Horizontal scaling of video conferencing applications in virtualized environments
Date
2016
Authors
Luo, Mante, author
Pallickara, Shrideep, advisor
Papadopoulos, Christos, committee member
Turk, Daniel, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Video conferencing is one of the most widely used services in the world. However, it usually requires dedicated hardware and expensive licenses. Cloud computing has helped many companies achieve lower operation costs, and many applications including video conferencing are being transitioned into the cloud. However, most video-conferencing applications do not support horizontal scaling as a built-in feature, which is essential to embrace the advantages of virtualized environments. The objective of this thesis is to explore horizontal scaling of video conferencing applications. We explore these ideas in the context of a Jitsi an open-source video-conferencing. The thesis develops a methodology for horizontal scaling in the Amazon EC2 cloud with the objective of ensuring quality of service such as per-packet latency (primarily), loss rates, jitter, and the number of participants per session. We build predictive models to inform our horizontal scaling decisions. Proactive scaling allows us to preserve several qualities of service metrics for video-conferencing. Scaling in the EC2 environment is fast and cost-effective with the added benefit of high availability, which helps us support large number of users consistently without much downtime.
Description
Rights Access
Subject
distributed systems
cloud computing
video-conferencing