What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
ISQED 2013: IBM lays it on the line
March 7th, 2013 by Peggy Aycinena
IBM’s Brad Brech gave one of the Tuesday morning keynotes at ISQED 2013 on March 5th. It was a thought-provoking talk and well received by a large audience of engineers at the Techmart in Santa Clara. Per Brech, like everywhere in computing the trend in data centers is “smaller, cheaper, and faster.”
To illustrate, he drew comparisons with developments in aeronautics over the last century, invoking the Wright Brothers’ aircraft, the Curtiss Flying Boat, the Boeing Stratoliner, the De Havilland Comet, the Boeing 707, and the Concorde. All of these aircraft, he said, addressed technology concerns related to speed. As tragically illustrated by the Concorde disaster in 2000, however, eventually speed as a lone metric of progress in commercial aviation proved unsustainable. Instead, after 2000 the industry became more concerned about efficiency and less about speed.
Similarly, Brech argued, now for the first time in the history of computing, “While CEOs are identifying technology as the single most important external force impacting their organizations, they’re not interested in the speed of the technology but how quickly and efficiently it can be brought online. Now the IT cycle is about speed to adoption and efficiency, not just about the latest/greatest software or hardware.”
“What is driving IT costs and value today? These escalating costs have shifted to the labor element. In the past, you could launch an app and if it grew stale, you rolled out something new. Nowadays however, we want to be able to reuse our platforms, both hardware and software; we’re all looking at software-defined networks, at software-defined environments. The system need to knows how to take care of itself, so the relaunch and upgrade can be done much more efficiently.”
While all of that is complex, Brech said, the problems associated with security are even worse: “Not only do we need speed in adopting new technologies, but mobile devices are now vastly extending the reach of our [data networks], particularly in specific industries such as financial and medical. People don’t like their medical secrets, or credit cards, being out there on the web.
“Yet, even as people want massive functionality in their devices, their systems, and their data centers, they don’t want to pay for the security features [they demand], especially if those features are seen as an extra expense. Therefore, we have no choice but to build the security right into the hardware.”
Brech moved on to a third area of concern: “Then there’s the problem of storage. Industries such as medicine are gathering more and more data today, with more and more resolution, and they’re finding that they need to keep every single data point.
“They can’t [afford] to compress an image or lose a single pixel that might stand between the life and death of a patient. And now that our society is seeing how much more can be done with all of this, we have no choice but to think about faster and more massive storage capabilities.”
He detailed some of the strategies for meeting those demands, and layered them on top of a complex road map that would also address the issued outlined earlier in his talk: “We will need advances in alternative design strategies, special purpose accelerators, fabric-interconnect technology, memory and storage technology, and all of this with an eye on security.”
Brech challenged his audience of hardware designers to get out there and make it happen through a combination of “consolidation and integration, optimization and standardization, simplification and automation, and dynamic optimization.”
In other words, Brech asked for no less than “sustainable innovation for smarter computing in the data center,” which coincidentally and happily was the title of his talk. It goes without saying, however, that sustainable innovation is the cornerstone in all computing, and aviation for that matter.
Brad Brech is a Distinguished Engineer in the Systems and Technology Division of IBM in the office of the CTO. He is currently leader of the Systems and Technology Architecture Board, where his responsibilities include technical strategy, energy efficient computing, intellectual property development and technical vitality for IBM STG. He joined IBM in 1982.
You can read more about the Curtiss Flying Boat here. The photo of Glenn Curtiss and Henry Ford is from that blog post in “Flying“.