Search

We study reconfigurations of wavelength-routed Wavelength Division Multiplexing (WDM) networks in response to lightpath demand changes, with the objective of servicing more lightpath demands without additional network resources from a long-term network operation point of view. For the reconfiguration problem under study, we assume WDM network operators are provided with lightpath demands in batches. With limited network resources, our problem has two unique challenges: balancing network resource allocations between current and future lightpath demands, and modeling future lightpath demands. The first challenge implies making tradeoffs between accepting as many current immediate lightpath demands as possible and reserving a certain amount of network resources for near future predicted lightpath demands. The second challenge implies modeling future predicted lightpath demands, which are not exactly known or certain as the current lightpath demands. Our proposed model allows a natural separation between the operation of the optical layer and the user traffic layer (predominantly the IP-layer), while supporting their interactions, for which we propose a new formulation for per-link congestion control, associated with a mathematical solution procedure. Our simulation results reveal that by properly controlling resource allocations in the current session using our proposed mechanism, rejections in future sessions are greatly reduced.

Several overlay-based live multimedia streaming platforms have been proposed in the recent peer-to-peer streaming literature. In most of the cases, the overlay neighbors are chosen randomly for robustness of the overlay. However, this causes nodes that are distant in terms of proximity in the underlying physical network to become neighbors, and thus data travels unnecessary distances before reaching the destination. For efficiency of bulk data transmission like multimedia streaming, the overlay neighborhood should resemble the proximity in the underlying network. In this paper, we exploit the proximity and redundancy properties of a recently proposed clique-based clustered overlay network, named eQuus, to build efficient as well as robust transport overlays for multimedia streaming. To combine the efficiency of content pushing over tree structured overlays and the robustness of data-driven mesh overlays, higher capacity stable nodes are organized in tree structure to carry the long haul traffic and less stable nodes with intermittent presence are organized in localized meshes. The overlay construction and fault-recovery procedures are explained in details. Simulation study demonstrates the good locality properties of the platform. The outage time and control overhead induced by the failure recovery mechanism are minimal as demonstrated by the analysis.

This paper presents a study on the Grade-of-Service (GoS) differentiation of static resource allocation in lightpath routed WDM networks, where lightpath requests between node pairs are given. Each request is associated with a service grade. The goal is to maintain certain service levels for the requests of all grades. The service levels are measured in terms of their acceptance ratios. We solve this network optimization problem by adopting a penalty-based framework, in which network design and operation goals can be evaluated based on cost/revenue. We propose a static GoS differentiation model as one minimizing the total rejection and cost penalty, in which the rejection penalty reflects the revenue of accepting a request, and the cost penalty reflects the resource consumption of providing a lightpath to a request. Then, a solution based on the Lagrangian relaxation and subgradient methods is used to solve the proposed optimization problem. Three different application scenarios are presented: static GoS differentiation of requests between the same node pair, static GoS differentiation of requests between different node pairs, and an integration of static GoS differentiation into the network profit objective. The fairness issues and the impact of relative penalty factors are discussed to provide guidelines for network planning.

Ubiquitous computing environment is defined by the shift of computing technology from the desktop to the background. One of its most notable attributes is its potential to extend the scope of service and personal mobility. This paper describes an agent-based architecture that brings personal and service mobility to the ubiquitous computing environment. A software agent, running on a portable device carried by the user, leverages the existing service discovery protocols to learn about all services available in the vicinity of the user. Short-range wireless technology such as Bluetooth can be used to build a personal area network connecting only devices that are close enough to the user. Acting on behalf of the user and based on a number of aspects, the software agent runs a quality of service (QoS) negotiation and selection algorithm to select the most appropriate available service(s) to be used for a given communication session. The software agent selects as well the configuration parameters for each service. The proposed architecture supports also service hand-off to recompense for service volatility during user movement. Copyright © 2004 John Wiley & Sons, Ltd.

Crawling Rich Internet Applications (RIAs) is important to ensure their security, accessibility and to index them for searching. To crawl a RIA, the crawler has to reach every application state and execute every application event. On a large RIA, this operation takes a long time. Previously published GDist-RIA Crawler proposes a distributed architecture to parallelize the task of crawling RIAs, and run the crawl over multiple computers to reduce time. In GDist-RIA Crawler, a centralized unit calculates the next task to execute, and tasks are dispatched to worker nodes for execution. This architecture is not scalable due to the centralized unit which is bound to become a bottleneck as the number of nodes increases. This paper extends GDist-RIA Crawler and proposes a fully peer-to-peer and scalable architecture to crawl RIAs, called PDist-RIA Crawler. PDist-RIA doesn’t have the same limitations in terms scalability while matching the performance of GDist-RIA. We describe a prototype showing the scalability and performance of the proposed solution.

