StreamBase

The Event Stream Processing (ESP) market is one that fascinates me. StreamBase barged into this market over a year ago talking in terms of tens or hundreds of thousands of transactions per second processing fast and vast streams of data. It occurred to me at the time that this could be a game changing development.

The point is that every 6 years or so Moore's law delivers a ten-fold improvement in performance, not just of CPU but memory, buses, disk and the rest. Because of that, every 6 years or so different applications become possible because the hardware has increased in power by one order of magnitude. The emergence of the ESP market is, if you like, the direct result of such progress.

Anyway, in its latest release StreamBase has added what you can think of as a data warehouse access capability. It is a fast and efficient indexing structure that can be applied to large heaps of data and which feeds StreamBase. Such indexing techniques are nothing new, of course, they've been around for about a decade at least, but nobody has previously fashioned one for use in ESP applications.

What StreamBase has done is widen the potential area of application for ESP. This is no bad thing and it is also inevitable. If StreamBase had not done it, someone else would have.

What I'm beginning to sense though, is that we could see a whole new software architecture emerge out of this. What I'm thinking is that we are entering the era of real-time BI. Real-time BI is way different to the trickle-feeding of data warehouses and the construction of data marts. Something new is going on.

EMC Acquires RSA

The big news this week was EMC's acquisition of RSA, for which they paid about $2 billion. Does it make sense?

Well, some of EMC's acquisition have puzzled me a little, but that may be because I'm not privy to the grand plan. The acquisition of Legato made sense. The acquisition of Smarts and ControlCenter also made sense in they were all about Storage Resource Management. The acquisitions of Rainfinity and Invista gave EMC a powerful storage virtualization capability. (SANs were only the first step in creating an independent data layer within the corporation). All of those acquisitions were logical.

However, Documentum seemed a strange acquisition to me. It was relevant to EMC's business because, well, data is data. But document management was previously peripheral to EMC's business lines. More surprising, however, was its acquisition of VMWare. I was told at the time that VMWare knew it would eventually be acquired by someone and that it had fended off interest from both Microsoft and IBM, so maybe it saw EMC as a safe pair of hands.

No matter. With annual revenues of about $10 billion, EMC is one of the powerhouses of the IT industry. Its software revenues are about $2 billion (20% of the total) and it clearly sees software as a growth area. Its acquisition of RSA just pushed the software figure up by about $1.2 billion.

But how does RSA's unparalleled encryption technology fit in with EMC's business? Well in some ways it doesn't. RSA pulls in very healthy revenues with strong authentication and this has very little to do with data. But that part of RSA is also a nice annuity business. It just keeps growing year-on-year. EMC in its “acquisition story” made a big play about the need for better data security—and I don't see anyone doubting the need. But still I have yet to see anyone spending big on this.

The problem with encryption is that the encrypt/decrypt cycle chews up a fair amount of CPU power so, although it would be possible to encrypt onto disk and decrypt on reads, it is rarely done. Having acquired RSA, EMC is in the situation where it could feasibly offer a kind of global end-to-end encryption capability so maybe they'll be able to make a market with this. But I'm not convinced.

SCO Attack Beaten Back

In case you've not been following it, in the excitement of the other World Cup action, SCO is in trouble in its knock-out match against the much fancied IBM. (The match is playing out now as I write these words). SCO, it has to be said, has not been awed by the opposition and has been nimble in its tactics, hoping to keep the score even in the hope of winning big in a $billion dollar penalty shoot out.

However, its nifty midfield lawyers could do little more than look on in dismay when Utah referee, Judge Wells, penalized SCO for “prejudicial vagueness” (which we believe to be the US soccer term for diving to the ground and holding your knee in order to give the impression that your opponent tripped you). SCO was clearly hoping for a severe penalty but its claims of foul (198 of them as it happens) were dismissed by the ref. There were load roars of approval from the Linux crowd in the East Wing of the stadium, while the IBM coach, gesticulating on the touchline, made no comment whatsoever, in typical IBM fashion.

What does it all mean?

A legal friend of mine informs me that the soccer analogy doesn't work too well. The truth is that SCO can still win, he tells me, but they are now unlikely to win big. There are still 10 SCO claims that IBM has to answer and a financial settlement is now a likely outcome, with IBM arguing that SCO was a failing business anyway and therefore the settlement should be low. But what do lawyers know about soccer?