Fazli Rahim.* et al. Int. Journal of Engineering Research and Application www.ijera.com ISSN : 2248-9622, Vol. 8, Issue 1, ( Part -V) January 2018, pp.57-60 www.ijera.com DOI: 10.9790/9622-0801055760 57 | Page Minimizing Delay in Heterogeneous Networks using Congestion Forecast 1 Fazli Rahim, 2 Muhammad Fawad 1-2 Department of Computer Science, Government Postgraduate College Dargai Malakand Khyber pukhtoonkhwa Pakistan * Corresponding Author: Fazli Rahim ABSTRACT Internet, a collection of large many packets, is expanding day by day. Every minute across the globe, billions of electronic devices, such as PDA, Tablets, mobiles, Personal Computer, laptops, are interconnected via this gigantic networking system called the internet. This interconnection of so many devices across the globe, however, is not without any problems, such as congestion. Congestion is very hot area of research because internet traffic is expanding tremendously with the passage of time because each second, billions of billions of packets may enter the network, which are controlled by buffers. The aim of the paper is to study the behavior of heterogeneous networks to achieve minimized delay through congestion forecast, to control congestion through probabilistic nature of congestion forecast and maximized throughput. All the work will be carried out through Modeler Opnet. Keywords: Congestion Control, congestion forecast, heterogeneous networks, IPV4, IPV6. I. INTRODUCTION A system is said to be in congestion if the required resources exceed than the holding. The resources may be bandwidth and buffer space etc which are usually expensive. Even for a small file much more time is needed to download. Solving congestion through the use of infinite buffer is not a good idea because it increases queue size and compels the packet to stay in the queue for an infinite time so for each packet a finite lifetime should be assigned. [1]. The size of the internet is increasing day by day as the number of users is increasing on hourly basis. But this results in performance degradation and delay, and the loss of data packets due to congestion [2]. An internet is said to be a queue of packets where some nodes add packets while other removes it. A situation when the number of packets added to this queue exceeds the number of packets removed, congestion occurs. So to control it the packets should be added upto a specific level, to maintain the performance. Packet switching is important for this purpose. The sources of congestion are usually internet because of its connectionless nature, routers with more traffic than its capacity and LANs Connecting WANs. The variable size of packets makes traffic prediction and guaranteed service provision difficult. So QoS is not obtained. Congestion is usually caused by low memory, slow processors, bandwidth line and speed mismatch of devices. Congestion may be controlled using two methods i.e. Open Loop and Close Loop [3]. In open loop, preventing the system to enter the congestion by knowing the current traffic is restricted to reach the peak value while in closed loop, congestion is handled after it has been occurred. Queuing helps in congestion resolution. Queuing is used in buffer management and hence can improve network performance. Router Algorithms may be classified into queue management and scheduling algorithms. Queue management algorithms are related to control traffic and the other is for performance and handling delay Congestion control has two steps:  Avoiding the congestion not to occur  Recovering from the Congestion. Let the congestion occur and handle it after that. II. LITERATURE REVIEW Heterogeneous network can be categorized according to network traffics (IPV4, IPV6, Audio, Video, Text etc) or type of network (wired or wireless) or according to speed of network (high or low speed). The best effort service model is proposed by P.Gevros et al, in [1]. This scheme is treating all the packets equally, which guarantee that no compromise upon quality of service (QOS). The best effort model has operational specification and expected that packet should be delivered while keeping congestion levels low. In this paper most of the mechanisms for congestion controls are router- based which were almost in the perspective of guaranteed real time and QoS traffic. There has been too little research in the best effort service framework, so there is a general misunderstanding RESEARCH ARTICLE OPEN ACCESS