Most work related to quality of service (QoS) is concerned with individual system components, such as the operating system or the network. However, to support distributed multimedia applications, the entire distributed system must participate in providing the guaranteed performance levels. In recognition of this, a number of QoS architectures have been proposed to provide QoS guarantees. The mechanisms and schemes proposed by those architectures are used in a rather static manner since the involved entities, e.g., the network, sender and receiver, are known before the connection (call) set-up phase. In contrast to these architectures, we propose a general QoS management framework which supports the dynamic choice of a configuration of system components to support the QoS requirements for the user of a specific application. We consider different possible system configurations and select the most appropriate one depending on the desired QoS and the available resources. In this paper we present an overview of this general frameworks especially, we concentrate on QoS negotiation and adaptation mechanisms. To show the feasibility of this approach, we designed and implemented a QoS manager for distributed multimedia presentational applications, such as news-on-demand. The negotiation and adaptation mechanisms which are supported by the QoS manager are specializations of the general framework. The proposed framework allows to improve the utilization of system resources, and thus to increase the system availability; it also allows to recover automatically, if this is possible, from QoS degradations. Furthermore, it provides the flexibility to incorporate different resource reservation schemes and scheduling policies, and to accommodate new system component technologies.

High-speed networks and powerful end-systems enable new types of applications, such as video-on-demand and teleconferencing. Such applications are very demanding on quality of service (QoS) because of the isochronous nature of the media they are using. To support these applications, QoS guarantees are required. However, even with service guarantees, violations may occur because of resources shortage, e.g., network congestion. In this paper we propose new adaptation approaches, which allow the system to recover automatically, if possible, from QoS violations (1) by identifying a new configuration of system components that might support the initially agreed QoS and by performing a user-transparent transition from the original configuration to the new one, (2) by redistributing the levels of QoS that should be supported, in the future, by the components, or (3) by redistributing the levels of QoS that should be supported immediately to meet end-to-end requirements based on the principle that (local) QoS violation at one component may be recovered immediately by the other components participating in the support of the requested service. The proposed approaches, together with suitable negotiation mechanisms, allow us (1) to reduce the probability of QoS violations which may be noticed by the user, and thus, to increase the user confidence in the service provider, and (2) to improve the utilization of the system resources, and thus to increase the system availability.

Distributed multimedia (MM) applications such as video-on-demand and teleconferencing provide services with different quality of service (QoS) requirements. Hence, the user should be able to negotiate the desired QoS depending on his/her needs, the end-system characteristics and his/her financial capacity. In response to a service request with the desired QoS, most QoS negotiation approaches return an acceptance or a simple rejection of the request. More specifically, they provide the user only with the QoS that can be supported at the time the request is made and assume that the service is requested for indefinite duration. This paper describes work on a new QoS negotiation approach with future reservations (NAFUR) that decouples the starting time of the service from the time the service request is made and requires that the duration of the requested service must be specified. NAFUR allows to compute the QoS that can be supported for the time the service request is made, and at certain later times carefully chosen. As an example, if the requested QoS cannot be supported for the time the service request is made, the proposed approach allows to compute the earliest time, when the user can start the service with the desired QoS. NAFUR will help to increase (a) the flexibility of the system by providing the user with more choices, and (b) the system resource utilization, and the availability of the system, by encouraging the sharing of the resources, e.g. multicast for video-on-demand systems. Furthermore, it provides the flexibility to incorporate (a) a range of resource reservation schemes and scheduling policies, and (b) a range of new system component technologies.

Distributed multimedia applications require a variety of levels of quality of service (QoS) from communication networks and end-systems which realize the multimedia interface with the human users or provide remotely accessed multimedia information. The management of these services has to take into account the available resources and the user's wishes concerning the desired quality and costs. Based on our experience with the construction of a News-on-Demand prototype, we present in this paper a few simple principles applying to QoS management for distributed multimedia applications. We discuss in particular ways in which an application can adapt to reduced network performance related to throughput, loss, delay or jitter, and we consider the situation where the configuration of the distributed application can be selected dynamically and be revised during the running of the application. We also comment on the handling of QoS in a layered system architecture and the use of performance models for QoS management.

