutions necessary for a company that will soon rival IBM in size.

For technology vendors, the M&A vs. organic growth debate boils down to common sense. Growth based on pure numbers is like yeast, rising only while the oven’s still warm. Growth based on strategy has better prospects, but obviously, only if the strategy makes sense. Whether growth comes via M&A or internally is incidental.