Flexible IoT Datapath Programming using P4

TL;DR

This paper explores using the P4 language to enable flexible and dynamic datapath programming for large-scale IoT data streaming, demonstrated through a case study on MST-based data collection in arbitrary network topologies.

Contribution

It introduces a novel approach to IoT data handling by leveraging P4 for dynamic datapath programming, enhancing flexibility and scalability.

Findings

Demonstrated dynamic MST-based datapath programming for IoT data collection

Showed feasibility of P4 in large-scale IoT streaming scenarios

Proposed a flexible, protocol-independent datapath design

Abstract

The progress of the network and device technologies enables any device to be connected to the Internet thereby forming an Internet of Things (IoT) architecture. Internet-related activities based on the IoT services can inevitably generate a lot of real-time data which leads a big data phenomenon. To efficiently handle such big IoT data, software-defined networking (SDN) techniques could come into play. Among SDN techniques, P4 is a highlevel language for programming protocol-independent packet processors. This paper presents our preliminary work on how we can use P4 to achieve flexible datapath programming for large scale IoT streaming data streaming. As a case study, we describe the implementation of dynamic MST-based datapath programing for IoT data collection in an arbitrary network topology.