International Journal of Computer Applications (0975 – 8887)
Volume 27– No.8, August 2011

Random Early Discard (RED-AQM) Performance Analysis in Terms of TCP Variants and Network Parameters:
Instability in High-Bandwidth-Delay Network
Mohammad Abu Obaida

Md. Sanaullah Miah

Md. Abu Horaira

Dept. of CSE, DUET
Gazipur-1700,
Bangladesh

Dept. of CSE, DUET
Gazipur-1700,
Bangladesh

Software Engineer
DataSoft BD Limited
Chittagong, Bangladesh

ABSTRACT
Conventional congestion control methods (e.g. DROP TAIL) discards all received packets after the queue is full moreover results in low-network performance. To address this problem,
RED was proposed to improve the performance of TCP connections. As a queue management mechanism, it drops packets in the considered router buffer to adjust the network traffic behavior according to the queue size. In application, TCP
Variants (Reno, NewReno, Vegas, Fack and Sack1) show oscillatory curve of packet reception if RED is considered for queuing, besides, some variants out performs in receiving packets over different network parameters that this paper analyzes and finds out. However, an increase in link capacity
(with the resulting increase of per-flow bandwidth) will cause significant degradation in TCP‟s performance, irrespective of the queuing scheme used. Hence the network is prone to instability with the rise in the number of High-bandwidth-delay product that is also attended to in this paper.

General Terms
Network Congestion Control, TCP Variants, Network
Parameters, Queuing, Drop Tail, High-bandwidth-delay and
Random Early Marking.

Keywords
RED, AQM, BW, TCP Variants, NS-2.

1. INTRODUCTION
Random Early Marking puts forward a great deal of improvement for the purpose of Congestion Control rather than conventional Drop Tail mechanism does. The rationale behind the discrepancy in RED performance for different TCP Variants is that each of the Variants type possesses some special decisive factors. Such as, the base TCP has become known as TCP
Tahoe. TCP Reno attaches one novel mechanism called Fast
Recovery to TCP Tahoe [6]. In addition, TCP Newreno employs the most recent retransmission mechanism of TCP Reno. [7].
The use of Sacks allows the receiver to stipulate several additional data packets that have been received out-of-order within one dupack, instead of only the last in order packet received [8]. TCP Vegas offers its own distinctive

retransmission and congestion control strategies. TCP Fack is
Reno TCP with forward acknowledgment [9]. Transmission
Control Protocol (TCP) Variants Reno, NewReno, Vegas, Fack and Sack1 are implemented in NS-2. RED supervises the average queue size and drops packets based on statistical likelihoods [2].

2. RANDOM EARLY MARKING
2.1 RED, an improvement over Drop Tail
Droptail network components discard packets when its FIFO queue is full. Under heavy load conditions, droptail routers grounds global synchronization, a phenomenon in which all senders sharing the same bottleneck router/link shut down their transmission windows at almost the same time, thereby causing a abrupt drop in the bottleneck link exploitation [10]. Besides, droptail routers are biased against bursty sources [11]. This is because, when a burst of packets from a sender arrives at a fully occupied queue, a sustained packet drop within the same window of data occurs. Floyd et al.
[12] and Xu and Ansari [13] showed that TCP [14], lacks the capability to recuperate from such multiple packet losses within the same window of data. Therefore, the TCP sender has to rely on retransmission timeouts to recover the lost packets that considerably slows down the transmission rate of a TCP flow. On the Contrary, RED monitors the average queue size and drops packets based on statistical probabilities [2]. As the queue grows, the probability for dropping an incoming packet grows too. RED is shown to effectively tackle both the global synchronization problem and the problem of bias against
bursty