As mobile information networks are expanding rapidly, we expect to integrate voice, paging, electronic mail and other wireless information services. Interworking units that perform protocol conversion at the boundaries of different networks will play an important role. In this paper, we propose an efficient algorithm for constructing optimized protocol converters to achieve interoperability between heterogeneous data networks. This algorithm first derives constraints from two given protocols, and apply the constraints to channel specifications, thus removing message sequences that do not contribute to system progress. Then, an optimized converter is generated from a given service specification, the two protocol specifications and the modified channel specifications. A reduction relation is used to compare the service specification and the constructed internetworking system in order to deal with the problem of nondeterministic services. Compared with related works, our method has two advantages: (1) it generates an optimized converter; (2) it can be applied to the case that the service specification is nondeterministic. The application of the method to mobile networks is given by an example.

Part 5: P2P and Aggregation Schemes; International audience; This paper presents a diffusive load balancing algorithm for peer-to-peer systems. The algorithm reduces the differences of the available capacities of the nodes in the system using service migrations between nodes in order to obtain similar performance for all nodes. We propose algorithms for handling homogeneous services, i.e., services with equal resource requirements, and for heterogeneous services, i.e., services with diverse resource requirements. We have investigated the effect of load balancing in a simulated peer-to-peer system with a skip-list overlay network. Our simulation results indicate that in case that the churn (nodes joining or leaving) is negligible, a system that hosts services with small resource requirements can maintain equal performance for all nodes with a small variance. In case that churn is high, a system that hosts homogeneous services with large resource requirements can maintain equal node performance within a reasonable variance requiring only few service migrations.

This paper presents a design for a general locality preserving routing overlay network based on geographic coordinates of nodes. For efficient use of network resources, it is important to follow some locality principles while routing bulk multimedia content. The proposed overlay network creates the interconnection among large number of highly dynamic nodes in a completely decentralized manner, based on adaptive hierarchical partitioning of the geographical space. A major benefit of using coarse grain geographic coordinates is that they are available at off-the-shelf databases and thus do not incur additional measurement overhead. The performance of the proposed geography based routing overlay, in terms of two locality properties - stretch and route-convergence, is evaluated by simulation studies. Real topology data of 65 ISPs with presence in 534 cities across the world, collected by the RocketFuel project, is used as the basis of the simulated network. The performance results are compared with corresponding results obtained from a simulated Pastry overlay constructed on the same underlying network. The performance results show that proposed overlay has significantly better stretch and route-convergence characteristics compared to a regular Pastry network. The locality properties of the proposed routing overlay are almost equivalent to those of an optimal Pastry network where all nodes use the most proximal neighbor for each routing table row. Thus the proposed geographical routing network provides good locality properties without the overhead of proximity neighbor selection. Moreover, the proposed scheme facilitates certain geographical search applications that are difficult to realize in a Pastry overlay.

The advantages and issues in deploying a fast photonic network on top of a reconfigurable WDM network are discussed. The agile photonic network is deployed as another user of the reconfigurable optical WDM network (RON), with the reconfigurable optical switches setting up the optical circuits that define the virtual topology for the agile network. The services provided by the agile network are then carried over the wavelengths that are assigned to it by the global control plane of the RON. Such deployment would allow the agile network to provide the fast optical time division multiplexing (OTDM) scheduling techniques warranted for fast-changing, low-capacity traffic flows typical of metropolitan and access networks; while sustained, high-capacity flows would remain in whole lightpaths provided at the RON level to other users. Connectivity options are described for edge and core nodes, as well as the functionality requirements of the global control plane that would manage such a deployment.

We present a study on the static resource allocation in lightpath routed WDM networks, where each request is associated with a service grade. The goal is to maintain certain acceptance ratios for the requests of all grades, as well as to minimize the resource consumption. We propose a model of static Grade-of-Service (GoS) differentiation as minimizing the total rejection and cost penalty. Then, we use the Lagrangian relaxation and subgradient methods to solve the problem. The results of using static GoS differentiation are presented.

