Abstract

Real-time communication is critical to emerging cloud applications from smart cities to industrial automation. The new class of latency-critical applications requires latency differentiation and performance isolation in a highly scalable fashion in a virtualized cloud environments. This dissertation aims to develop novel cloud architecture and services to support real-time communication at both the platform and infrastructure layers. At the platform layer, we build SRTM, a scalable and real-time messaging middleware (platform) that features (1) latency differentiation, (2) service isolation through rate limiting, and (3) scalability through load distribution among messaging brokers. A key contribution of SRTM lies in the exploitation of the complex interactions among rate limiting and load distribution. At the infrastructure layer, we develop VATC, a virtualization-aware traffic control framework in virtualized hosts. VATC provides a novel network I/O architecture that achieves differentiated packet processing with rate limiting while being scalable on multi-core CPUs. The research is evaluated in a cloud testbed in the context of Internet of Things applications.

Committee Chair

Chenyang Lu

Committee Members

Chenyang Lu, Roch Guerin, Christopher Gill, Sanjoy Baruah,

Comments

Permanent URL: https://doi.org/10.7936/ws5z-cj66

Degree

Doctor of Philosophy (PhD)

Author's Department

Computer Science & Engineering

Author's School

McKelvey School of Engineering

Document Type

Dissertation

Date of Award

Spring 5-15-2020

Language

English (en)

Author's ORCID

http://orcid.org/0000-0001-8957-7110

Download

Share

COinS