• Journal of KIISE:Information Networking
• /
• v.32 no.4
• /
• pp.443-451
• /
• 2005

Stateful inspection devices must maintain flow information. These devices create the flow information also for network attack packets, and it can fatally inflate the dynamic memory allocation on stateful inspection devices under network attacks. The memory inflation leads to memory overflow and subsequent performance degradation. In this paper, we present a guideline to set the flow entry timeout for a stateful inspection device to remove harmful embryonic entries created by network attacks. Considering Transmission Control Protocol (TCP) if utilized by most of these attacks as well as legitimate traffic, we propose a parsimonious memory management guideline based on the design of the TCP and the analysis of real-life Internet traces. In particular, we demonstrate that for all practical purposes one should not reserve memory for an embryonic TCP connection with more than (R+T) seconds of inactivity where R=0, 3, 9 and $1\leqq{T}\leqq{2}$ depending on the load level.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.2 no.6
• /
• pp.333-351
• /
• 2008

Transmission control protocols have to overcome common problems in wireless networks. TCP employing both packet loss discrimination mechanism and available bandwidth estimation algorithm, known as the good existing solution, shows significant performance enhancement in wireless networks. For instance, TCP New Jersey which exhibits high throughput in wireless networks intends to improve TCP performance by using available bandwidth estimation and congestion warning. Even though it achieves 17% and 85% improvements in terms of goodput over TCP Westwood and TCP Reno, respectively, we further improve it by exploring maximized available bandwidth estimation, handling bit-error-rate error recovery, and effective adjustment of sending rate for retransmission timeout. Hence, we propose TCP NJ+, showing that for up to 5% packet loss rate, it outperforms other TCP variants by 19% to 104% in terms of goodput when the network is in bi-directional background traffic.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7B
• /
• pp.667-674
• /
• 2004

The performance of transmission control protocol (TCP) is largely dependent upon its loss recovery. Therefore, it is a very important issue whether the packet losses may be recovered without retransmission timeout (RTO) or not. Although TCP SACK can recover multiple packet losses in a window, it cannot avoid RTO if a retransmitted packet is lost again. In order to alleviate this problem, we propose a simple change to TCP SACK, which is called TCP SACK+ in simple. We use a stochastic model to evaluate the performance of TCP SACK+, and compare it with TCP SACK. Numerical results evaluated by simulations show that SACK+ can improve the loss recovery of TCP SACK significantly in presence of random losses.

• Journal of KIISE:Information Networking
• /
• v.32 no.3
• /
• pp.382-390
• /
• 2005

As today's networks evolve towards an If-based integrated network, the role of transmission control protocol(TCP) has been increasing as well. As a well-known issue, the performance of TCP is affected by its loss recovery mechanism that is comprised of two algorithms; fast retransmit and fast recovery. Although retransmission timeout(RTO) caused by multiple packet losses can be avoided by using selective acknowledgement(SACK) option, RTO cannot be avoided if a retransmitted packet is lost. Therefore, we propose a simple modification to make it possible for a TCP sender using SACK option to detect a lost retransmission. In order to evaluate the proposed algorithm, simulations have been performed for two scenarios where packet losses are random and correlated. Simulation results show that the proposed algorithm can improve TCP performance significantly.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.6B
• /
• pp.396-405
• /
• 2005

Snoop protocol is one of the efficient schemes to compensate TCP packet loss and enhance TCP throughput in wired-cum-wireless networks. However, Snoop protocol has a problem; it cannot perform local retransmission efficiently under the bursty-error prone wireless link. In this paper, we propose Enhanced Snoop(E-Snoop) protocol to solve this problem of Snoop protocol. With E-Snoop protocol, packet losses can be noticed by receiving new ACK packets as well as by receiving duplicate ACK packets or local retransmission timeout. Therefore, TCP throughput can be enhanced by fast recognition of bursty packet losses and fast local retransmissions. From the simulation results, E-Snoop protocol can improve TCP throughput more efficiently than Snoop protocol and can yield more TCP improvement especially in the channel with high packet loss rates.

• Journal of KIISE:Information Networking
• /
• v.34 no.6
• /
• pp.435-445
• /
• 2007

Snoop protocol is one of the efficient schemes to compensate TCP packet loss and enhance TCP throughput in wired-cum-wireless networks. However, Snoop protocol has a problem: it cannot perform local retransmission efficiently under the bursty-error prone wireless link. To solve this problem, SACK-Aware-Snoop and SNACK mechanism have been proposed. These approaches improve the performance by using SACK option field between base station and mobile host. However in the wireless channel with high packet loss rate, SACK-Aware-Snoop and SNACK mechanism do not work well because of two reason: (a) end-to-end performance is degraded because duplicate ACKs themself can be lost in the presence of bursty error, (b) energy of mobile device and bandwidth utilization in the wireless link are wasted unnecessarily because of SACK option field in the wireless link. In this paper, we propose a new local retransmission scheme based on Cross-layer approach, called Cross-layer Snoop(C-Snoop) protocol, to solve the limitation of previous localized link layer schemes. C-Snoop protocol includes caching lost TCP data and performing local retransmission based on a few policies dealing with MAC-layer's timeout and local retransmission timeout. From the simulation result, we could see more improved TCP throughput and energy efficiency than previous mechanisms.