In this paper, we propose a distributed approach for the least constraining slot allocation scheme in all-optical TDM networks (LC) that was introduced in a previous work. The driving force behind our proposal is the employment of the LC scheme in a GMPLS context. After describing the basic data model and messaging parameters, we focus on defining an efficient LC resource status update scheme, which is essential to achieve compatibility with GMPLS' periodic link state update standards. Basically, we reduce the rate of resource status updates from once per call to once per few calls, and measure the impact on network performance.

Node clustering is an effective solution for achieving good performance and high reliability for peer-to-peer (P2P) systems. To improve the performance of a clustered P2P system, it is important to balance the service load among the clusters in the system. In this paper, we describe a diffusive load balancing scheme for clustered P2P systems, which dynamically adjusts the size of the clusters, by moving nodes among the clusters, based on their service demands and node resource capacities. Our simulations show that the proposed load balancing scheme significantly improves the performance of a P2P system in terms of balanced available capacity.

Several overlay-based live multimedia streaming platforms have been proposed in the recent peer-to-peer streaming literature. In most of the cases, the overlay neighbors are chosen randomly for robustness of the overlay. However, this causes nodes that are distant in terms of proximity in the underlying physical network to become neighbors, and thus data travels unnecessary distances before reaching the destination. For efficiency of bulk data transmission like multimedia streaming, the overlay neighborhood should resemble the proximity in the underlying network. In this paper, we exploit the proximity and redundancy properties of a recently proposed clique-based clustered overlay network, named eQuus, to build efficient as well as robust overlays for multimedia stream dissemination. To combine the efficiency of content pushing over tree structured overlays and the robustness of data-driven mesh overlays, higher capacity stable nodes are organized in tree structure to carry the long haul traffic and less stable nodes with intermittent presence are organized in localized meshes. The overlay construction and fault-recovery procedures are explained in details. Simulation study demonstrates the good locality properties of the platform. The outage time and control overhead induced by the failure recovery mechanism are minimal as demonstrated by the analysis., 10 pages

We propose a novel Internet Exchange (IX) architecture, namely OSN-IX, which adopts an overlaid-star network (OSN) as an IX. OSN can be considered as a "distributed" switch, which combines the advantages of the network and switch. Compared to other IX architectures, e.g., LAN-based IX, MPLS IX, Photonic IX, etc., OSN-IX has good properties of scalability, resilience, and widely distributed access points. Particularly, for the first time, OSN-IX introduces traffic engineering (TE) into the IX world. Based on the TE framework we developed for OSN-IX, OSN-IX can provide optimized dynamic inter-ISP (Internet Service Provider) routing while requiring no change, hardware or software, on existing traditional IP/MPLS routers. We have shown by simulation that our TE framework outperforms several existing inter-AS TE schemes.

The purpose of this study is to define principles that apply to the design of communication gateways between heterogeneous computer systems, and to identify strategies to solve incompatibilities in a systematic manner. The importance of communication service common properties is explored in detail. The modification of the service available for interworking due to subset selection and service concatenation is discussed. Various methods of adaptation are described. Two architectural options for the design of communication gateways are explored, namely, conversion at the service level or at the PDU (protocol data unit) level. While the former is conceptually simpler, various optimizations are possible through the latter approach. A method for deriving the specification of a protocol adapter from the two protocol specifications is given. All these issues are illustrated by a number of simple examples. >

"Agile All-Photonic Networks" (AAPN) is the theme of a Canadian research collaboration. An AAPN is a wavelengthdivision- multiplexed network that consists of several overlaid stars formed by edge nodes that aggregate traffic, interconnected by bufferless optical core nodes that perform fast switching in order to provide bandwidth allocation in sub-wavelength granularity. Specific issues addressed in this context are (a) efficient bandwidth allocation, (b) routing of MPLS flows over the AAPN, (c) allocation of protection paths, and (d) development of a demonstration prototype. This paper high-light research results and design choices related to these issues.

We study the resilience of MPLS flows over an agile all-photonic star WDM network (AAPN). On the basis of our previous inter-area optimal routing architecture, we propose and develop a dynamic inter-area shared segment-based protection (SSP) framework. We consider the dynamic protection for optimal inter-area working paths and improve the recovery time by segment-based protection. We develop a distributed partial routing information management to increase the scalability in multi-area networks while maintaining good performance compared with the case of complete information. By simulation, we show that our framework outperforms existing scheme. Furthermore, our approach shows its good potential to be a protection solution for inter-AS protection.

