Martin Banks

Google recently raised a few eyebrows by snapping up a small software company called PeakStream, which specialises in tools that help single-threaded applications run on multi-core processors. The purchase should help Google develop better multi-threaded code for running on x86 processors, which makes perfect sense from both a technological and commercial perspective.

If Google is to maintain its leadership in the potentially massive search-and-every-service-that-spins-off-it market, then it needs to keep its edge in software development. And moving to multi-threaded applications should also provide a welcome performance boost for future services.

PeakStream’s assimilation by the search giant is arguably a loss for the broader user community, though, because the majority of applications out in the real world are still single threaded and developers are not well schooled in any other technology model.

That we have four-core processors already upon us, with eight-core variants trundling over the horizon cannot be ignored. Nor will it be long before 16-core processors will be around, so any developer that still thinks a single-threaded approach is going to cut the mustard is likely to be sadly disappointed.

If the PeakStream technology could provide a technological bridge between what developers can already do and what the new multi-threaded technology will allow them to do, then it is a sad thing that Google has gobbled up its expertise before it could be shared with everybody else.

Google might just want PeakStream’s technology to help it target its adverts more accurately. But I am sure that it has identified all sorts of other ways that PeakStream could come in handy.

If you look at what the company is doing with the basic concept of search, all you see are services. It has already sussed out that

software-as-a-service (SaaS) is, above all, about the service, and so is now busy adding new services to its portfolio.

Google now has the technology to offer SaaS on a scale that few other vendors, if any, can match. PeakStream gives it access to parallel programming skills that will deliver mind-boggling power when applied to 50,000-plus, eight-core, dual-processor servers – a SaaS platform that may make all the competition give up in despair.

It could also give Google the ability to deliver compelling services directly to users – be they individuals or major corporations. Such is Google’s awesome marketing and financial clout, it could even afford to give away end-user client systems pre-configured to work with its services.