Solving The Real Time Problem For Twitter And Facebook


Real-time performance has been aptly defined as the ability to automatically deliver a digital message in a consistently predictable fraction of a second.

It is a definition Joe Ward, CEO of Groovy, eagerly embraces.

That’s because Ward believes he has solved a key bottleneck to the predictable delivery of online data, whether the rapid-fire messages are updates to Facebook news feeds or tweets on the Twitter network.

Groovy on Friday began shipping its SQL switch, the GSX 100, and with it its hope of bypassing the relational databases in the real time infrastructure.

The database is the slow link in the chain, says Ward, whose company recently moved from Australia to Silicon Valley. Instead of forwarding data immediately, or in real time, database feeds must be refreshed at intervals, meaning the flow of information is periodic instead of continual.

Database companies hope to compensate by using caches of memory chips to capture and feed information. But Groovy argues the technology it has spent three years developing is better suited for the task.

The GSX 100 has 24 computer chip cores that it uses in parallel to process information, and Ward claims it can save 20 percent of what a company spends to operate its databases and web and applications servers.

Interest in Groovy appears to be high. Companies such as Twitter and Facebook are finally realizing the constraints the database is having on real-time streaming, says Ward. “This is the one thing people haven’t been willing to say.”

With the launch of the GSX 100, this once hidden discussion may come out into the open. Most certainly it will in venture circles. Groovy, which has raised a couple million dollars in private funding up to now, plans to raise a $5 million to $10 million in financing from strategic investors.

If the GSX 100 proves a success, it may do that in real time.