In this paper, we propose a timeslot allocation scheme in all-optical TDM networks with no traffic buffering. The purpose of the new scheme is to take the blocking rate to an optimal level close to what can be achieved with the use of buffers. Previous works considered the first fit and least loaded approaches to reserve timeslots in networks that include buffers and multi-fibers. Our proposed scheme applies to more basic networks that have single unidirectional fibers and no buffers.

Today's Internet suffers from the problem of heterogeneity in client devices, network connectivity, content format, and users' preferences. The framework presented in this paper tackles this problem using the approach of service composition to support distributed multimedia applications. The discussed framework for trans-coding multimedia streams uses self-organizing, resilient data distribution algorithms. The framework takes into consideration the profile of communicating devices, network connectivity, exchanged content formats, context description, and available adaptation services to find a chain of adaptation services that could be applied to the content. Part of the framework is a selection algorithm that finds the best sequence of adaptation services that can maximize the user's satisfaction with the delivered content.

Fault detection and localization is a crucial issue in all-optical networks. Since most commercially-available all-optical space switches are incapable of detecting the loss of optical signals along the data paths between its input ports and output ports, fault localization becomes a challenge for providing service survivability in such networks. This paper proposes a fault detection and localization scheme for an all-optical overlaid-star TDM network. The proposed scheme employs a fault localization technique that identifies the location of a failure by detecting the power loss of optical signals in data and control channels. Two alternatives are proposed. One requires a control channel on each wavelength of a fiber link while the other requires a small data block to be transmitted in each non-allocated data channel. Based on the proposed fault localization technique, a fault advertisement protocol is further presented, which can be incorporated into the signaling protocol used in the network to facilitate the provisioning of static protection or dynamic restoration. The data loss, fault detection time, and connection recovery time are analyzed for the different failure scenarios.

Load balancing is an effective solution to relieving network congestion and achieving good network performance. This paper investigates routing strategies for load balancing in all-optical overlaid-star TDM networks. A random routing strategy and a least-congested-path routing strategy are first presented, based on which a weighted-least-congested-path routing strategy is then proposed. The proposed strategy takes into account both load balancing and end-to-end delay in path selection, and thus can achieve better delay performance while maintaining the same blocking performance under low traffic load as compared with the other strategies. The performance of the routing strategies is evaluated through simulation results.

A semi-analytical method based on alternate projections on a linear vector space is used to construct a service matrix from a traffic matrix, where the traffic matrix represents the bandwidth requested by the edge nodes and the service matrix represents how the bandwidth will be distributed by the core of an optical star network that operates in a Time Division Multiplexing mode. The algorithm iterates over a mathematical expression of complexity O(N^2), where N denotes the number of edge nodes. The complexity of the method is therefore O(kN^2) where k denotes the number of iterations needed to converge. With N large enough one observes that k\le\leN and hence this expression tends to O(N^2). Results show that the service matrices obtained with this projection method have very high measures of similarity to the original traffic matrix, with an average similarity greater than 95% for N \geqslant 32 . The method is robust to inadmissible/bursty traffic and yields equal or improved delay performance in the optical network compared to other allocation methods.

Multi-domain resource sharing is the fundamental feature of computing and data grid applications. In this paper, we address the issue of multi-domain optical network resource sharing and present a network management system designed for end-to-end lightpath provisioning across multiple independently managed domains. Our network management system, called user-controlled lightpath provisioning (UCLP) system, is based on the grid services concept and built on the Jini and JavaSpaces technologies. The UCLP system provides a traffic engineering approach to control network infrastructure for computing and data grid applications.

Summary form only given. Shared protection aims to provide the same level of protection, against a single link failure, as the dedicated one while using less network resources. In this paper, we present the issue of survivability in a time slotted optical networks deploying DWDM. To guarantee the recovery and to maintain the performance of the service, sufficient resource needs to be available at the setup time of the protection. However it is possible to optimize the protection capacity. Indeed the primary traffic is composed of a set of flows, which may be going through different paths. Therefore a protection could be found using just enough resources by sharing the backup among many flows. We propose here a technique to identify and provision the protection using the minimum necessary resources. In particular, we present an algorithm that computes the optimal protection required for a primary traffic from a source to a destination, and the maximum capacity that could be used, on each link, for the protection. We prove through simulation results that this shared mesh protection scheme can significantly reduce the required network protection capacity.

