• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10c
• /
• pp.175-177
• /
• 2004

무선망에서의 TCP 성능은 않은 연구가 되어 왔지만 손실율이 큰 경우의 TCP의 성능은 아직도 개선의 여지가 많다. 본 논문은 두 가지 점에서 새로운 TCP를 제안한다. 첫째, 혼잡제어(Congestion Control)와 손실제어(Loss Recovery)가 분리(Decoupled)되었다. 기존의 연구들은 무선손실과 혼잡손실을 구분하여야 한다는 점에는 많은 연구를 해왔지만 무선 손실에 대하여 어떻게 대처할 것 인가에 관한 문제는 상대적으로 간과하였다. 둘째, 기종의 TCP-Westwood와 TCP-Jersey에서 사용되는 ABE(Available Bandwidth Estimation)가 무선네트워크에 의해 발생하는 패킷 손실로 인해 부정확해지는 것을 지적하고 새로운 ABE방법을 제시한다. 시뮬레이션을 통하여 우수한 성능을 보여주었다.

• Journal of Communications and Networks
• /
• v.6 no.4
• /
• pp.376-386
• /
• 2004

In ad hoc networks, loss-based congestion window progression by the traditional means of duplicate ACKs and timeouts causes high network buffer utilization due to large bursts of data, thereby degrading network bandwidth utilization. Moreover, network-oriented feedbacks to handle route disconnection events may impair packet forwarding capability by adding to MAC layer congestion and also dissipate considerable network resources at reluctant intermediate nodes. Here, we propose a new TCP scheme that does not require the participation of intermediate nodes. It is a purely end-to-end scheme using TCP timestamps to deduce link conditions. It also eliminates spurious reductions of the transmission window in cases of timeouts and fast retransmits. The scheme incorporates a receiver-oriented rate controller (rater), and a congestion window delimiter for the 802.11 MAC protocol. In addition, the transient nature of medium availability due to medium contention during the connection time is addressed by a freezing timer (freezer) at the receiver, which freezes the sender whenever heavy contention is perceived. Finally, the sender-end is modified to comply with the receiver-end enhancements, as an optional deployment. Simulation studies show that our modification of TCP for ad hoc networks offers outstanding performance in terms of goodput, as well as throughput.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.11a
• /
• pp.1482-1484
• /
• 2010

A new TCP protocol can succeed for large bandwidth delay product when it meets network bandwidth utilization efficiency and fair sharing. We introduce a novel congestion control algorithm which employs queueing delay information in order to calculate the amount of congestion window increment in increase phase, and reduces congestion window to optimal estimated bound as packet loss occurs. Combination of such methods guarantees that the proposal utilizes fully network bandwidth, recovers quickly from packet loss in wireless link, and preserves fairness for competing flows mixed short RTT and long RTT. Our simulations show that features of the proposed TCP meet the desired requirements.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.121.1-121
• /
• 2001

Congestion control is a critical issue in TCP networks, Recently, active queue management (AQM) was proposed for congestion control at routers. The random early detection RED algorithm is widely known in the AQM algorithms, We present an adaptive nonlinear RED (NRED) algorithm, which has nonlinear drop probability profile. The proposed algorithm enhanced the performance of the RED algorithm by the self-parameterization based on the traffic load Furthermore, the proposed algorithm can effectively adapt itself between he RED and the drop-tail queue management by adopting proper nonlinearity in the drop probability profile. Through simulation, we show the effectiveness of the proposed algorithm comparing with the drop-tail and the original RED algorithm.

• Journal of Internet Computing and Services
• /
• v.8 no.1
• /
• pp.15-23
• /
• 2007

Wireless LAN (IEEE 802.11) uses traditional TCP for reliable data transmission, But it brings the unintentional packet loss which is not congestion loss caused by handoff, interference, and fading in wireless LAN. In wireless LAN, TCP experiences performance degradation because it consumes that the cause of packet loss is congestion, and it decrease the sending rate by activating congestion control algorithm. This paper analyzes that correlation of throughput and buffer size for wireless buffer tuning. We find MBT (Maximum Buffer Threshold) which does not increase the throughput through the analysis, For calculation of MBT, we experiment the throughput by using high volume music data which is creased by real-time performance of piano. The experiment results is shown that buffer tuing based on MBT shows 20.3%, 21.4%, and 45.4% throughput improvement under 5ms RTT, 10ms RTT, and 20ms RTT, respectively, comparing with the throughput of operation system default buffer size, In addition, we describe that The setting of TCP buffer size by exceeding MBT does not have an effect on the performance of TCP.

