Showing papers in "Photonic Network Communications in 2007"
TL;DR: This paper verifies the necessary and sufficient condition for achieving the complete fault localization (CFL) in MFDL, which is defined as the case that every single network fault can be located to a unique link.
Abstract: With the soaring channel speed and density in all-optical networks (AONs), the risk of high data loss upon network faults increases quickly. To manage network faults efficiently, an m-cycle based fault detection and localization (MFDL) scheme has been introduced recently. This paper verifies the necessary and sufficient condition for achieving the complete fault localization (CFL) in MFDL, which is defined as the case that every single network fault can be located to a unique link. We model the m-cycle construction as a new mathematical problem: the variant version of the constrained cycle-cover problem (vCCCP) and explore its formal expression. The model includes the consideration of the cycle-length limit, cycle number, and wavelength cost, while also keeps the CFL achievable. A two-phase branch-and-bound (B&B) algorithm was developed for solving the vCCCP, which guarantees to find near-optimal solutions. This algorithm is then applied to four typical and four random network examples to validate and assess the performance. The results are analyzed and compared with some previously reported algorithms, in terms of fault localization degree, cycle number, wavelength overhead, and cost reduction. The performance evaluation and comparison reveal that the new model and algorithm could significantly reduce the MFDL cost, including both the cost of monitoring devices and reserved wavelengths.
29 citations
TL;DR: This work characterises the burst-assembly delay distribution of each packet in a burst arisen by the main assembly algorithms found in the literature and introduces a new burst- assembly strategy that takes into account the particular delay constrains of packets in the formation of optical bursts.
Abstract: In Optical Burst Switching (OBS), packets travel through the network core as part of longer-size optical bursts, which do not suffer electronic conversion until they reach an eggress point. Typically, such optical bursts comprise tens or hundreds of packets, which are assembled/deassembled at border nodes. During the burst-formation process, each arriving packet must wait until the final burst is complete, which clearly adds an extra delay on each packet in the burst, especially on those arriving earlier. However, such burst-assembly delay may be excessive for the appropriate performance of certain applications, mainly real-time interactive ones. This work’s findings are twofold: first, it characterises the burst-assembly delay distribution of each packet in a burst arisen by the main assembly algorithms found in the literature; and, second, it introduces a new burst-assembly strategy that takes into account the particular delay constrains of packets in the formation of optical bursts, along with a detailed study of its properties.
17 citations
TL;DR: A novel ring-based local fault recovery mechanism, Multiple Ring-based Local Restoration (MRLR), for point-to-multipoint multicast traffic based on the minimum spanning tree (MST) in WDM mesh networks is proposed in this article.
Abstract: Optical networks with DWDM (Dense Wavelength Division Multiplex) can provide multiple data channels to supply high speed, high capacity to perform bandwidth-intensive multicast transmission service. Light-tree is a popular technique applied to support point-to-multipoint multicast services. Any failure during a multicast session would cause severe service loss or disruptions, especially when the faults occur near the source node. A novel ring-based local fault recovery mechanism, Multiple Ring-based Local Restoration (MRLR), for point-to-multipoint multicast traffic based on the minimum spanning tree (MST) in WDM mesh networks is proposed in this article. The MRLR mechanism dismembers the multicast tree into several disjoint segment-blocks (sub-trees) and reserves preplanned spare capacity to set up multiple protection rings in each segment-block for providing rapid local recovery. The MRLR scheme outperforms other methodologies in terms of the blocking probability, recovery time, and average hop count of protection path per session for different network topologies.
15 citations
TL;DR: A greedy algorithm is presented to reserve backup bandwidth considering multiple (F > 2) link (SRLG) failure scenarios and a bandwidth-saving joint selection scheme of working and protection paths is presented for protection against random multiple-link failures under dynamic traffic.
