One often gets caught up in the realities of customer pain-points and current market maturity in devising solutions for the next 9-18 month time-frame. I find it refreshing to have an ongoing parallel thread to collaborate with academia and understand the longer term horizon (3-5 years). Interestingly, the longer term perspective helps shape Blue Market strategies and new market verticals, which will eventually put a dent into the install-base of competitors as well as those with a first-mover advantage.
OK, back to the topic of this blog. I attended the Berkeley EECS Annual Research Symposium (BEARS'10) which has an awesome collection of talks and great networking opportunities. Michael Franklin (a pioneer in database research) shared some of his interesting thoughts on the future of clouds. I encourage you to check-out the video on the BEARS web-site.
There were two interesting take-away points from my perspective:
1) The role people will play in cloud (an excellent example of Crowd Computing is Amazon's Mechanical Turk program). I would consider this as a Blue Market with tons of new applications.
2) The idea of using Machine Learning to model the Data Center Computer (aka Cloud). I have worked in machine learning, and realize its a complex problem. But the pay-off is huge. Imagine the potential to reduce your operating costs with machine learning techniques pin-pointing problems/root-causes such that our current administrator can manage 10 or 100 times more storage and compute foot-print.
An interesting food for thought!
OK, back to the topic of this blog. I attended the Berkeley EECS Annual Research Symposium (BEARS'10) which has an awesome collection of talks and great networking opportunities. Michael Franklin (a pioneer in database research) shared some of his interesting thoughts on the future of clouds. I encourage you to check-out the video on the BEARS web-site.
There were two interesting take-away points from my perspective:
1) The role people will play in cloud (an excellent example of Crowd Computing is Amazon's Mechanical Turk program). I would consider this as a Blue Market with tons of new applications.
2) The idea of using Machine Learning to model the Data Center Computer (aka Cloud). I have worked in machine learning, and realize its a complex problem. But the pay-off is huge. Imagine the potential to reduce your operating costs with machine learning techniques pin-pointing problems/root-causes such that our current administrator can manage 10 or 100 times more storage and compute foot-print.
An interesting food for thought!
No comments:
Post a Comment