• International Journal of Control, Automation, and Systems
• /
• v.5 no.1
• /
• pp.70-78
• /
• 2007

We prove that TCP Vegas globally converges to its equilibrium point in finite time assuming no feedback information delay. We analyze a continuous-time TCP Vegas model with discontinuity and high nonlinearity. Using the upper right-hand derivative and applying the comparison lemma, we cope with the discontinuous signum function in the TCP Vegas model; using a change of state variables, we deal with the high nonlinearity. Although we ignore feedback information delay in analyzing the model of TCP Vegas, the simulation results illustrate that TCP Vegas in the presence of feedback information delay shows very similar dynamic trends to TCP Vegas without feedback information delay. Consequently, dynamic properties of TCP Vegas without feedback information delay can be used to estimate those of TCP Vegas in the presence of feedback information delay.

• Journal of Communications and Networks
• /
• v.9 no.3
• /
• pp.236-246
• /
• 2007

This paper investigates the performance of the transmission control protocol (TCP) transport protocol over IEEE 802.11 infrastructure based wireless networks. A wireless link is generally characterized by high transmission errors, random interference and a varying latency. The erratic packet losses usually lead to a curbing of the flow of segments on the TCP connection and thus limit TCP's performance. This paper examines the impact of the lossy nature of IEEE 802.11 wireless networks on the TCP performance and proposes a scheme to improve the performance of TCP over wireless links. A negative acknowledgment scheme, selective negative acknowledgment (SNACK), is applied on TCP over wireless networks and a series of ns-2 simulations are performed to compare its performance against that of other TCP schemes. The simulation results confirm that SNACK and its proposed enhancement SNACK-S, which incorporates a bandwidth estimation model at the sender, outperform conventional TCP implementations in 802.11 wireless networks.

• Journal of KIISE:Information Networking
• /
• v.30 no.5
• /
• pp.649-655
• /
• 2003

Network-based computing such as cluster computing requires a reliable high-speed transport protocol. TCP is a representative reliable transport protocol on the Internet, which implements many mechanisms, such as flow control, congestion control, and retransmission, for reliable packet delivery. This paper, however, finds out that Myrinet does not incur any packet losses caused by network congestion. In addition, we ascertain that Myrinet supports reliable and ordered packet delivery. Consequently, most of reliable routines implemented in TCP produce unnecessarily additional overheads on Myrinet. In this paper, we show that we can attain the reliability only by flow control on Myrinet and propose a new reliable protocol based on UDP named RUM (Reliable UDP on Myrinet) that performs a flow control. As a result, RUM achieves a higher throughput by 45% than TCP and shows a similar one-way latency to UDP.

• Journal of KIISE:Information Networking
• /
• v.34 no.6
• /
• pp.466-474
• /
• 2007

Asymmetric networks that the downlink bandwidth is larger than the uplink bandwidth may cause the degradation of the TCP performance due to the uplink congestion. In order to solve this problem, this paper designs and implements the Dynamic Segment Size Control mechanism which offers a suitable segment size for current networks. The proposed mechanism does not require any changes in customer premises but suppress the number of ACKs using segment reassembly technique to avoid the uplink congestion. The gateway which adapted the Dynamic Segment Size Control mechanism, detects the uplink congestion condition and dynamically measures the bandwidth asymmetric ratio and the packet loss ratio. The gateway reassembles some of segments received from the server into a large segment and transmits it to the client. This reduces the number of corresponding ACKs. In this mechanism, the SACK option is used when occurs the bit error during the transmission. Based on the simulation in the GEO satellite network environment, we analyzed the performance of the Dynamic Segment Size Control mechanism.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.3B
• /
• pp.487-492
• /
• 2000

The ABR(available bit rate) service has been developed to support data application over ATM(asynchronous Transfer mode) networks, and TCP is the most widely used transport layer protocol in existing data networks. Thus, the performance improvement of TCP over ABR is an issue of cardinal importance. In this paper, we propose algorithm of the adjustment of congestion window size utilizing RM(resource Management) cell information. The motivation is to use more reliable information in ER(explicit rate) field within RM cell to determine TCP window size. Simulations are performance using the suggested algorithm is improved, as compared to TCP-reno.