Abstract: Survivability is of critical importance in high-speed optical communication networks. A typical approach to the design of survivable networks is through a protection scheme that pre-determines and reserves backup bandwidth considering single/double link failure scenarios. In this article, a greedy algorithm is presented to reserve backup bandwidth considering multiple (F > 2) link (SRLG) failure scenarios. A bandwidth-saving joint selection scheme of working and protection paths is presented for protection against random multiple-link failures under dynamic traffic. Simulation shows that the algorithm can achieve maximum sharing of backup bandwidth for protection against random multiple-link failure with significant amount of bandwidth saving.
12 citations
TL;DR: Compared with simulation, the method was proved to be very accurate and much faster (up to four orders of magnitude faster than simulation in the studied cases).
Abstract: A novel, accurate, numerically stable and fast mathematical method for network-wide blocking probability evaluation of end-to-end optical burst switching (OBS) networks with heterogeneous link capacity under non-uniform traffic is proposed. Unlike most of previous works, which have used the Poisson model for the burst traffic, this paper considers a non-uniform ON–OFF model being a more realistic choice for OBS networks. Compared with simulation, the method was proved to be very accurate and much faster (up to four orders of magnitude faster than simulation in the studied cases). These features make the proposed method very useful for network analysis, especially for large size networks where simulation time can be prohibitively high.
12 citations
TL;DR: This work proposes placement algorithms that are based on network topology and the relative importance of a node in routing multicast sessions, which is measured by the proposed metrics.
Abstract: Most existing algorithms for the problem of optical signal splitter placement or multicast splitting-capable node placement in a WDM network are based on the performance of attempting a large set of randomly generated multicast sessions in the network. Experiments show that placement of multicast capable nodes based on their importance for routing one set of multicast sessions may not be a right choice for another set of multicast sessions. In this work, we propose placement algorithms that are based on network topology and the relative importance of a node in routing multicast sessions, which is measured by our proposed metrics. Since a network topology is fixed once given, the proposed algorithms are essentially network traffic independent. We evaluate the proposed placement algorithms given static sets of multicast sessions as well as under dynamic traffic conditions, which are routed using our splitter constrained multicast routing algorithm. Our results show that the proposed algorithms perform better, compared to existing algorithms.
12 citations
TL;DR: Three novel algorithms for solving the “Impairment Aware Optimal Path Pair” problem are proposed, and their performance is evaluated through extensive simulations.
Abstract: In wavelength-routed optical networks, a lightpath connection can be provisioned only if a path, or a pair of paths in case of path protection is required, can be found which satisfies multiple constraints while simultaneously achieving optimal primary cost. The primary cost can be any metric set by network administrators, and the constraints concerned in optical networks include wavelength continuity constraint, accumulation effect of some transmission impairments in the optical domain, and SRLG-disjoint requirement in survivable networks. In this paper, the impact of these constraints on the optimal path calculation algorithms is studied. Three novel algorithms for solving this problem, which we call “Impairment Aware Optimal Path Pair” problem, are proposed, and their performance is evaluated through extensive simulations.
12 citations
TL;DR: A novel graph model is proposed for describing the states of the multi-granularity switching WDM networks and it is proved that its algorithms yield better performance than those deal with the routing and wavelength assignment separately.
Abstract: For the purpose of reducing the complexity and cost of optical large-scale cross-connect, wavelengths are grouped into wavebands or fiber to be switched as a single entity, which is called multi- granularity switching However, it introduces more complexity into the routing and wavelength assignment problem In this paper, we propose a novel graph model for describing the states of the multi-granularity switching WDM networks Based on the model, the dynamic routing and wavelength assignment problems for multi-granularity traffic can be solved jointly, and different on-line wavelength grooming policies can be achieved simultaneously By simulation, we compared the performance of our algorithms under different policy and different percent of fibers for fiber switching The result proved that our algorithms yield better performance than those deal with the routing and wavelength assignment separately
11 citations
TL;DR: A novel Lagrangean relaxation (LGR) algorithm is proposed to perform constraint relaxation and derive a set of subproblems to solve the tradeoff between traffic grooming and IP QoS routing in IP over WDM networks.
