In the large-scale distributed storage systems - cloud storage and big databases - servers fail rather frequently for variety of reasons ranging from link overload to software crash. These networks of storage must allow quick repair and support fast updates and queries at the same time. In the classical optimization problems of information/coding theory, these "local" processing parameters are missing. In this talk we describe new (graphical) models for large scale repairable networked storage systems that account for the topology of storage networks and construct reliable codes that are both update-efficient and locally repairable.
Arya Mazumdar is an assistant professor in the University of Minnesota - Twin Cities (UMN) since 2013. Before coming to UMN, he was a postdoctoral scholar (2011-12) at the Massachusetts Institute of Technology. Arya received his Ph.D. from the University of Maryland, College Park, in 2011. He spent the summers of 2008 and 2010 at the Hewlett-Packard Laboratories, Palo Alto, CA, and IBM Almaden Research Center, San Jose, CA, respectively.
Arya is a recipient of NSF CAREER award (2015), the IEEE ISIT Best Student Paper Award (2010), and the 2011 Distinguished Dissertation Award (at University of Maryland). Arya’s research interests include information and coding theory and their applications to storage, security and biology.