• The KIPS Transactions:PartC
• /
• v.17C no.6
• /
• pp.451-458
• /
• 2010

To detect and recover packet losses over wireless links is very important in terms of reliability in packet transmission. Most wireless communication systems adopt an automatic repeat request (ARQ) protocol operating at link layer. However, it has been constantly addressed that the interaction not harmonized sufficiently between ARQ and TCP rather degrades TCP performance. In this paper, therefore, we propose an improved scheme from the aspect of the interaction with TCP loss recovery mechanism that can be applied to selective repeat ARQ (SR-ARQ) protocol and prove that the proposed scheme improves TCP performance significantly by OPNET simulations.

• Journal of KIISE:Information Networking
• /
• v.27 no.3
• /
• pp.247-255
• /
• 2000

최근 이동통신 서비스의 보급이 날로 증가됨에 따라 무선 접속 인터넷 서비스이 사용에 대한 요구가 급증하고 있다. 그러나 인터넷에서 사용되는 TCP 프로토콜은 에러 발생율이 낮은 유선망을 고려하여 설계되었기 때문에 망에서 발생되는 패킷 손실을 망내의 폭주로 인한 것으로 가정하고 폭죽제어 알고리즘을 동작시켜 윈도우 크기를 줄인다. 그러나 무선통신망과 같이 에러 발생율이 높은 환경에서는 패킷 손실이 주로 엘 발생에 기인하는데, 이 경우 기존의 TCP 프로토콜을 사용하면 폭주제어 알고리즘이 동작되어 TCP의 성능을 저하시키는 문제점이 발생된다. 따라서 본 논문에서는 유무선 복합망에서 TCP 프로토콜의 성능을 개선하기 위한 Split ACKs 기법을 제안하였다. 이 기법은 기지국에서 무선링크의 패킷 손실 이후에 수신된 ACK 패킷을 여러 개로 쪼개서 TCP 송신측으로 전달한다. 따라서 여러 개의 ACK 패킷을 수신한 TCP 송신측은 폭죽제어 알고리즘이 동작되어 감소시킨 윈도우 크기를 빠르게 복귀시켜 주기 때문에, 저하된 TCP 프로토콜의 성능을 신속히 향상시킬 수 있다. 아울러 제안된 기법은 기존 TCP 프로토콜을 그대로 사용할 수 있으며, TCP의 End-to-end Semantics가 유지되는 장점이 있다. 시뮬레이션을 통한 성능분석 결과 이 기법은 기존의 TCP 프로토콜에 비해 약 20%의 성능향상을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.11a
• /
• pp.978-981
• /
• 2007

TCP(Transmission Control Protocol)는 흐름제어와 오류제어 그리고 혼잡제어 등을 통해 목적지까지 효율적으로 데이터를 전송하는 프로토콜이다. 최근 라우터 성능 및 라우팅 알고리즘 등의 발달로 인해 패킷이 비순차적으로 전송되는 패킷 재배치(packet reordering)가 빈번한 발생하여 TCP 성능 저하의 주요 원인이 되고 있다. 패킷 재배치가 TCP 로 하여금 불필요한 재전송(Spurious Retransmission) 및 혼잡제어를 수행하도록 하기 때문이다. 본 논문에서는 TCP 의 신속 재전송(Fast Retransmission)의 중복 응답 임계값(dupthresh)를 동적으로 조절함으로써, 불필요한 재전송 및 혼잡제어의 발생횟수를 효율적으로 줄이는 TCP-DAD(TCP Dynamically Adjusted Dupthresh)기법을 제안한다. 본 논문은 NS-2를 이용한 시뮬레이션을 통해 TCP-DAD 가 다른 TCP 메커니즘보다 평균적으로 약 34%의 성능 향상을 보임을 증명하였다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.10
• /
• pp.789-797
• /
• 2003

An implementation scheme and some improvements are proposed to adopt public-licensed operating system, Linux and de-facto world-wide network standard, TCP/IP into the field of behavior-based autonomous mobile robots. To demonstrate the needs of scheme and the improvement, an analysis is performed on a server/client communication problem with real time Linux previously proposed, and another analysis is also performed on interactions among TCP/IP communications and the performance of Linux system using them. Implementation of behavior-based control architecture on real time Linux is proposed firstly. Revised task-scheduling schemes are proposed that can enhance the performance of server/client communication among local tasks on a Linux platform. A new method of TCP/IP packet flow handling is proposed that prioritizes TCP/IP software interrupts with aperiodic server mechanism as well. To evaluate the implementation scheme and the proposed improvements, performance enhancements are shown through some simulations.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.1286-1287
• /
• 2007

무선 네트워크에서 TCP Vegas는 다양한 원인으로 발생하는 패킷 손실로 인해 전송률이 떨어진다. 다수의 패킷 손실이 발생하면 TCP Vegas의 혼잡제어 기법은 패킷 손실로 인해 혼잡윈도우가 증가되지 않는다. 본 논문은 TCP Vegas의 혼잡제어 기법을 수정하여 무선 네트워크에서 TCP Vegas의 성능을 향상시킨 wirelessVegas를 제안한다. wirelessVegas는 NS-2를 이용하여 시뮬레이션 한 결과 TCP Vegas 보다 1%의 무선 링크 에러율에서 약 20% 높은 성능을 보인다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04d
• /
• pp.469-471
• /
• 2003