Abstract: In this article, we consider traffic grooming and integrated routing in IP over WDM networks. The challenges of this problem come from jointly considering traffic grooming, IP routing, and lightpath routing and wavelength assignment (RWA). Due to the high bandwidth of optical fiber, there exists a mismatch between the capacity needed by an IP flow and that provided by a single lightpath. Traffic grooming is therefore used to increase the network utilization by aggregating multiple IP flows in a single lightpath. However, traffic grooming incurs additional delays that might violate Quality-of-Service (QoS) requirements of IP users. In this work, the tradeoff between traffic grooming and IP QoS routing is well-formulated as a mixed integer and linear optimization problem, in which the revenue from successfully provisioning IP paths is to be maximized. Problem constraints include IP QoS, routing, optical RWA, and the WDM network capacity. We propose a novel Lagrangean relaxation (LGR) algorithm to perform constraint relaxation and derive a set of subproblems. The Lagrangean multipliers are used in the proposed algorithm to obtain a solution in consideration of grooming advantage and resource constraints simultaneously. Through numerical experiments and comparisons between the proposed algorithm and a two-phase approach, LGR outperforms the two-phase approach under all experimental cases. In particular, the improvement ratio becomes even more significant when the ratio of IP flow to the wavelength capacity is smaller.
11 citations
TL;DR: A feedback-based OBS network architecture in which core nodes send messages to source nodes requesting them to reduce their transmission rate on congested links is described, with a new congestion control mechanism called congestion control with explicit reduction request (CCERQ).
Abstract: In this article we describe a feedback-based OBS network architecture in which core nodes send messages to source nodes requesting them to reduce their transmission rate on congested links. Within this framework, we introduce a new congestion control mechanism called congestion control with explicit reduction request (CCERQ). Through feedback signals, CCERQ proactively attempts to prevent the network from entering the congestion state. Basic building blocks and performance tradeoffs of CCERQ are the main focus of this article.
10 citations
TL;DR: An efficient sparse splitting constrained multicast routing algorithm called Multicast Capable Node First Heuristic (MCNFH) is proposed and results and analysis show that MCNFH builds multicast trees with the least wavelength channel cost and with the smallest number of wavelengths used per link.
Abstract: The advances in wavelength-division multiplexing (WDM) technology are expected to facilitate bandwidth-intensive multicast applications through light splitting. Due to complexity and cost constraints, light splitting (or optical multicast) nodes are sparsely configured in a practical WDM network. In this article, we investigate the multicast routing problem under the sparse light-splitting constraint. An efficient sparse splitting constrained multicast routing algorithm called Multicast Capable Node First Heuristic (MCNFH) is proposed. The key idea of MCNFH is to include the shortest path, that includes most of the multicast capable nodes, for configuring the multicast tree. Simulations and comparisons are used to demonstrate the performance of MCNFH. Simulation results and analysis show that MCNFH builds multicast trees with the least wavelength channel cost and with the smallest number of wavelengths used per link. In addition, MCNFH requires only one transmitter at the source node.
TL;DR: A novel asymmetric traffic model is presented, suitable for studying the effects of non-uniform traffic in slotted buffer-less OPS networks and to act as a tutorial for teletraffic engineering in such networks.
Abstract: Optical packet switching (OPS) has emerged as a promising architecture for the future all-optical network scenario. In order to have a successful deployment of OPS networks, several networking issues must be resolved, e.g., how to resolve contentions and how to provide Quality-of-Service (QoS) differentiation. Such networking studies often rely on teletraffic analysis in order to quantify the performance of the OPS network. In this paper, we introduce new and review existing traffic models for slotted buffer-less OPS networks. In particular, we present a novel asymmetric traffic model, suitable for studying the effects of non-uniform traffic. Optical packet switches with and without wavelength conversion are studied. Numerical evaluations and a comparison study of the presented traffic models are reported. The main contributions of this paper are to advance the field of establishing a theoretical framework for slotted OPS networks and to act as a tutorial for teletraffic engineering in such networks.
TL;DR: A novel contention avoidance technique is presented by using the parallel link server (PLS) architecture, which may overcome the lack of information at the edge node and the absence of global coordination among nodes to obtain burst blocking and data dropping probabilities.
