Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier

https://orcid.org/0000-0002-0967-1033

AccessType

Open Access Dissertation

Document Type

dissertation

Degree Name

Doctor of Philosophy (PhD)

Degree Program

Electrical and Computer Engineering

Year Degree Awarded

2020

Month Degree Awarded

September

First Advisor

Russell Tessier

Subject Categories

Systems and Communications

Abstract

With the application of virtualization technology in computer networks, many new research areas and techniques have been explored, such as network function virtualization (NFV). A significant benefit of virtualization is that it reduces the cost of a network system and increases its flexibility. Due to the increasing complexity of the network environment and constantly improving network scale and bandwidth, it is imperative to aim for higher performance, extensibility, and flexibility in the future network systems. In this dissertation, hybrid NFV platforms applying virtualization technology are proposed. We further explore the techniques used to improve the performance, scalability and resilience of these systems. In the first part of this dissertation, we describe a new heterogeneous hardware-software NFV platform that provides scalability and programmability while supporting significant hardware-level parallelism and reconfiguration. Our computing platform takes advantage of both field-programmable gate arrays (FPGAs) and microprocessors to implement numerous virtual network functions (VNFs) that can be dynamically customized to specific network flow needs. Traffic management and hardware reconfiguration functions are performed by a global coordinator which allows for the rapid sharing of network function states and continuous evaluation of network function needs. With the help of state sharing mechanism offered by the coordinator, customer-defined VNF instances can be easily migrated between heterogeneous middleboxes as the network environment changes. A resource allocation algorithm dynamically assesses resource deployments as network flows and conditions are updated. In the second part of this thesis document, we explore a new session-level approach for NFV that implements distributed agents in heterogeneous middleboxes to steer packets belonging to different sessions through session-specific service chains. Our session-level approach supports inter-domain service chaining with both FPGA- and processor-based middleboxes, dynamic reconfiguration of service chains for ongoing sessions, and the application of session-level approaches for UDP-based protocols. To demonstrate our approach, we establish inter-domain service chains for QUIC sessions, and reconfigure the service chains across a range of FPGA- and processor-based middleboxes. We show that our session-level approach can successfully reconfigure service chains for individual QUIC sessions. Compared with software implementations, the distributed agents implemented on FPGAs show better performance in various test scenarios.

DOI

https://doi.org/10.7275/t03j-3g74

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS