Title: Algorithms and System-Theoretic Foundations for Edge Computing Systems
Speaker: Dr. Jaya Prakash Varma Champati, Information Science and Engineering, KTH Royal Institute of Technology, Sweden
Date, Time: 6th June 2019, 11.30 AM
Venue: IEOR Seminar Room (2nd Floor, IEOR Building)
Abstract: Edge computing or fog computing, where computational resources (e.g. edge servers, cloudlets, micro-datacenters) are placed close to devices that offload computational tasks or data for processing, is seen as a key architectural component of 5G and future wireless networks. Placing computational resources close to the devices will reduce latency, thereby enabling future networks to respond in real time, which further paves way for a new wave of time-critical-control applications from the domain of Cyber-Physical Systems (CPS) including process automation systems, autonomous driving, smart grid, smart cities etc.
To realize the full potential of edge computing several research questions need to be addressed. In this talk, we focus on two key questions, 1) how to efficiently offload (computational offloading)? and 2) how to analyze real-time response for time-critical control applications? We study different aspects of computational offloading using new paradigms for scheduling on parallel machines and propose algorithms. We propose algorithms with provable competitive ratios, and use simulation for average performance evaluation. To analyze the real-time response, we argue that the traditional stationary metrics such as average delay per packet are not suitable and propose transient analysis framework for multi-hop wireless networks. We further focus on the freshness of data in time-critical control applications. In particular, we study Age of Information (AoI), a metric for quantifying the freshness of data at the receiver. Given an age limit, we analyze and optimize AoI age limit violation probability for different queuing systems.
I will conclude my talk by highlighting two of my future research directions, 1) Use of AoI in networked control systems, and 2) Scalable machine learning at the edge.
Speaker Bio: Jaya Prakash Champati is a post-doctoral researcher from the division of Information Science and Engineering, EECS, KTH Royal Institute of Technology, Sweden. He finished his PhD in Electrical and Computer
Engineering, University of Toronto, Canada in 2017. He obtained his master of technology degree from Indian Institute of Technology (IIT) Bombay, India, and bachelor of technology degree from National Institute of
Technology Warangal, India. His general research interest is in the design and analysis of algorithms for scheduling problems that arise in networking and information systems. Currently, his focus is on theoretical
foundations for Edge computing, Networked control for Cyber-Physical Systems (CPS) and the Internet of Things (IoT). Prior to joining PhD he worked at Broadcom Communications, where he was part of developing LTE MAC
layer. He was a recipient of the best paper award at IEEE National Conference on Communications, 2011.