THE EFFECT OF TRAFFIC SHAPING AND DYNAMIC BANDWIDTH ALLOCATION AS CONGESTION CONTROL MEASURES IN INTERNET PROTOCOL NETWORKS

Abstract

ABSTRACT Internet protocol networks are experiencing increased traffic load, resulting in congestion. This congestion can have detrimental effects on the network performance and hence the need for congestion control. This research assessed the effectiveness of traffic shaping and dynamic bandwidth allocation as congestion control measures in Internet Protocol (IP)networks. A typical IP network topology was designed with end devices configured. End devices used were five routers, three switches, Hyper-Text Transfer Protocol (HTTP) server, Domain Name Server (DNS), File Transfer Protocol (FTP) server, email server, a network controller, twenty-four personal computers (PCs), six IP phones and two printers. A hybrid of ring-star topology was used for the designed network and it was done in the Cisco packet tracer virtual laboratory. Devices’ IP addresses and routing protocols were configured and verified using appropriate commands. Four classes of traffics, HTTP, FTP, email and Internet Control Message Protocol (ICMP), were generated and simulated to assess the performance of the designed network. The Quality of Service (QoS) policy was implemented on the designed network through class-map, policy-map and service-policy. The QoS policy was imposed on routers 2, 4 and 5 while routers 1 and 3 were left because the two routers served as dynamic host configuration protocol servers for IP phones. The results showed that on router 2, the HTTP traffic experienced packets delay of about 33% while 1018 out of 1917 matched packets were successfully transmitted in FTP but with 3972 delayed packets. The ICMP traffic had all 1807 matched packets successfully sent without any drop. The email traffic experienced highest congestion as only 470 out of 2118 packets were successfully sent. Traffic analysis on router 4 revealed that HTTP traffic at 5 minutes offered rate of 1.854 kbps was associated with 1739 delayed packets while 815 packets were transmitted. The FTP traffic had 759 packets transmitted out of 759 matched while the ICMP traffic had 780 and 781 matched and transmitted packets, respectively. The email traffic was the worst hit by the imposed QoS on router 4 as out of 2095 matched packets, a paltry 81 were successfully transmitted. The relative values of matched to transmitted packets showed that all class of traffics fared better on router 5. The drop rates for HTTP, FTP, ICMP and email traffics were 0.418, 0.424, 11.6244 and 0.100 kbps, respectively, on router 5. These results suggested that the QoS favoured ICMP traffic followed by FTP traffic on router 2 while on router 4, the QoS supported HTTP, FTP and ICMP traffics than the email traffic. Observations on router 5 might be due to assigned bandwidths to each class of traffics considered. The results showed that, with imposed QoS, each of the HTTP, FTP, ICMP and email traffics experienced varying degrees of congestion at different routers while traversing the IP network. It can be concluded that the traffic congestion in Internet protocol networks could be mitigated via appropriate traffic shaping and dynamic bandwidth allocation.