Codes for (Un)Expected Loads

# 205






Abstract

Distributed computing systems strive to maximize the concurrent data access requests they can support with fixed resources. Replicating data objects according to their relative popularity and access volume helps achieve this goal. However, these quantities are often unpredictable. In emerging applications such as edge computing, the instantaneous and expected numbers of users and their data interests fluctuate extensively. Therefore, data storage schemes should support such dynamics. Erasure coding is emerging as an efficient and robust form of redundant storage. This talk asks which data access rates erasure-coded systems can handle. It introduces the notion of access service rate region. It argues that it should be an essential consideration in designing efficient distributed systems that must remain stable for a wide range and multiple combinations of mean loads. The problems associated with the service rate region metric generalize several notions in coding theory for distributed systems. We address them using math tools such as combinatorial optimization on graphs and finite geometry.

Prof. Emina Soljanin, Distinguished Professor, Rutgers University.

Emina Soljanin is a Distinguished Professor of Electrical and Computer Engineering at Rutgers. Before moving to Rutgers in January 2016, she was a (Distinguished) Member of Technical Staff for 21 years in various incarnations of the Mathematical Sciences Research Center of Bell Labs. Her interests and expertise are broad and currently range from distributed computing to quantum information science. She is an IEEE Fellow, an outstanding alumnus of the Texas A&M School of Engineering, the 2011 Padovani Lecturer, a 2016/17 Distinguished Lecturer, and the 2019 IEEE Information Theory Society President. In 2023, Emina received the IEEE Information Theory Society Aaron D. Wyner Distinguished Service Award and Mrs Urmila Agrawal Distinguished Visiting Chair Professorship at the Indian Institute of Sciences.