The tremendous growth of the Internet has introduced a number of interoperability problems for distributed multimedia applications. These problems are related to the heterogeneity of client devices, network connectivity, content formats, and user's preferences. The purpose of this paper is to present a framework for transcoding multimedia streams. The proposed infrastructure takes into consideration the profile of communicating devices, network connectivity, exchanged content format, context description, and available customization services to find a chain of services that could be applied to adapt the content to the required needed format. Part of the framework is a QoS-based selection algorithm that finds the best sequence of adaptation services which can maximize users' satisfaction with the delivered content.

Shared protection aims to provide the same level of protection, against failure, as the dedicated one while using less network resources. In this paper we present the issue of survivability in a time slotted optical networks deploying DWDM. To guarantee the recovery, sufficient resource needs to be available at the setup time of the protection. However it is possible to optimize the protection capacity. Indeed the primary traffic is composed of a set of flows, which may be going through different paths. Therefore a protection could be found using just enough resources by sharing the backup among many flows. We propose here a technique to identify and provision the protection using the minimum necessary resources. We prove through simulation results that this shared mesh protection scheme can significantly reduce the required network protection capacity

Wavelength Routing (WR) and Optical Burst Switching (OBS) are two optical network techniques that have received enormous attention over the last decade. However, the two techniques are plagued with many problems. The main concern with WR is the inefficient bandwidth utilization. On the other hand, the problem with OBS is resource contention and burst dropping. In this paper, we propose a new scheme to share network resources using Time Division Multiplexing (TDM) instead of the statistical multiplexing employed in optical burst switching. To avoid contention and improve bandwidth utilization, we resort to a simple reservation scheme that guarantees timeslot deliveries. In addition, we propose the deployment of a new device that we call Sequencer, a simplified form of Optical Time Slot Interchangers (OTSIs), to assist in mapping incoming timeslots to some available outgoing ones. Our goal is to achieve a contention free network, and improve performance.

This paper describes an agent-based architecture that extends personal mobility to ubiquitous environment. A software agent, running on a portable device, leverages the existing service discovery protocols to learn about all services available in the vicinity of the user. Short-range wireless technology such as Bluetooth can be used to build a personal area network connecting only devices that are close enough to the user. Acting on behalf of the user, the software agent runs a QoS negotiation and selection algorithm to select the most appropriate available service(s) to be used for a given communication session, as well as the configuration parameters for each service, based on session requirements, the user preferences and the constraints of the devices that provide the service(s). The proposed architecture supports also service hand-off to take account of service volatility as a result of user mobility.

With the advent of smaller and ever faster computing and communications, more computing devices will be embedded in our life artifact. The explicit management of these embedded computers is definitely a burden for the regular person, especially when more than one device is required to complete a single task. This paper describes an architecture that integrates the use of agent technology, wireless communications and context-awareness to support service and personal mobility. A software agent, running on a portable device, leverages the existing service discovery protocols to learn about all services available in the vicinity of the user. The software agent runs a QoS negotiation and selection algorithm that sets the values for the session configurations parameters based on the session requirements, the user preferences, and the constraints of the devices running the services. The proposed architecture supports also service hand-off between devices to take account of service volatility in case of context change.

We present the Preemptive Bandwidth Allocation Protocol; a distributed QoS-aware signaling protocol that adapts the users’ bandwidth requirement to the limited resources of the network by preempting bandwidth of less prioritized streams. We assume that a connection will have different streams with predefined quality requirements and each stream will have a priority level assigned to it. When a new connection request comes to the network and there is a lack of bandwidth, the communication service will try to preempt some streams of existing connections to satisfy the new requests without disconnecting the basic stream of any of these existing connections. The aim is to accommodate the maximum number of users within the network with at least their minimum requirements (e.g. I frames in a MPEG-2 media data). We also present a performance evaluation that compares this distributed preemptive approach with a centralized version and the traditional non-preemptive approach.

