Packet Loss Prevention Systems for Failover Incident on Network Infrastructure

Dharmawan, Richard and Samuel, Samuel (2019) Packet Loss Prevention Systems for Failover Incident on Network Infrastructure. In: 2019 5th International Conference on New Media Studies (CONMEDIA), 9-11 Oct. 2019, Bali.

[img]
Preview
Text
Packet Loss Prevention Systems for Failover Incident on Network Infrastructure.pdf

Download (1MB) | Preview

Abstract

There are many external factors that can cause a link failure between network appliances. Thermal, electrical/magnetic field or any harmful physical object is a common example. On traditional layer 2 network appliances, the failover mechanism decided internally by Spanning Tree Protocol (STP) calculation. Unfortunately, in every occurrence of failover, the communication will suffer from packet losses. When detecting a link failure, layer 2 appliances will calculate STP to decide which backup link to be selected and in that time period all of the packets suppose to be forwarded in the link will lose. On the other hand, OpenFlow capable network appliances (Software Defined Networks) can be orchestrated with a sensor to detect earlier any possible harmful external factors before a link failure happen. This initiation research only focuses to overcome an extreme thermal condition that can cause a link failure. OpenFlow Switch connected to a thermal sensor with a Raspberry Pi as a bridge. When the sensor detects temperature beyond its limit, it will tell the OpenFlow Switch to use a backup link immediately before the link failure occurs. With this mechanism packet loss can be prevented.

Item Type: Conference or Workshop Item (Paper)
Keywords: C# languages , Media , Packet loss
Subjects: 000 Computer Science, Information and General Works > 000 Computer Science, Knowledge and Systems > 003 Systems (Computer Modeling and Simulation)
Divisions: Faculty of Engineering & Informatics > Computer Engineering
Depositing User: Administrator UMN Library
Date Deposited: 03 Mar 2023 06:15
Last Modified: 03 Mar 2023 06:15
URI: https://kc.umn.ac.id/id/eprint/23324

Actions (login required)

View Item View Item