Last year I visited GE Software -- the Bay Area's largest startup -- General Electric's new software research center in San Ramon. It was hotter than 110 degrees and I vowed never to complain about San Francisco's cold summers.

The new business group has about 700 engineers and is primarily focused on the industrial Internet, or the "Internet of things." It has an ambitious roadmap. One of its goals is to use Big Data to help its massive jet engines (above) run more efficiently and reliably.

This week it revealed Perdix, a single common platform designed to be used for many different types of industrial systems monitoring and analytics, rather than the thousands of separate, incompatible systems that are in use today. 

When I met with Bill Ruh, head of GE Software, (above) he said that the development of a common architecture for the diverse number of industrial systems that GE produces and operates was a vital first step. It will enable quicker development of new apps and make it easier to share important capabilities between applications and thus better scale the work of 700 software engineers -- a massive payroll that constantly needs to prove itself. 

Kyle Vanhemert, reporting for Wired: GE's Radical Software Helps Jet Engines Fix Themselves

Predix, a flexible software platform intended to dramatically streamline the monitoring and maintenance for all the industrial technologies GE provides. Whereas field engineers currently wrestle with idiosyncratic systems and separate interfaces for all the different hardware they service-sometimes armed with little more than a briefcase full of paper manuals-Predix and its card-based UI will gradually become the interface for all those machines.

But the software's promise goes beyond cohesiveness and convenience. The real vision is to link all these diverse machines to the cloud, quantifying their performance and benchmarking them against each other-all in the name of eliminating the number one bugaboo of all these industries: unscheduled down time. Imagine an engineer walking up to a wind turbine with tablet in hand, and the turbine telling him that a fuse should be replaced not because it's broken but because it will give out in three weeks time. That's the future GE's really looking toward.

A 1 percent efficiency in jet engine performance can provide a $15 billion savings in jet fuel. For an arline industry that operates on narrow margins, that humble 1 percent is very significant.

GE this week added 14 new analytics packages to the the ten launched last year, which have generated $290m so far, with orders for $400m. It expects to grow its software revenues by more than 15% annually.

Jeff Immelt, chairman and CEO of GE, said, "Industrial data is not only big, it's the most critical and complex type of big data. Our greatest challenge and opportunity is to manage and analyze this data in a highly secure way."

I met Mr. Immelt briefly when he visited late last year for an event (he brought along a GE jet engine, top). He said that GE Software is looking for help from Silicon Valley startups.

That might be a tall order because few have the domain experience to be able to figure out how to analyze the trillions of data points that a jet engine produces, or any other industrial kit. But the Silicon Valley area can certainly provide GE Software with the engineers it needs to implement the heavy duty industrial apps that GE and its customers require.