• Journal of KIISE:Information Networking
• /
• v.36 no.5
• /
• pp.383-388
• /
• 2009

In this paper, we focus on the retransmission following transmission failure in impulse radio ultra wideband (IR-UWB). The reasons of transmission failure are classified and a new 'selfish' retransmission protocol is proposed because time hopping can support multiple transmissions at the same time. Selfish retransmission protocol retransmits packets immediately without any kind of timeout or channel observation. Simulation results show that the proposed protocol improves throughput up to 50% and decreases retransmission delay also up to 70%, compared to a conventional retransmission system in IR-UWB.

• 대한공업교육학회지
• /
• v.33 no.1
• /
• pp.233-248
• /
• 2008

Stream Control Transmission Protocol(SCTP) is a new transport protocol that is known to provide improved performance than Transmission Control Protocol(TCP) in multi-homing environment that is having two and more IP addresses. But currently single-homed computer is used primarily that is having one IP address. To identify whether mean transfer time for SCTP is faster that for TCP in single-homed environment considering packet loss, we make up real testbed regulating the bandwidth, delay time and packet loss rate on router and observe the transfer time. We write server and client applications to measure SCTP and TCP mean transfer time by C language. Analysis of these experimental results from the testbed implementation shows that mean transfer time of SCTP is not better than performance of TCP in single homed environment exceptional case. Main reasons of performance are that SCTP compared to TCP stops transmitting data by timeout and data transmission is often delayed when SACK congestion happens. The result of study shows that elaborate performance tuning is required in developing a new SCTP module or using a implemented SCTP module.

• Journal of Communications and Networks
• /
• v.6 no.3
• /
• pp.235-244
• /
• 2004

There have been a lot of approaches to evaluate and predict transmission control protocol (TCP) performance in a numerical way. Especially, under the recent advance in wireless transmission technology, the issue of TCP performance over wireless links has come to surface. It is because TCP responds to all packet losses by invoking congestion control and avoidance algorithms, resulting in degraded end-to-end performance in wireless and lossy systems. By several previous works, although it has been already proved that overall TCP performance is largely dependent on its loss recovery performance, there have been few works to try to analyze TCP loss recovery performance with thoroughness. In this paper, therefore, we focus on analyzing TCP's loss recovery performance and have developed a simple model that facilitates to capture the TCP sender's behaviors during loss recovery period. Based on the developed model, we can derive the conditions that packet losses may be recovered without retransmission timeout (RTO). Especially, we have found that TCP Reno can retransmit three packet losses by fast retransmits in a specific situation. In addition, we have proved that successive three packet losses and more than four packet losses in a window always invoke RTO easily, which is not considered or approximated in the previous works. Through probabilistic works with the conditions derived, the loss recovery performance of TCP Reno can be quantified in terms of the number of packet losses in a window.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.9
• /
• pp.2323-2340
• /
• 2012

Recent advances in cognitive radio technology have drawn immense attention to higher layer protocols above medium access control, such as transmission control protocol (TCP). Most proposals to improve the TCP performance in cognitive radio (CR) networks have assumed that either all nodes are in CR networks or the TCP sender side is in CR links. In those proposals, lower layer information such as the CR link status could be easily exploited to adjust the congestion window and improve throughput. In this paper, we consider a TCP network in which the TCP sender is located remotely over the Internet while the TCP receiver is connected by a CR link. This topology is more realistic than the earlier proposals, but the lower layer information cannot be exploited. Under this assumption, we propose an enhanced TCP protocol for CR networks called TCP for cognitive radio (TCP-CR) to improve the existing TCP by (1) detection of primary user (PU) interference by a remote sender without support from lower layers, (2) delayed congestion control (DCC) based on PU detection when the retransmission timeout (RTO) expires, and (3) exploitation of two separate scales of the congestion window adapted for PU activity. Performance evaluation demonstrated that the proposed TCP-CR achieves up to 255% improvement of the end-to-end throughput. Furthermore, we verified that the proposed TCP does not deteriorate the fairness of existing TCP flows and does not cause congestions.