Abstract: Optical burst switching (OBS) is thought to be the best way to adapt bursty traffic of IP-based next generation network services. However, there are a lot of challenges to make OBS networking a reality. Of most concern is burst contention avoidance. The major contention avoidance resolutions in literature are wavelength conversion, fiber delay lines, and deflecting routing. They are very vulnerable to network load and may suffer from severe data loss in case of heavy traffic. Even at moderate traffic, contention caused by using these methods lead to burst blocking and data losses. In this article, a novel contention avoidance technique is presented by using the parallel link server (PLS) architecture, which may overcome the lack of information at the edge node and the absence of global coordination among nodes. Using Poisson and Self-Similar traffic arrival models, the proposed mechanism is compared with the traditional single link server architecture through simulation in the 14-node national science foundation network. The numerical results show that this architecture, without additional other methods, can obtain burst blocking and data dropping probabilities with almost two orders of magnitudes less than those in the single link server architecture. Gains are achieved by less than 7% increase in end-to-end delay when carrying coordinated traffic and a load under 0.45.
TL;DR: A new layered-routing approach to address the problem of all-optical multicast over wavelength-routed wavelength division multiplexing (WDM) networks and proposes two generic frameworks to facilitate heuristic development.
Abstract: We have developed a new layered-routing approach to address the problem of all-optical multicast over wavelength-routed wavelength division multiplexing (WDM) networks. We model the WDM network as a collection of wavelength layers with sparse light- splitting (LS) and wavelength conversion (WC) capabilities. We apply the degree constraint technique to solve the problem. The approach is capable of completing multicast routing and wavelength assignment (MCRWA) in one step. We propose two generic frameworks to facilitate heuristic development. Any heuristic that is derived from either Prim’s or Kruskal’s algorithm can be easily imported to solve the MCRWA problem. One example is given for each framework to demonstrate heuristic development. Extensive simulations were carried out to measure the performance of heuristics developed from the frameworks. The results show that the STRIGENT scheme is suitable for hardware design and it is advisable to deploy light splitters and wavelength converters to the same node for better performance.
TL;DR: This experiment answers the question “To what extent can a selection of trees compliment a cycle-based design, or vice-versa?” and demonstrates the intrinsic merit of cycles over trees for pre-planned protection.
Abstract: As high-speed networks grow in capacity, network protection becomes increasingly important. Recently, following interest in p-cycle protection, the related concept of p-trees has also been studied. In one line of work, a so-called “hierarchical tree” approach is studied and compared to p-cycles on some points. Some of the qualitative conclusions drawn, however, apply only to p-cycle designs consisting of a single Hamiltonian p-cycle. There are other confounding factors in the comparison between the two, such as the fact that, while the tree-based approach is not 100% restorable, p-cycles are. The tree and p-cycle networks are also designed by highly dissimilar methods. In addition, the claims regarding hierarchical trees seem to contradict earlier work, which found pre-planned trees to be significantly less capacity-efficient than p-cycles. These contradictory findings need to be resolved; a correct understanding of how these two architectures rank in terms of capacity efficiency is a basic issue of network science in this field. We therefore revisit the question in a definitive and novel way in which a unified optimal design framework compares minimum capacity, 100% restorable p-tree and p-cycle network designs. Results confirm the significantly higher capacity efficiency of p-cycles. Supporting discussion provides intuitive appreciation of why this is so, and the unified design framework contributes a further theoretical appreciation of how pre-planned trees and pre-connected cycles are related. In a novel further experiment we use the common optimal design model to study p-cycle/p-tree hybrid designs. This experiment answers the question “To what extent can a selection of trees compliment a cycle-based design, or vice-versa?” The results demonstrate the intrinsic merit of cycles over trees for pre-planned protection.
TL;DR: An average assembly delay metric is defined, which represents the assembly delay experienced by a random arrival at the burst assembler of an edge OBS node, and this metric is used to define and configure a two-class burst-assembly policy, which gives preference to high-priority traffic over low-priority packet arrivals.
