Date of Award

4-2023

Document Type

Thesis

Publisher

Santa Clara : Santa Clara University, 2022.

Degree Name

Master of Science (MS)

Department

Computer Science and Engineering

First Advisor

Behnam Dezfouli

Abstract

WiFi wireless access points must adapt to complex network operations as connected devices and their bandwidth requirements continue to rise. The use of software switches supporting Programming Protocol-independent Packet Processors (P4) in wireless access points has drawn significant attention due to their versatility, flexibility, and potential to overcome the limitations of conventional network designs. T4P4S and BmV2 are the two most widely used P4 programmable software switches, and comparing the performance of these switches allows network operators to select the most suitable switch for specific use cases, such as minimizing latency or maximizing throughput while reducing resource consumption of the underlying hardware.

This thesis examines the performance and resource consumption of T4P4S and BmV2. We first examine their design spaces and then evaluate their performance in the presence of various traffic types. Based on our experimental findings, the efficient packet processing pipeline of the T4P4S switch makes it a superior alternative to the BmV2 switch for executing P4 programs. The presented results and analysis provide a better understanding of design decisions and identify potential performance bottlenecks for P4 programmable software switches.


Download

Share

COinS