• Journal of Communications and Networks
• /
• 제16권5호
• /
• pp.548-558
• /
• 2014

Load distribution is vital to the performance of multipath transport. The task becomes more challenging in real-time multimedia applications (RTMA), which impose stringent delay requirements. Two key issues to be addressed are: 1) How to minimize end-to-end delay and 2) how to alleviate packet reordering that incurs additional recovery time at the receiver. In this paper, we propose sub-packet based multipath load distribution (SPMLD), a new model that splits traffic at the granularity of sub-packet. Our SPMLD model aims to minimize total packet delay by effectively aggregating multiple parallel paths as a single virtual path. First, we formulate the packet splitting over multiple paths as a constrained optimization problem and derive its solution based on progressive approximation method. Second, in the solution, we analyze queuing delay by introducing D/M/1 model and obtain the expression of dynamic packet splitting ratio for each path. Third, in order to describe SPMLD's scheduling policy, we propose two distributed algorithms respectively implemented in the source and destination nodes. We evaluate the performance of SPMLD through extensive simulations in QualNet using real-time H.264 video streaming. Experimental results demonstrate that: SPMLD outperforms previous flow and packet based load distribution models in terms of video peak signal-to-noise ratio, total packet delay, end-to-end delay, and risk of packet reordering. Besides, SPMLD's extra overhead is tiny compared to the input video streaming.

• 한국컴퓨터정보학회논문지
• /
• 제11권1호
• /
• pp.157-165
• /
• 2006

본 논문에서는 출력 버퍼형 탠덤크로스포인터 멀티캐스트 ATM 스위인 MTCOS 스위치를 제안한다. MTCOS 스위치는 라우팅 구조가 간단한 다수의 크로스포인터 스위치 패브릭으로 구성된 TCSF와 효율적 멀티캐스팅을 위한 집중화기 출력 버퍼로 구성된다. TCSF는 셀프 라우팅 크로스바 스위치가 갖는 셀 지연 편차 문제를 개선하고, 또한 하나의 입력에서 다수 출력 포트들로 다수의 동시 경로를 제공하며, 간단한 소프트웨어적 설정을 통해 다중 채널 스위칭을 제공하며, 확장성, 고성능, 모듈화 특성을 갖는다. MTCOS에서 제공되는 공유 트래픽 집중화 및 출력 큐잉 방식은 다양한 멀티캐스트 트래픽에 대해 낮은 셀 손실율과 낮은 지연시간을 보인다. 또한 동일 Knockout 집중화율을 달성하기 위해 SCOQ와 Knockout 멀티캐스트 스위치 보다 낮은 하드웨어 복잡도를 갖는다. 해석적 및 컴퓨터 시뮬레이션을 통해 임의의 트래픽에 대해 제안된 스위치가 높은 성능을 달성할 수 있음을 보였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.777-781
• /
• 2004

This paper presents an algorithm for multichannel slotted-ring topology medium access protocol (MAC) using in wavelength division multiplexing (WDM) networks. In multichannel ring, there are two main previously proposed architectures: Tunable Transmitter - Fixed Receiver (TTFR) and Fixed Transmitter - Tunable Receivers (FTTR). With TTFR, nodes can only receive packets on a fixed wavelength and can send packets on any wavelengths related to destination of packets. Disadvantage of this architecture is required as many wavelengths as there are nodes in the network. This is clearly a scalability limitation. In contrast, FTTR architecture has advantage that the number of nodes can be much larger than the number of wavelength. Source nodes send packet on a fixed channel (or wavelength) and destination nodes can received packets on any wavelength. If there are fewer wavelengths than there are nodes in the network, the nodes will also have to share all the wavelengths available for transmission. However the fixed wavelength approach of TTFR and FTTR bring low network utilization. Because source node with waiting data have to wait for an incoming empty slot on corresponding wavelength. Therefore this paper presents Tunable Transmitter - Tunable Receiver (TTTR) approach, in which the transmitting node can send a packet over any wavelengths and the receiving node can receive a packet from any wavelengths. Moreover, the self-similar distributed input traffic is used for evaluation of the performance of the proposed algorithm. The self-similar traffic performs better performance over long duration than short duration of the Poison distribution. In order to increase bandwidth efficiency, the Destination Stripping approach is used to mark the slot which has already reached the desired destination as an empty slot immediately at the destination node, so the slot does not need to go back to the source node to be marked as an empty slot as in the Source Stripping approach. MATLAB simulator is used to evaluate performance of FTTR, TTFR, and TTTR over 4 and 16 nodes ring network. From the simulation result, it is clear that the proposed algorithm overcomes higher network utilization and average throughput per node, and reduces the average queuing delay. With future works, mathematical analysis of those algorithms will be the main research topic.