Abstract: In Optical Burst-Switched (OBS) networks, the limitation of optical buffering devices make it impractical to deploy conventional delay-based differentiation algorithms such as Active Queue Management, Weighted Fair Queuing, etc. Furthermore, only the delay that appears due to the burst-assembly process constitutes a variable quantity (all the other sources of delay are mostly fixed), it is then reasonable to make use of the burst-assembly algorithm to provide class-based delay differentiation. The aim of the following study is twofold: first it defines an average assembly delay metric, which represents the assembly delay experienced by a random arrival at the burst assembler of an edge OBS node; and second, this metric is used to define and configure a two-class burst-assembly policy, which gives preference to high-priority traffic over low-priority packet arrivals. The results show that, (1) tuning the parameters of the two-class assembly algorithm, the two classes of traffic exhibit different burst-assembly delay; and, (2) such parameters can be adjusted to provide a given differentiation ratio in the light of the proportional QoS differentiation approach proposed in the literature. A detailed analysis of the two-class assembly algorithm is given, along with an exhaustive set of experiments and numerical examples that validate the equations derived.
TL;DR: The capability that core nodes can split incoming light signals to support multicast to achieve more efficient burst grooming is explored and two effective burst grooming algorithms are proposed, including a no over-routing waste approach (NoORW); and a minimum relative total resource ratio approach (MinRTRR).
Abstract: Optical burst switching (OBS) is a switching concept which lies between optical circuit switching and optical packet switching. Both node switching time and burst size can impact the resource efficiency of an OBS network. To increase resource utilization, burst grooming has been proposed where numerous data bursts are coalesced to form a larger burst that will be switched as one unit in order to reduce the resource waste and switching penalty. In this article, assuming burst grooming can only be realized at edge nodes, we study the burst grooming problem where sub-bursts originating from the same source may be groomed together regardless of their destinations under certain conditions. We explore the capability that core nodes can split incoming light signals to support multicast to achieve more efficient burst grooming. Specifically, core nodes can transmit the groomed burst to multiple downstream nodes if the sub-bursts in the groomed burst have different destinations. The groomed burst will traverse a tree which spans the source and all the destinations of the sub-bursts in the groomed burst. The destination egress nodes recognize, de-burstify, and drop the sub-bursts destined to these nodes, i.e., the sub-bursts destined to these egress nodes are removed from the groomed burst. At the same time, the remaining sub-bursts may be groomed with sub-bursts at these egress nodes subject to burst grooming criteria. We propose two effective burst grooming algorithms, (1) a no over-routing waste approach (NoORW); and (2) a minimum relative total resource ratio approach (MinRTRR). Our simulation results have shown that the proposed algorithms are effective in terms of the burst blocking probability, the average burst end-to-end delay, the number of sub-bursts per groomed burst, as well as the resource waste.
TL;DR: This work forms a unified framework for studying QoS in OBS networks with a GMPLS-based control plane and addresses the problem of planning a whole OBS network given a series of QoS constraints for each one of the Forwarding Equivalency Classes in the network.
Abstract: The design of OBS networks that guarantee QoS provisioning for different classes of traffic is a major topic under current research. In this work we formulate a unified framework for studying QoS in OBS networks with a GMPLS-based control plane. We use this framework in order to investigate two problems. First, the configuration of the parameters of an aggregation strategy so that a given Forwarding Equivalency Class observes its corresponding QoS requirements. Second, we address the problem of planning a whole OBS network given a series of QoS constraints for each one of the Forwarding Equivalency Classes in the network. The presented QoS framework constitutes a valuable tool for studying QoS-related issues in OBS networks.
TL;DR: A preemption-based scheme for the concurrent improvement of network throughput and burst fairness in optical burst switched networks and is enhanced by combining it with a fairness-improving scheme previously proposed in the literature.
