A completely new kind of approach to innovation is required to ensure EMC doesn't lose its way as storage needs grow and change, in light of companies' adopting Web 2.0 applications, service-oriented architectures and software-as-a-service offerings.

That was the message at this week's Red Herring East conference in Boston from EMC executive vice-president and chief development officer, Mark Lewis, during an interview about the vendor's product direction.

"It isn't all about us," Lewis said, who joined EMC from HP five years ago and maintains a blog to share his thoughts on a regular basis. "But it is about building communities and ecosystems around different business needs . . . It's not just about a bunch of engineers inside the firewall or here in Hopkinton [ EMC world headquarters]."

Fostering innovation wasn't about overly structured processes either, he said.

"We have a great Six Sigma program and while I love it as applied properly, I told those guys if they ever come near our creative process I'm going to shoot 'em," Lewis said. "Those kinds of processes for creativity are some of the worst [things] you can do."

EMC had been able to innovate in recent years through more than two dozen acquisitions and upwards of $US10 billion in R&D spending, Lewis said.

Those efforts let EMC diversify well beyond being a one-trick pony selling Symmetrix boxes to large organisations and become a horse of many colours whose revenue is split pretty evenly between hardware and software and services. Addressing new storage, security and data management needs brought about by customer adoption of new collaborative and rich media applications would require new offerings from EMC, Lewis said.

In the light of such developments, big changes have been taking place in such areas as EMC's developer environment, which "used to be characterized by a small cadre of folks developing around our product sets such as Documentum or Smarts", Lewis said.

EMC had evolved that into a hugely collaborative environment that now had more than 25,000 people writing Documentum code, he said.

The company last month introduced a formal effort to coordinate its internal R&D personnel better with those at university labs and elsewhere.

A big goal for EMC is to shortcut the traditional product development model (do some research, product development, roll out support and services, deliver to customers, and go from there) by involving potential partners and customers much earlier in the process.

Under the new model, proofs of concept and research prototypes (technologies in pre-alpha and beta stages) would be put into customers' hands, Lewis said. Products would evolve from there.

To encourage participation from a broader development community, EMC plans to run more contests that will reward code writers with prizes for their efforts (Lewis cited a contest to build metatagging for photos).

He said EMC also was innovating through development of a common platform strategy to better integrate the many technologies it has developed and obtained through acquisition. This should greatly cut the cost of building security, management and other components used by many of its products, Lewis said.

"Inside EMC we are community-sourcing more and more code for our internal community, just like you would open source on the outside world," he said. "It saves us in R&D [and provides] more consistency for our customer base."

EMC expected the concept of software appliances, or virtual machines that run on various platforms, would gain momentum through this effort, Lewis said.