New Method of Measuring TCP Performance of IP Network using Bio-computing

TL;DR

This paper introduces a novel bio-computing based method to measure TCP performance in real-time IP networks, achieving high accuracy compared to traditional approaches like Little's law.

Contribution

It proposes a new bio-computing approach, specifically molecular calculation, for real-time TCP performance measurement, addressing weaknesses of existing methods.

Findings

The bio-computing method closely matches Little's law results with only 0.13% difference.

Application on BIOKM model for FTP and IRCD protocols demonstrates effectiveness.

Provides a new avenue for accurate, real-time TCP performance analysis using phylogenetic analysis.

Abstract

The measurement of performance of Internet Protocol IP network can be done by Transmission Control Protocol TCP because it guarantees send data from one end of the connection actually gets to the other end and in the same order it was send, otherwise an error is reported. There are several methods to measure the performance of TCP among these methods genetic algorithms, neural network, data mining etc, all these methods have weakness and can't reach to correct measure of TCP performance. This paper proposed a new method of measuring TCP performance for real time IP network using Biocomputing, especially molecular calculation because it provides wisdom results and it can exploit all facilities of phylogentic analysis. Applying the new method at real time on Biological Kurdish Messenger BIOKM model designed to measure the TCP performance in two types of protocols File Transfer Protocol FTP and Internet Relay Chat Daemon IRCD. This application gives very close result of TCP performance comparing with TCP performance which obtains from Little's law using same model (BIOKM), i.e. the different percentage of utilization (Busy or traffic industry) and the idle time which are obtained from a new method base on Bio-computing comparing with Little's law was (nearly) 0.13%. KEYWORDS Bio-computing, TCP performance, Phylogenetic tree, Hybridized Model (Normalized), FTP, IRCD