Abstract: In this article, we propose and evaluate a preemption-based scheme for the concurrent improvement of network throughput and burst fairness in optical burst switched networks. The scheme uses a preemption weight consisting of two terms: the first term is a function of the size of the burst and the second term is the product of the hop count times the length of the lightpath of the burst. The second term is adjusted by a minimum function to prevent the problem of reverse unfairness. Extensive simulation tests showed that the application of the scheme results in higher throughput and at the same time improves the fairness coefficient. The scheme is further enhanced by combining it with a fairness-improving scheme previously proposed in the literature. Detailed performance tests are presented and analyzed.
TL;DR: An enhancement to the current GMPLS RSVP-TE signaling protocol with offset time-based provisioning that minimizes the lightpath setup, improving the overall network performance in terms of blocking probability and setup delay is proposed.
Abstract: In wavelength-routed optical networks, the high-delay introduced by the optical switching fabric for resource reservation increases critically the lightpath setup delay. In order to minimize the setup delay, Generalized Multi-protocol Label Switching (GMPLS) introduced the concept of Suggested Label Object (SL), which allows to start reserving and configuring the hardware with a proposed wavelength from the source node to the destination node. This solution is not optimal in wavelength selective networks (WSN) (i.e., without wavelengths converters). The need of guaranteeing the wavelength continuity constraint for end-to-end optical connections, combined with the lack of global wavelength-based link information (the source node is not aware of which wavelengths are available on each link), makes that the likelihood of establishing a lightpath using the proposed suggested label may be minimum. In this article, we propose an enhancement to the current GMPLS RSVP-TE signaling protocol with offset time-based provisioning that minimizes the lightpath setup, improving the overall network performance in terms of blocking probability and setup delay. Experimental performance evaluation has been carried out in ADRENALINE testbed, a GMPLS-based intelligent all-optical transport network.
TL;DR: The complete network study shows that ORION can absorb temporal traffic overloads, as intended, provided sufficient buffering is present and the restriction of simultaneous packet insertions/extractions do not deteriorate performance and thus the use of traffic concentrators assure ORION’s economic viability.
Abstract: In this article, we present a detailed performance evaluation of a hybrid optical switching (HOS) architecture called Overspill Routing in Optical Networks (ORION) The ORION architecture combines (optical) wavelength and (electronic) packet switching, so as to obtain the individual advantages of both switching paradigms In particular, ORION exploits the possible idle periods of established lightpaths to transmit packets destined to the next common node, or even directly to their common end-destination Depending on whether all lightpaths are allowed to simultaneously carry and terminate overspill traffic or overspill is restricted to a sub-set of wavelengths, the architecture limits itself to constrained or un-constrained ORION To evaluate both cases, we developed an extensive network simulator where the basic features of the ORION architecture were modeled, including suitable edge/core node switches and load-varying sources to simulate overloading traffic conditions Further, we have assessed various aspects of the ORION architecture including two basic routing/forwarding policies and various buffering schemes The complete network study shows that ORION can absorb temporal traffic overloads, as intended, provided sufficient buffering is present We also demonstrate that the restriction of simultaneous packet insertions/extractions, to reduce the necessary interfaces, do not deteriorate performance and thus the use of traffic concentrators assure ORION’s economic viability
TL;DR: This article proves the correctness of segment protection used for avoiding traps, and investigates the effect of the number of segments on the blocking performance in several typical real networks, finding that, in most cases, it is sufficient to use two backup segments to protect a given active path.
Abstract: Segment protection has been recognized as an efficient way to avoid “traps” in survivable WDM mesh network. In this article, we present some insights on this kind of protection. We first prove the correctness of segment protection used for avoiding traps. We then investigate the effect of the number of segments on the blocking performance in several typical real networks (includes middle and large networks). We find that, in most cases, it is sufficient to use two backup segments to protect a given active path. Based on this observation, we then propose a novel and efficient segment protection algorithm, called two-segment (TS), which has a lower-computational complexity and comparative performance than existing algorithms.
TL;DR: This article proposes two novel ILP formulations, which drastically reduce the number of integer variables compared to existing ILPs, which leads to much more efficient formulations and presents a simple heuristic that may be used for larger networks for which ILP formulation become computationally intractable.
