The market for streaming software bloomed into existence when it spawned a whole series of start-ups, led by Apama (now part of Progress Software) and StreamBase. Larger companies including TIBCO, Avaya and BEA joined the fray either with their own offerings or through acquisition; and some companies, which had defined themselves in related markets, particularly the BAM (Business Activity Monitoring) market, repositioned themselves as streaming software in one way or another. It could be viewed as a case study on the genesis of a new software market.

This market is still very young with commentators estimating its size at no more than $30 million. Typically you get new entrants while a market is still young and the latest entrant to the streaming market is TruViso (until recently known by the rather confusing pre-launch name of Amalgamated Insight). TruViso is not just another streaming company with yet another set of slide-ware. Its approach is radical and, if I'm not mistaken, it will shake the market up.

Until now the market has been defined by the idea of quickly developing applications that scan a rolling window of transactions and mining that rolling window for useful and, possibly, highly profitable information. Think of it in terms of financial traders formulating new trading strategies to test on the financial markets in real time and then quickly developing applications to implement those strategies. Naturally you can extend such a paradigm to other streams of data such as telecomms activity or the data generated by sensors on an oil pipeline, but the most profitable stream of data known to man comes out of the financial markets, so that's where the new applications cluster.

TruViso's streaming technology does not work like that. Its engine works in a different way and it therefore enables streaming applications that work in a slightly different way. TruViso offers a continuous query engine. To understand how it works you need to understand the engine. The engine is not built to run specific queries against a moving window, it is built to execute large numbers of concurrent queries against such a window. There's some "technology magic" in this, because the engine delivers you better performance-per-query the more queries you include. (It sounds like a paradox, but it isn't. There are a number of in-memory data structuring techniques that can be used which will deliver such a performance profile). Conceptually, you can think of it as "multithreading its way through the queries" but it's got more to do with effective data structuring and management than multithreading.

Instead of adding a 4GL-like capability to build applications for the engine, TruViso provides an ability to define new and additional queries and a dashboard style BI-like interface which allows you to monitor the results of the queries that are running, in real-time. Naturally you can write code to trigger activity should specific conditions occur—as you can with other streaming software. In fact you can carry out BAM activity, CEP (Complex Event Processing) activity and ESP (Event Stream Processing) activity. In other words you can do the things you can do with other streaming products. However, on top of that you can execute multiple continuous queries and act on the collective results.

Think in terms of a patient in a hospital having multiple monitors attached to read various biological signals as they occur and having a single engine to interpret all of these signals at once. That's what TruViso adds to the streaming market. You can get the whole picture as well as any part of it.

It's radical and it's compelling and, in my view, it is destined to gain traction very quickly.