My Photo

« Writing Distributed Programs: Why Not Message Passing? | Main | IBM has a secret island headquarters hideaway inside a computer game »

October 10, 2006


du xu

By the similar mechanism, we can analyze the traffic characteristic and behavior of the user online, and then get semi-real-time popularity distribution. Based on the result, we can provide adaptive/self-learning in data placement services, in other words, we can push data and functions properly over the network. Is that possible?

The comments to this entry are closed.