Abstract: Path protection in WDM networks is one of the popular ways to design resilient WDM networks. Although complete ILP formulations for optimal design of WDM networks have been proposed in literature, the computational cost of actually solving such formulations make this approach impractical, even for moderate sized networks. This high computational cost arises mainly due to the large number of integer variables in the formulations, which increases the complexity exponentially. As a result, most practical solutions use heuristics, which do not provide any guarantees on the performance. In this article, we propose two novel ILP formulations, which drastically reduce the number of integer variables compared to existing ILPs. This leads to much more efficient formulations. We also present a simple heuristic that may be used for larger networks for which ILP formulations become computationally intractable.
TL;DR: This article evaluated the performance of several different multi-board switches, based on better understanding the effects of different connection schemes between switching boards and optical buffers and investigating possible schemes for achieving comparable performance to that of the ideal, strictly non-blocking switches.
Abstract: With the development of optical packet-switching (OPS) technologies, multi-fiber OPS networks will play an important role in the future data transmissions. In such networks, instead of constructing some extremely expensive node configurations with strictly non-blocking switching function, a more practical solution is multi-board switches that contain a number of small-sized switching boards. In this article, we have evaluated the performance of several different multi-board switches, based on the following two main objectives: (i) better understanding the effects of different connection schemes between switching boards and optical buffers and (ii) investigating possible schemes for achieving comparable performance to that of the ideal, strictly non-blocking switches. Extensive simulation results have shown that unlike circuit-switched net- works, multi-board OPS cannot easily perform comparably to the strictly non-blocking switch by having slightly more fibers per link. Also, such a problem can be tackled by several different approaches. The most efficient one is to equip the switch with more buffers rather than to increase the switching-board size or to enhance the buffer sharing between different switching boards.
TL;DR: A new random access protocol for OC-DMA networks is proposed, employing a new mathematical model, namely the processor-sharing system, that outperforms other existing protocols and is truly applicable to model OCDMA networks.
Abstract: A New random access protocol for OC-DMA networks is proposed in this paper. Employing a new mathematical model, namely the processor-sharing system, steady-state throughput ratio and average delay of OCDMA networks are calculated. The results reveal that our protocol outperforms other existing protocols. Meanwhile, we investigate the performance of OCDMA networks by altering the code parameters, i.e., length and weight, and the maximum number of active users in the system, corresponding results are indeed consistent with the practical situation. In addition, the analysis is simplified compared with the conventional Markov chain model. Thus the processor-sharing system is truly applicable to model OCDMA networks.
TL;DR: A dynamic load-aware congestion control scheme that operates based on the highest of the loads of all links over the path between each pair of ingress and egress nodes in an OBS network, and an algorithm that dynamically determines a load threshold for adjusting burst sending rate, according to the traffic load in a network.
Abstract: The most important design goal in Optical Burst Switching (OBS) networks is to reduce burst loss resulting from resource contention. Especially, the higher the congestion degree in the network is, the higher the burst loss rate becomes. The burst loss performance can be improved by employing an appropriate congestion control. In this paper, to actively avoid contentions, we propose a dynamic load-aware congestion control scheme that operates based on the highest (called ‘peak load’) of the loads of all links over the path between each pair of ingress and egress nodes in an OBS network. We also propose an algorithm that dynamically determines a load threshold for adjusting burst sending rate, according to the traffic load in a network. Further, a simple signalling method is developed for our proposed congestion control scheme. The proposed scheme aims to (1) reduce the burst loss rate in OBS networks and (2) maintain reasonable throughput and fairness. Simulation results show that the proposed scheme reduces the burst loss rate significantly, compared to existing OBS protocols (with and without congestion control), while maintaining reasonable throughput and fairness. Simulation results also show that our scheme keeps signalling overhead due to congestion control at a low level.
TL;DR: The use of wavelength division multiplexing, in combination with clockwork routing, is proposed to provide significant improvements in network scalability and bisection bandwidth, resulting in higher throughput and reduced latency.
