Title: Systems Techniques for Enhancing Content Delivery in the face of Network Variability in the Edge-Cloud Continuum
Date: Monday, May 8, 2023
Time: 14:00 – 15:30 EST
Location: Virtual (Click here to join)
Manasvini Sethuraman
Ph.D. Student
School of Computer Science
College of Computing
Georgia Institute of Technology
Committee:
Dr. Umakishore Ramachandran (Advisor) – School of Computer Science, Georgia Institute of Technology
Dr. Ashutosh Dhekne – School of Computer Science, Georgia Institute of Technology
Dr. Raghupathy Sivakumar – School of Electrical and Computer Engineering, Georgia Institute of Technology
Dr. Anand Sivasubramaniam – Dept. of Computer Science & Engineering, Pennsylvania State University
Dr. Ellen Zegura – School of Computer Science, Georgia Institute of Technology
Abstract:
Internet traffic today is largely dominated by video, both on-demand video from sources like Netflix and YouTube, as well as live-streaming and video conferencing. Increasingly, users who consume such content are mobile. The last mile of network wireless connectivity for delivering such content to mobile users varies widely, be it cellular or Wi-Fi. Wireless links exhibit temporal and spatial variations in bandwidth, which can cause disruptions and degradation in the user's quality of experience. At the same time, there are emerging network technologies that could be exploited to increase the quality of content delivery. This thesis proposal investigates a variety of techniques for both mitigating disruptions due to network variability and enhancing content delivery: (a) by adapting to spatio-temporal variations in bandwidth availability, (b) by exploiting short-range high bandwidth mmWave links, and (c) by orchestrating resource allocation in the face of bandwidth variations.
We first examine how knowledge of the mobile user's trajectory can be utilized to inform the user's device of the available bandwidth in the near future, and thus plan content download ahead of time. We then examine how content delivery for mobile users can be enhanced by leveraging 60 GHz short-range mmWave links for "just in time" delivery of stored content using the evolving geo-distributed edge infrastructure and the user's travel route. Finally, we look at a more general problem of resource orchestration on a wirelessly connected edge infrastructure hosting bandwidth-intensive and latency-sensitive services for a community of users (both mobile and stationary). Placement of services not only needs to consider CPU and memory as the available resources, but also bandwidth and latency, which may change over time, thereby necessitating service migration. We propose efficient algorithms for initial service placement, as well as for service migration, when application service-level-objectives (SLOs) are violated because of network variability.