With the growing demand of bandwidth and the development being made in the optical components technology, the IP over DWDM using optical burst switching seems to be the best solution to take advantage of the huge capacity of the fiber and accommodate high traffic of information. In this architecture the optical network is seen as an optical cloud with intelligent edges capable to interpret the IP address and store the information in electronic domains as well as checking and correcting errors. Optical burst switching [1,3] achieves better bandwidth exploitation because all the fiber wavelengths are shared among the bursts without resource pre-allocation and the whole wavelength capacity can be used by a burst. However with higher load, contention increases, and hence the number of dropped burst increases leading to a big lost of performance Several methods can be used to lower the burst-dropping probability such as wavelength conversion [5] and the use of buffers [6], but these solutions are still not there due to the high cost and the immaturity of technology. In this paper we will propose another method where the burst is segmented into several parts of equal length and in case of contention only the parts at the beginning, causing the conflict, will be discarded. We will analyze and compare this method with other practical methods: delayed burst, deflection routing. The segments of burst can be used to carry different class of service since the dropping probability of each segment depends on its position in the burst. The parts at the end have the smallest probability to be dropped. We prove trough analysis and simulation that the segmented burst improves the performance and is more suitable for traffic with several class of service.

ISO (International Standardization Organization) developed the ISO distributed test architecture for testing layered protocols. Furthermore, a general distributed test architecture where the IUT (implementation under test) contains several distributed ports is used for testing distributed systems, based on the Open Distributing Processing (ODP) Basic Reference Model (BRM). In this architecture, the testers cannot communicate or synchronize with one another unless they communicate through the IUT; and no global clock is available in the system. This architecture could model a test architecture of a communication network with n accessing nodes, where the testers reside in these nodes. When n=2, this general distributed test architecture reduces to the ISO distributed test architecture. The authors develop a test selection method with respect to this general distributed test architecture.

In this paper, we present a protocol that allocates resources in communication networks in order to assure specific QoS characteristics as requested by new connections. The design takes into consideration the possibility for the network allocation to adapt to application requirements.The proposed protocol uses a Bandwidth Preemptive Algorithm that permits adaptive bandwidth allocation in multicast, multi-stream environments. This design has been inspired by the one proposed by Sakate [1] where a centralized methodology is used. In our approach, we use a distributed methodology where we change the behavior of the communication service and allow the continuation of the service under more severe conditions. In other words, when there is a lack of bandwidth for a new connection, the communication service will try to find the missing bandwidth within the existent connections (or streams) when looking for a feasible path on a hop-by-hop basis, starting from the destination to an a on-tree node.

Most adaptive multimedia multicast applications require the source to select the number of streams to transmit as well as the QoS parameters for each stream. If the receivers have different bandwidth limits for their devices and have various preferences for the quality of the data, selecting the QoS parameters that generate the best average satisfaction for all receivers is a challenging problem. In this paper, we developed a selection algorithm that is based on the user profiles and the device capabilities. Receivers are required to send their profiles and the bandwidth limitation on their devices to the source once before the session starts. To avoid the implosion problem and have a constant running time for the selection algorithm, we partition the receivers according to the bandwidth limit of their devices into classes and use a virtual representative for each class of receivers.

In the context of distributed multimedia applications involving multicast to a large number of users, a single quality of service level may not be appropriate for all participants. It is necessary to distribute part of the QoS management process and allow each user process to make certain QoS decisions based on its local context. In order to allow for different QoS options, we assume that each source provides, for each logical multimedia stream, several different stream variants, representing different choices of user-level QoS parameters. The paper presents the design of a teleteaching system which uses this paradigm for QoS negotiation, and explains how the Delivery Multimedia Integration Framework (DMIF) of MPEG-4 can be adapted as a session protocol for such an application. In fact, it appears that this DMIF protocol, which is now being extended by ISO (DMIF Version 2) to the context of multicasting, provides some general session management functions which are quite useful for many distributed multimedia applications using broadcasting.

Gateways are introduced for interworking between several, possibly heterogeneous, distributed computer systems. A gateway has to provide for the necessary adaptation between the communication protocols used in the interconnected networks. The adaptation problem is best handled by considering the communication services of the interconnected systems. Once the problem is solved at this level, the remaining problem of conversion between the incompatible communication protocols used in the different systems can be solved automatically, as demonstrated for the case of a simple example of data transmission service from a sender to a receiver process. >

We developed a new QoS management scheme called Cooperative QoS management which is especailly suitable for multicast multimedia communication. In this paper, we discuss how this new scheme influences the design of applications based on it.

