Recently, the transport networks must cope with the ever increasing traffic inflation of the high capacity and reliable systems. The demand is increased by many different factors, such as the rapid growth of the Internet, voice, data, videoconferencing, and private networking services. All of them consume large amount of bandwidth which results in dynamically-increased need for bandwidth in networks. Fortunately, optical network [1] which provides high-speed processing and huge amount of capacity based on WDM technology becomes the best solution.Most researches in the field of optical networks have focused on the mesh topology systems. But in recent years, various researches have proposed the optical networks using WDM ring topologies, such as Synchronous Optical Network/Synchronous Digital Hierarchy (SONET/SDH) rings [2], Resilient Packet Ring (RPR) and etc. The ring topologies are attractive for their simple router interface and high-speed communication. In addition, ring topology allows easy addition and removal of nodes at arbitrary locations [3] [4], which supplies good expandability.There are several kinds of node architectures that can be applied to the optical ring networks. The node architectures can be classified into FT-FR (Fixed Transmitter-Fixed Receiver), FT-TR (Fixed Transmitter-Tunable Receiver), and TT-TR (Tunable Transmitter-Tunable Receiver). In the FT-FR, each node can only access the predetermined channel and network cost is increased because many transceivers are needed to guarantee the efficiency of network resources. The FT-TR suffers from a waste of bandwidth because the receiver collision is occurred if multiple bursts arrive at the same destination simultaneously. The TT-TR has a switching overhead to coordinate between source and destination and may have problems due to resource contention. In recent, the Collision-free Optic-burst Ring Network (CORNet) has been proposed to overcome these problems and to use network resource more efficiently [5].In CORNet, multiple nodes share the same drop-wavelength for receiving data burst by using FR and each node can use any data wavelength for transmission by TT. In addition, a fixed-tuned transmitter is dedicated to each node's assigned drop wavelength. The CORNet employs the Multi-token Access Control (MACA) to avoid the channel collision. In this protocol, multiple tokens are used to manage the accessibility of all data wavelengths. One problem in CORNet is that the utility of network resources is limited as tokens are synchronously transmitted around the ring and each node must wait until an empty available bandwidth between two adjacent tokens before transmitting bursts. Thus the collision-free is guaranteed by the synchronous management. However, the network performance, such as the throughput or the resource utilization, is limited.Therefore, in order to improve the throughput and the average queuing delay of the CORNet, this paper focuses on improving the MAC protocol to reuse the drop-wavelengths in the network. We propose new protocol named PRADA which stands for Protocol for Reusing Available Drop-wavelength Appropriately. The PRADA can be implemented based on three different queue selection schemes to reuse the drop-wavelengths: Best-Fit (BF), Farthest sub-Burst (FB), and Neighbor sub-Burst (NB). The performance of the PRADA with different queue selection schemes is evaluated and compared in terms of throughput and average queuing delay through OPNET simulation.
인터넷, 화상통신 및 VoIP (Voice over IP) 같은 다양한 응용 서비스의 사용량이 증가하고 있으며, 이러한 응용 서비스의 증가를 수용하기 위하여 차세대 네트워크는 고속 및 효율적인 전송능력이 요구된다. 그러나 기존의 메트로 네트워크는 음성 트래픽에 최적화되어 있기 때문에 고속의 백본 네트워크와 가입자망 사이에서 병목현상이 발생한다. 따라서 이러한 문제를 해결하기 위하여 링 토폴로지 기반의 광 버스트 교환방식(OBS)을 이용하는 네트워크에 대한 연구가 활발히 진행되고 있다.기존 OBS 링 네트워크의 충돌 문제를 해결하기 위하여 Collision-free Optic-burst Network (CORNet)이 제안되었다. CORNet은 TT-FT-FR (Tunable Transmitter-Fixed Transmitter -Fixed Receiver)를 사용하는 노드 구조를 가지며 파장 충돌 문제를 해결하기 위하여 다중 토큰을 이용하여 모든 파장에 접근을 관리하기 위한 다중 토큰 접근 제어 (Multi-token Access Control: MACA) 방식을 사용하였다. 그러나 MACA는 토큰이 동기적으로 링에 전송되며 각 노드는 데이터 버스트를 전송 하기 전에 두 인접 토큰 사이에 사용 가능한 파장이 있을 때까지 기다려야만 하기 때문에 CORNet은 네트워크 자원을 효율적으로 사용하지 못하고 큐잉 지연시간이 증가하는 문제점을 갖는다.따라서 본 논문에서는 CORNet에서 효율적 네트워크 자원 사용과 평균 큐잉 지연 시간을 향상시키기 위하여 파장 재사용을 위한 유효 수신 파장 재사용 매체 접근 제어 프로토콜 (Protocol for Reusing Available Drop-wavelength Appropriately: PRADA)을 제안한다. 제안한 프로토콜은 최적 버스트(Best-Fit: BF), 장거리 버스트 (Farthest sub-Burst: FB) 및 근거리 버스트(Neighbor sub-Burst : NB)를 큐에서 선택하여 전송하는 세가지 큐 선택 알고리즘을 사용하여 설계한다. 또한 제안한 프로토콜의 성능 평가를 위하여 OPNET 시뮬레이터를 이용한 평균 큐잉 지연과 망 성능을 측정을 수행하고 기존의 기법과 비교 분석한다.