Yahoo! Ratchets! Open! Source! Grid! Platform!
Hoping to further the Google-battling ways of an open source cluster computing platform named for a stuffed elephant, Yahoo! has decided to share a 4,000-CPU, 1.5 petabyte grid with some boffins at Carnegie Mellon University.
Context & Ripple Effects
With no internal search-ad engine to rival Google's, Yahoo! is playing the one card it holds at scale: infrastructure. The Register reports the company is opening a 4,000-CPU, 1.5-petabyte Hadoop cluster — the open source platform named for a stuffed elephant — to Carnegie Mellon University researchers, turning an internal batch-processing tool into a shared academic resource.
First-order effects
- Carnegie Mellon researchers gain access to a 4,000-CPU, 1.5-petabyte grid, giving Hadoop a high-profile university proving ground outside Yahoo!'s own data centers.
Second-order effects
- Rival consumer-web platforms face pressure to answer with comparable open-source infrastructure gestures, since Google keeps its equivalent grid technology proprietary.
Third-order effects
- If large web operators keep donating commodity clusters to universities, open-source distributed computing becomes the default substrate for academic computer-science research — and a recruiting pipeline back into the donors' engineering ranks.
The trend: Web-scale operators are converting proprietary internal grids into open-source platforms as a competitive lever against Google's closed stack.