Abstract: Optical networks based on the ‘clockwork’ routing mechanism have significant advantages for high-speed optical packet switched applications in which latency is critical, such as parallel and distributed computing. The main advantages are that no optical buffers and no centralized contention arbitration are required. However, the network performance decreases rapidly with increasing network size. This article proposes the use of wavelength division multiplexing, in combination with clockwork routing, to provide significant improvements in network scalability and bisection bandwidth, resulting in higher throughput and reduced latency.
TL;DR: A high-performance Efficient Message Prioritization and Scheduling (EMPS) protocol, for intelligent message scheduling in Wavelength-Division Multiplexing (WDM) star networks is introduced and extensive simulations show that the proposed protocol always brings about a significant performance improvement.
Abstract: A high-performance Efficient Message Prioritization and Scheduling (EMPS) protocol, for intelligent message scheduling in Wavelength-Division Multiplexing (WDM) star networks is introduced. The performance of the well-known EATS and MSL schemes is noticeably degraded in practical networks with non-uniform destinations and non-negligible transceiver tuning latencies. Under these realistic conditions, it is common that two or more messages with the same destination have to be scheduled consecutively or at close times. In most cases, this brings about some performance penalty, owing to the delayed availability of the destination’s receiver for the second (and beyond) of the consecutive messages. As the frequency of such occurrences increases, the performance degradation of the existing schemes becomes more prominent. EMPS is proposed to deal with this problem. It simultaneously considers multiple messages from different transmitting nodes and gives priority to messages intended for the least used destinations each time. By balancing the offered load in this way, EMPS minimizes the probability of having to schedule two or more messages with the same destination consecutively or at close times. Additionally, by incorporating the Minimum Scheduling Latency algorithm for channel selection, the protocol also minimizes the actual performance penalty incurred, when scheduling of consecutive messages with the same destination cannot be avoided. Extensive simulations are carried out in order to study the performance of EMPS and compare it to other efficient schemes under various conditions. The simulation results show that the proposed protocol always brings about a significant performance improvement.
TL;DR: Network performance with respect to connection unavailability, traffic loss, spare capacity consumption, and dual failure restorability is investigated in a case study for an optical span-restorable long-haul network.
Abstract: Dual-span failures are the key factor of the system unavailability in a mesh-restorable network with full restorability of single-span failures. Availability analysis based on reliability block diagrams is not suitable to describe failures of mesh-restorable networks with widely distributed and interdependent spare capacities. Therefore, a new concept of restoration-aware connection availability is proposed to facilitate the analysis. Specific models of span-oriented schemes are built and analyzed. By using the proposed computation method and presuming dual-span failures to be the only failure mode, we can exactly calculate the average connection unavailability with an arbitrary allocation rule for spare capacity and no knowledge of any restoration details, or the unavailability of a specific connection with known restoration details. Network performance with respect to connection unavailability, traffic loss, spare capacity consumption, and dual failure restorability is investigated in a case study for an optical span-restorable long-haul network.
TL;DR: From numerical analysis, the optimal number of wavelengths in a link, which provides the lowest blocking probability, is obtained to suggest a guideline for node dimensioning.
Abstract: Network dimensioning should be progressed for pursuing the ultimate efficiency of network system resources in order to satisfy target performance. This article studies node dimensioning as a method of resource optimization in optical burst switching (OBS) networks. OBS is a new switching technology for pursuing bufferless transparent optical networks by sending control packets prior to data burst in order to provision resources for the burst. However, the basic assumption of a bufferless node implies burst contention at a core node when more than two bursts attempt to move forward the same output simultaneously. Thus, burst contention is a critical performance metric and this article takes it into account as a constraint on node dimensioning and target performance. In this article, we first present node dimensioning issues for OBS networks. Two constraints from the transport plane and the control plane which affect burst contention are then introduced. The effect of the burst assembly process on node dimensioning is also presented. From numerical analysis, the optimal number of wavelengths in a link, which provides the lowest blocking probability, is obtained to suggest a guideline for node dimensioning.