The methods developed and made practical by Kannan Ramchandran for distributed source coding and distributed storage coding are benefiting image and video communications and large data storage systems. Ramchandran connected distributed source coding theory to channel coding approaches that could be applied to real applications, such as video. To overcome the unreliability of nodes in large distributed systems where data is stored over multiple nodes for redundancy, Ramchandran created regenerating codes. With these codes, he demonstrated how considerably less data is needed to be transferred over the network when a failed node is repaired, while maintaining minimal storage overhead. Variants of these codes have saved companies like Microsoft hundreds of millions of dollars in data center costs and will be part of future releases of the Apache Hadoop open-source framework.
An IEEE Fellow, Ramchandran is a professor of electrical engineering and computer science with the University of California, Berkeley, Berkeley, CA, USA.