Session

Precision Time Protocol optimization using genetic algorithm

Speakers

Marta A Plantykow
Milena Olech
Maciej Machnikowski

Label

Moonshot

Session Type

Talk

Contents

Description

This paper represents the work in developing an automated methodology of tuning the PI controller used by linuxptp applications to even out timing errors in its implementation of the Precision Time Protocol (PTP).
In most cases, the hardware-accelerated PTP protocol synchronizes the PTP Hardware Clock (PHC) running on the Network controller. Consequently, system time synchronization requires an additional step that synchronizes the system time to the PHC.
PI controller optimization is needed to fulfill the strict phase-synchronization requirement of telecommunication systems introduced by the LTE E-ULTRA TDD and 5G technologies.
This paper explores the use of a genetic algorithm (GA), a stochastic algorithm used in Artificial Intelligence, to optimize the parameters of the PI controller. GA mimics the process of natural evolution and has shown to be capable of locating high-performance areas in complex domains.
This paper presents different approaches to solve the described problem, various data evaluation methods, and the results achieved in the process.
This paper is also a continuation of the Performance optimization using artificial intelligence methods talk presented on the NetDev 0x14.