Algorithmic Challenges in the Design of Efficient Data Centers

March 10, 2017 - 11:15am to 12:15pm
Duke University
Keller Hall 3-180
Ravi Janardan

ABSTRACT: One of the main reasons for the enormous success of cloud and big data applications is the ability to store and process vast amounts of data rapidly. Cloud computing heralded a paradigm shift in computing which resulted in an unprecedented growth of data centers. Hence, there is an urgent need to build cloud infrastructure that is fast, simple, and cost-effective.  

In this talk, I will focus on algorithmic challenges in building efficient data centers. ​In particular,  I will talk about two main components of a data center: 1) Big data clusters 2) Data center interconnect. I will highlight the interdisciplinary nature of the problems that arise in these systems, and the interplay ideas from online algorithms, optimization theory, and economics in the design of algorithms.

BIO: Janardhan Kulkarni is a postdoctoral researcher in the theory group at Microsoft Research, Redmond. He obtained his Ph.D. from the department of computer science at Duke University, Durham in 2015.  Kulkarni's primary research interest is in the design of algorithms with provable performance guarantees for problems that arise in building large systems.  His research has appeared in top theory and system conferences such as  STOC, FOCS, SPAA, EC, IPCO, SIGCOMM, and OSDI. He is a recipient of Duke CS Best Thesis Award 2015, Duke CS Teaching and Mentoring Award 2013, and a gold medal from the Indian Institute of Science in 2010.