I've said it before: Microsoft should buy BEA

The dark clouds over BEA are deepening with news that activist Carl Icahn is buying up tangible influence in the software vendor (probably with the idea of a sale). BEA faces lingering unknowns on its financial statements -- not to mention entrenched concerns about sales, growth and its ability to compete against Microsoft, IBM and myriad open source alternatives.

The Icahn news bodes badly for a healthy, wealthy and independent BEA. Whether through an outright sale or false-value run-up of BEA's value and then abandonment at a false peak (remember Time Warner and Motorola), the status quo won't hold.

A few years ago when BEA was at another crossroads (like, most of the technical management tier walked out), I suggested that Microsoft should buy them. It was a novel declaration, and some snickered. Others found surprising logic in the argument but for the Microsoft religion thing around Windows Everywhere.

I think the idea of a Microsoft-BEA mashup deserves merit more now than ever. And it has more to do with Microsoft than BEA, largely because Microsoft has changed with the times in the past several years while BEA has mostly stayed the course, just renaming theirs a SOA value while voraciously resisting the middleware moniker while selling predominantly ... middleware.

As we see Microsoft truly surrounded by vipers -- despite its huge strength, influence and tenacious installed base -- it recognizes that the effects of open source on software economics is not going away. Nor is IBM. And it also knows that advertising revenue -- even at the clip that Google assembles and attracts it -- will not make Microsoft live as well as it has the past 10 years for the next 10 years.

Microsoft is not in any where near the trouble that BEA is, but BEA's history may be worth studying -- for it could be a fast-forward version of Microsoft's.

Windows Everywhere used to make great sense (and cents), but the key to remaining an IT supplier growth engine for the global 2000 enterprises and legions of SMBs is not, alas, Windows Everywhere. IT departments everywhere have a different goal than to make Microsoft their Daddy. They need to continuously cut costs while providing better services, more interoperability, and to elevate agility to the level of malleable business processes. The divergence between Microsoft's business goals and the goals of its core customers is wide and growing.

IBM gets this. HP gets this. Oracle gets this. The global SIs get this. Microsoft could get it -- and perhaps solve it substantially -- if it bought BEA now and became a true enterprise IT solutions provider; if it took the best tools and a great framework and made them the overlay to developing and managing the best combination of mainstream IT architectures, platforms and frameworks.

What's happened in the past five years is that the development, ISV, and hosting communities have divvied up into separate camps around .NET and Java. Many need to try and play nice together, but it's not been easy. It's expensive. And not only is it technology that separates these camps -- they think differently. They see the role and use of open source much differently, and they view costs of IT differently, as in where should the costs be? An astute observer mentioned to me yesterday that .NET developers are from Venus and Java framework developers are from Mars. And so it goes.

But with a deep shift on the part of Microsoft toward really embracing and extending the Java side of the universe, then we're talking about dominating not just the .NET world, but perhaps the Java one, too. Even better, how about becoming the best way to leverage the mainframe, client/server, Web and SOA paradigms? How about building on that leverage with the best management, modeling, and integration for the best price? And you could bring the advertising, media, and consumer reach into the mix. Microsoft could become a one-stop shop, rather than a proponent of a re-hashed one-shop sop about Windows Everywhere.

The strategic risks are now such that an embrace of open source and Java-centric computing by Microsoft is at par with trying to foist Windows Everywhere on everyone (while eroding the boundaries slowly over time). Cut bait, I say, and fish anew -- and to hell with an instance on one rod over another.

We're really talking about cutting total costs and complexity, not purporting to cut them both while ignoring a major portion of the real world (and letting CIOs deal with it). BEA would allow Microsoft to redefine the distinct orbits of the .NET world and the distributed Java/open source tools world. The abstraction on where to seek lock-in nowadays is at the inflection point of RIAs, SOA and ALL the legacy.

A Microsoft-BEA combination would signal true neutrality as to underlying platform requirements -- while making the new differentiation point all about total management of IT costs and complexity. This is the value that IT buyers need in a partner, not just a vendor.