P4 based Load Balancing Strategies for Large Scale Software-Defined Networks

Title

P4 based Load Balancing Strategies for Large Scale Software-Defined Networks

Creator

Kulkarni M.; Goswami B.; Paulose J.

Description

To meet the large demands of future networks, several large-scale Software Defined Networking (SDN) test-beds have been designed. The increasing complexity of networks has resulted in convoluted methods for managing and orchestrating efficiently across a wide range of network environments. The load balance function is impaired when the controller fails to connect with the switches. A traditional Load Balancer (LB) must decapsulate layers one by one and get the information needed to run load balancing algorithms. For instance, OpenFlow, NetConf, Programming Protocol-independent Packet Processors (P4), and Data Plane Developement Kit (DPDK) provide network programmability at both the control and data plane levels. In this paper, authors implement load balancing using the P4 programming language without the need of a controller, the P4 load balancer can operate on its own. Controller's support is used to keep track on the health of the web servers. In this situation, the controller can identify a server failure and notify the P4 load balancer, which will restrict requests to the malfunctioning server, lowering the dispatching failure rate. A detailed investigation of various load balancing mechanisms is analysed in this paper followed by the identification of four potential approaches to large-scale SDN tests, including connection hash, weighted round-robin, DPDK technique, a Stateless Application-Aware Load-Balancer (SHELL). 2022 IEEE.

Source

2022 2nd International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies, ICAECT 2022

Date

2022-01-01

Publisher

Institute of Electrical and Electronics Engineers Inc.

Subject

DPDK; Load Balancing mechanisms; P4 Language; SHELL; Software-Defined Network

Coverage

Kulkarni M., Christ University, Department of Computer Science, Bangalore, India; Goswami B., School of Computer Science Queensland, University of Technology, Brisbane, Australia; Paulose J., Christ University, Department of Computer Science, Bangalore, India

Rights

All Open Access; Green Open Access

Relation

ISBN: 978-166541120-2

Format

Online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1109/ICAECT54875.2022.9807999

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85134262159&doi=10.1109%2fICAECT54875.2022.9807999&partnerID=40&md5=998f0a43128b1f58a2d60f5f06a2a1e8