• 한국통신학회논문지
• /
• 제28권11A호
• /
• pp.867-875
• /
• 2003

기존의 단대단 혼잡제어는 단순히 중복된 ACK 신호를 이용하여 혼잡을 처리하며 일반적으로 혼잡을 제어하는데 시간이 많이 소요된다. 이러한 메커니즘은 혼잡발생시에 TCP의 혼잡 윈도의의 크기가 동기화 되는 현상을 피할 수 없으며, 더욱이 RTT가 증가될 경우 중복된 ACK 패킷을 받기 전에 혼잡이 사라지거나 받기 전까지 전송되는 패킷으로 인해서 정확한 혼잡제어를 할 수 없다. 최근에 이러한 단대단 혼잡제어의 문제점을 개선하는 다양한 AQM(Active Queue Management)를 소개되고 있으며, 기본적으로 RED의 혼잡 제어 메커니즘을 사용한다. 본 논문에서는 우선 다양한 전송률과 혼잡에 반응하는 흐름(TCP)과 반응하지 않는 흐름(UDP)을 갖는 트래픽 상에서 RED, CHOKe, FRED 그리고 SRED등과 같은 AQM의 효과를 평가한다. 특히, CHOKe 메커니즘의 경우 혼잡에 반응하지 않는 흐름이 증가 할 경우 공평성이 떨어지는 단점을 갖는다. 따라서, 본 논문에서는 UDP 트래픽으로 부터 TCP 트래픽을 보호하기 위해 UPD 트래픽과 TCP 트래픽을 두개의 독립된 논리적인 큐에서 서로 다른 CHOKe 메커니즘으로 처리하는 이중 큐 CHOKe 메커니즘을 제안한다. 각 흐름의 정보를 유지하기 위해 LRURED에서 제안한 부분상태 정보(Partial state information)를 이용하여 트래픽 유형별로 구별하고 격리하여 보다 효율적인 혼잡제어를 제시한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제3권6호
• /
• pp.575-596
• /
• 2009

Facing limited network resources such as bandwidth and processing capability, the Internet will have congestion from time to time. In this paper, we propose a scheme to maximize the total utility offered by the network to the end user during congested times. We believe the only way to achieve our goal is to make the scheme application-aware, that is, to take advantage of the characteristics of the application. To make our scheme scalable, it is designed to be class-based. Traffic from applications with similar characteristics is classified into the same class. We adopted the RED queue management mechanism to adaptively control the traffic belonging to the same class. To achieve the optimal utility, the traffic belonging to different classes should be controlled differently. By adjusting link bandwidth assignments of different classes, the scheme can achieve the goal and adapt to the changes of dynamical incoming traffic. We use the control theoretical approach to analyze our scheme. In this paper, we focus on optimizing the control on two types of traffic flows: TCP and Simple UDP (SUDP, modeling audio or video applications based on UDP). We derive the differential equations to model the dynamics of SUDP traffic flows and drive stability conditions for the system with both SUDP and TCP traffic flows. In our study, we also find analytical results on the TCP traffic stable point are not accurate, so we derived new formulas on the TCP traffic stable point. We verified the proposed scheme with extensive NS2 simulations.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 1999년도 가을 학술발표논문집 Vol.26 No.2 (3)
• /
• pp.423-426
• /
• 1999

In offering a statistical end-to-end bandwidth guarantee service, typically called Assured Service, in Differentiated Serviced (Diff-Serv) framework, the biggest issue is its inconsistency. Larger profile TCP flows fail to achieve the guaranteed rate when competing with many smaller profile flows. This phenomenon, which we call "bandwidth skew", stems from the fact that larger profile flows take longer time to recover from the congestion window size backoff after a packet drop. Proposed solutions to this problem, therefore, are focused on modifying the TCP behavior. However, TCP modification is not practicable, mainly due to its large installation base. We look to other mechanisms in the Diff-Serv framework to find more realistic solutions. In particular, we demonstrate that RIO, the de facto standard packet differentiation mechanism used for Assured Service, also contributes to the bandwidth skew. Based on this new finding, we design a modified RIO mechanism called RI+O. RI+O uses OUT queue length in addition to IN and IN+OUT queue length to calculate OUT packet drop probability. We show through extensive simulation that RI+O significantly alleviates the bandwidth skew, expanding the operating regime for Assured Service.d Service.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권1호
• /
• pp.149-165
• /
• 2013

Current group key agreement protocols, which are often tree-based, have unnecessary delays that are caused when members with low-performance computer systems join a group key computation process. These delays are caused by the computations necessary to balance a key tree after membership changes. An alternate approach to group key generation that reduces delays is the dynamic prioritizing mechanism of queue-based group key generation. We propose an efficient group key agreement protocol and present the results of performance evaluation tests of this protocol. The queue-based approach that we propose is scalable and requires less computational overhead than conventional tree-based protocols.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권7호
• /
• pp.1737-1753
• /
• 2013

Current group key agreement protocols, which are often tree-based, have unnecessary delays that are caused when members with low-performance computer systems join a group key computation process. These delays are caused by the computations necessary to balance a key tree after membership changes. An alternate approach to group key generation that reduces delays is the dynamic prioritizing mechanism of queue-based group key generation. We propose an efficient group key agreement protocol and present the results of performance evaluation tests of this protocol. The queue-based approach that we propose is scalable and requires less computational overhead than conventional tree-based protocols.