TCP는 패킷손실이 발생할 경우에 이를 혼잡에 의한 것이라고 판단하여 혼잡제어를 수행하는 프로토콜이다. 그러나 무선망은 그 특성상 에러가 많기 때문에 기존의 TCP프로토콜을 그대로 적용할 경우 불필요한 혼잡제어를 수행하게 되어 결국 전체 망의 성능을 떨어뜨리게 된다. 본 논문에서는 링크상태에 따라 송신률을 조절하는 TCP-ELSA 메커니즘 이용하여 링크상태 예측 방법을 개선하였다. TCP-ELSA는 무선링크의 링크상태를 예측하여 송신률을 조절하여 대역폭의 효율을 증가시키고 링크의 공평성을 보장하는 메커니즘이다. 본 논문에서 제안하는 개선된 TCP-ELSA는 링크상태 예측방법에 있어서 오류가 발생할 수 있는 경우를 고려함으로써 공평성을 유지하면서도 좀더 정확한 링크상태의 예측이 가능하도록 한다. 그리고 실시간 패킷을 비실시간 패킷과 구분하여 전송하도록 하여 Qos를 제공하였다.

• Journal of KIISE:Information Networking
• /
• v.29 no.5
• /
• pp.501-509
• /
• 2002

When using TCP over a mobile network, TCP responds to a handoff by invoking a congestion control algorithm, thereby resulting in a degraded end-to-end performance in a mobile network. In this paper, two schemes are proposed, TCP-MD and TCP-R. TCP-MD can detect the movement of s mobile host early on, whereas TCP-R can force the source to freeze data transmission during registration. The proposed schemes maintain end-to-end TCP semantics, making it possible to fully interoperate with the existing infrastructure. Only a small change is required in the mobile host, plus the implementation is simple because some Mobile IP messages are used to notify the handoff, eliminating the need for any additional messages. Simulations confirmed that the proposed schemes give an excellent performance under various environments.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.11B
• /
• pp.900-910
• /
• 2004

Route failure is mainly caused by mobility of mobile host in ad hoc networks. Route failure, which may lead to sudden packet losses and delays, is losing the route from source to destination. In this situation, TCP assumes that congestion has occurred within the network and also initiates the congestion control procedures. Congestion control algorithm provides the means for the source to deal with lost packets. TCP performance in ad hoc environments will be degraded as TCP source cannot distinguish congestion from route failure. In this paper, we propose TCP-P as pre-detection approach to deal with route failure. TCP-P freezes TCP through pre-detection of route failure. Route failure information of the proposed mechanism is obtained not by routing protocol but by MAC protocol. The intermediated node, obtaining route failure information by its MAC layer, relays the information to TCP source and lets TCP source stop the congestion control algorithm. Results reveal that TCP-P responding with proactive manner outperforms other approaches in terms of communication throughput under the presence of node mobility.

• Journal of Communications and Networks
• /
• v.8 no.4
• /
• pp.391-400
• /
• 2006

With the on-going wireless access technologies, the Internet has become accessible anytime anywhere. In wireless networks, link errors significantly degrade the performance of the transmission control protocol (TCP). To cope with this issue, this paper improves the recently-proposed terrestrial REFWA (T-REFWA) scheme by adding a new error recovery mechanism to its original design. In the T-REFWA scheme, senders are acknowledged with appropriate sending rates at which an efficient and fair utilization of network resources can be achieved. As the feedback values are computed independently of link errors, senders can keep transmitting data at high rates even in case of link error occurrences. Using this feature, the proposed error recovery mechanism can achieve high throughput in environments with high bit error rates. The throughput is further improved by disabling the exponential back-off algorithm of TCP so that long idle times are avoided in case of link errors. We show through simulations that the proposed method improves TCP performance in high bit error rates. Compared with several TCP variants, the proposed error recovery scheme exhibits higher link utilization and guarantees system fairness for different bit error rates.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.5
• /
• pp.257-266
• /
• 2006

Recently TCP/IP Offload Engine (TOE) technology, which processes TCP/IP on a network adapter instead of the host CPU, has become an important approach to reduce TCP/IP processing overhead in the host CPU. There have been two approaches to implementing TOE: software TOE, in which TCP/IP is processed by an embedded processor on a network adapter; and hardware TOE, in which all TCP/IP functions are implemented by hardware. This paper proposes a hybrid TOE that combines software and hardware functions in the TOE. In the hybrid TOE, functions that cannot have guaranteed performance on an embedded processor because of heavy load are implemented by hardware. Other functions that do not impose as much load are implemented by software on embedded processors. The hybrid TOE guarantees network performance near that of hardware TOE and it has the advantage of flexibility, because it is easy to add new functions or offload upper-level protocols of TCP/IP. In this paper, we developed a prototype board with an FPGA and an ARM processor to implement a hybrid TOE prototype. We implemented the hardware modules on the FPGA and the software modules on the ARM processor. We also developed a coprocessing mechanism between the hardware and software modules. Experimental results proved that the hybrid TOE prototype can greatly reduce the load on a host CPU and we analyzed the effects of the coprocessing mechanism. Finally, we analyzed important features that are required to implement a complete hybrid TOE and we predict its performance.