Not everybody remembers a little outfit called Wily Technology. It was a Silicon Valley startup that caught my eye because it did something that made eminent common sense: it watched a running Java application the way an end user would experience it on the Internet. In January 2006, CA acquired the eight-year-old company for $390 million.How can you know when a startup will pan out? You don't, but one of my indicators is to look at what the founders have done previously. When someone approached me to say I should take a look at the team at BlueStripe Software, I asked what made them different from everybody else. After several false starts, word gradually filtered through various spokesmen that this company included several early builders of Wily. Chris Neal, BlueStripe CEO, was VP of field operations for Wily and its early builder of revenue. John Bley, BlueStripe's chief architect, was an early engineering employee at Wily who eventually went into product management; ditto for John Whittington, BlueStripe VP of business development, who also filled that role at Wily.

The other principals had lived through decisive product periods at either Cisco, IBM's Tivoli system management, or Relicore, which eventually merged into Symantec. This isn't a sure sign of continued success. I've seen too many executives succeed once and announce they were about to do it again by divine right. It ain't necessarily so.

But expertise once applied to the demands of one era can sometimes be reapplied in another. Wily's principals understood the blind spots to running Java applications. You might be able to see the server and software running and you might even be able measure user response times. But if you didn't know the user was only getting an error message, it was all for naught. Wily's Introscope could detect that the application wasn't returning the result that the user needed and then alert someone who should care.

In talking to BlueStripe CEO Chris Neal, he says the problem with application operation now is that it may be split up into separate parts, with several parts running in virtual machines. I've heard this is a hard-to-manage setup and virtualization obfuscates the attempt to get a view of how the total application is running. Another way of putting it is applications these days are segmented to reflect varied business processes, but you don't know how well the business process is running because there are so many moving parts in isolated containers.

BlueStripe purports to address this issue, which places it in the midst of the drive toward virtualization combined with the trend to try to get to service-oriented architecture, or at least an application that consists of manageable modules of code. I don't know that BlueStripe will be able to bring any secret sauce to the process, but judging by its founders' focus of the past, they at least have excellent knowledge of what's gone wrong before and what might be done to correct it.