• 한국통신학회논문지
• /
• 제32권8B호
• /
• pp.547-552
• /
• 2007

IEEE는 E-PON에 대한 구현 표준으로 IEEE802.3ah을 제정하였으나 구체적인 상향대역할당 메커니즘에 대해서는 별도로 규정하지 않고 제조업체의 몫으로 남겨 두었다. 이에 관련 전문가들은 E-PON에서의 상향 전송에 대한 성능 개선 및 다양한 서비스 수용을 위해서 여러 방식의 상향대역할당 기법들을 연구하였다. 본 논문에서는 상향 트래픽 전송 성능을 개선하는 동적대역할당 메커니즘을 제안한다. 이 메커니즘은 ONU의 상향전송 큐별로 최소보장 상향대역을 다양하게 설정할 수 있도록 하였으며, 큐에 대기 중인 패킷을 전송하기 위하여 큐별로 가중치를 부여한 DRR(Deficit Round Robin) 기법을 적용하였다. 시뮬레이션에 의해 상향대역 이용률이 기존방식 대비 10% 이상 개선됨을 보여 주었다.

• Journal of Communications and Networks
• /
• 제7권3호
• /
• pp.337-352
• /
• 2005

In this paper, a fuzzy logic implementation of the random early detection (RED) mechanism [1] is presented. The main objective of the proposed fuzzy controller is to reduce the loss probability of the RED mechanism without any change in channel utilization. Based on previous studies, it is clear that the performance of RED algorithm is extremely related to the traffic load as well as to its parameters setting. Using fuzzy logic capabilities, we try to dynamically tune the loss probability of the RED gateway. To achieve this goal, a two-input-single-output fuzzy controller is used. To achieve a low packet loss probability, the proposed fuzzy controller is responsible to control the $max_{p}$ parameter of the RED gateway. The inputs of the proposed fuzzy controller are 1) the difference between average queue size and a target point, and 2) the difference between the estimated value of incoming data rate and the target link capacity. To evaluate the performance of the proposed fuzzy mechanism, several trials with file transfer protocol (FTP) and burst traffic were performed. In this study, the ns-2 simulator [2] has been used to generate the experimental data. All simulation results indicate that the proposed fuzzy mechanism out performs remarkably both the traditional RED and Adaptive RED (ARED) mechanisms [3]-[5].

• 전자공학회논문지CI
• /
• 제45권5호
• /
• pp.125-133
• /
• 2008

본 논문은 다양한 서비스 클래스가 제공되는 BcN (Broadband convergence Network) 환경에서 상위 우선순위 서비스의 품질을 보장하기 위한 능동 큐 관리 메커니즘 (Active-WRED)을 제안한다. 네트워크 혼잡 상황에서 제안된 메커니즘은 상대적으로 낮은 우선순위를 갖는 서비스의 drop 확률을 증가시키고 이에 따라 상위 우선순위 서비스의 drop 확률을 감소시킴으로써 상위 우선순위 서비스에 대해 패킷 손실로부터 품질을 보장할 수 있다. 제안된 능동 큐 관리 메커니즘의 성능을 분석하기 위해 이산 시간 큐잉 시스템의 통계적 분석 방법을 도입하였다. 두 서비스 클래스 (GS: Guaranteed Service, BS: Best Effort Service)만을 고려하여 제안 방안의 성능 분석을 위해 MMBP-2 (two-state Markov-Modulated Bernoulli arrival Process) 트래픽 소스 모델과 2차 이산시간 Markov 체인을 도입하였다.

• Management Science and Financial Engineering
• /
• 제1권1호
• /
• pp.1-19
• /
• 1995

In this paper, we analyze an M / M / 1 queueing system where there are incentive conflicts among customers. Self-interested customers' decisions whether to join the system or not may not necessarily induce a socially optimal congestion level. As a way to alleviate the over-congestion, toll imposition was used in Naor's paper [3]. Instead of using a toll mechanism, we study the usefulness of imperfect information on system state (queue size, for example) as a way to reduce the over-congestion by self-interested customers. The main conclusion of this paper is that by purposefully giving fuzzy or imperfect information on the current queue size we can improve the congestion in the system. This result might look contradictory to rough intuition since perfect information should give better performance than imperfect information. We show how this idea is verified. In deriving this result, we use the concept of Nash equilibrium (pure and mixed strategy) as introduced in game theory. In some real situations, using imperfect information is easier to apply than imposing a toll, and thus the result of this paper has practical implications.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1994년도 추계학술대회 논문집
• /
• pp.664-669
• /
• 1994

This paper focuses on development and implementation of a performance management algorithm for IEEE802.4 token bus networks to serve large-scale integrated manufacturing systems. Such factory automation networks have to satisgy delay constraints imposed on time-critical messages while maintaining as much network capacity as possible for non-time-critical messages. This paper presents a network perfomance manager that adjusts queue apacity as well as timers by using a set of fuzzy rules and fuzzy inference mechanism. The efficacy of the performance management has been demonstrated by a series of simulation experiments.