The packet level of the CCITT Recommendation X.25 describes how the virtual circuit communication service of public packet-switched data networks can be used. The recommendation contains in particular a specification of how to establish and clear virtual circuits through a given DTE-DCE interface. We comment in this paper on the following three aspects of these procedures for call establishment and clearing: (a) The recommendation contains state diagrams to clarify the operation of the DTE-DCE interface. The reader may assume that the diagrams show, at each instant in time, the unambiguous state of the interface. This is, however, not always true since the DTE and the DCE may be in different states. (b) We have analysed the operation of the interface using a finite state validation method, and have found that in some rare cases a conflicting situation (i.e. when the DTE and the DCE are in different states) may sustain through the repetitions of some cyclic behavior. (c) The procedures of X.25 are very symmetrical. However, the packet level call establishment procedure contains a non-symmetrical element for the handling of call collisions. We consider as alternatives some symmetrical procedures for this purpose. For each of these aspects we try to explain the problem and give some suggestions for its solution.

This document is a specification of the Message Link Protocol (MLP). The MLP creates a transport service between session control entities via transport connections, and some additional session control service. It is based on virtual circuits a la X.25. It is concerned with transport connection establishment and termination, and handles resets and clears to provide a reliable transport service. Furthermore it provides synchronization support related to session establishment, termination and negotiation. The allocation of transmission resources to connections is also considered. It is assumed that the flow control capabilities of the X.25 virtual circuits can be used for transport connection flow control. If, for other transmission media, flow control may not be assumed, an additional function to provide flow control at the transport service level has to be introduced.

The higher-level communication protocols in computer networks, such as belonging to the OSI Transport through Application layers, are usually considered to have an end-to-end significance, that is the entities executing the protocol reside at the two respective ends of the connection, close to the application users. This question of end-to-end significance was in particular an issue in the discussion of the meaning of acknowledgements in the Transport layer, and in the distinction between the OSI Network and Transport layers. The OSI protocols are developed for the interworking of "Open Systems". It seems that an Open System may be, in particular, a distributed computing system which presents to its environment, for example through a gateway, the protocol behavior defined by OSI. In this case the Open System would use internally, possibly over a local area network, specific protocols, which may or may not be structured according to the OSI Reference Model, but which are different to the protocols defined for OSI. In this case the OSI higher-level protocols would be executed between the entities in the environment of the Open System (e.g. system of user A1) and the gateway of the Open System, which is not end-to-end, as indicated in the figure below. The purpose of this contribution is to point out that this situation can be considered normal (i.e. end-to-end significance is irrelevant) provided that the OSI services have a property, which we call in the following "concatenation property". A service specification has the concatenation property if two service providing subsystems may be concatenated such that the system (consisting of the concatenated subsystems) provides a service which has the same logical properties as the original services. (Clearly, the performance of the concatenated service would in general be lower). It is important to note that we assume that the services of the non-OSI system (e.g. service provider B) are logically identical to those defined for OSI (e.g. service provider A). In this case the concatenation entity has a straightforward function (see below). In the case that they are different, some adaptation may be performed which has, however, an impact on all users and could sometimes result in quite complex gateway functions. The OSI Transport service has the concatenation property to a large extent. It is sufficient to use a concatenation entity that generates service requests to the service provider B for each service indication received from service provider A (with same parameter values). A problem arising, however, is one of addressing. It is necessary that the addressing space within the service provider B be foreseen by the addressing conventions used within the service provider A (and inversely) such that a user connected to the service provider A may select an address which identifies a user connected to the service provider B. It is then necessary that all service indications generated by service provider A destined to users connected to service provider B be generated over the interface to the concatenation entity, and similarly in the opposite direction. Another problem is related to error reporting. It seems that a distinction between connection aborts generated by the user or service provider, respectively, is not possible in the manner defined for OSI. It seems that the services specified for the other higher layers of OSI have similar concatenation properties. We can see two conclusions from the above discussion: (1) If services have the concatenation property the question of end-to-end significance is irrelevant. Gateways may be constructed which realize the concatenation of two or more distributed system services. (2) The service specifications of OSI are the main documents on which the construction of OSI gateways is based. These specifications are therefore essential, not only as a basis for checking the consistency of several protocol layers within the OSI hierarchy, but also as as basis for interworking between existing systems according to OSI.

As explained in section 4.2, higher-level function oriented protocols rely on an appropriate process-to-process message transport service, which is provided at the interface of the transport subsystem. Depending on the applications, different communication requirements are encountered. These requirements may be characterized (i) by the way in which a communication partner is identified and selected, and (ii) by the following points

In this work we propose an optical burst switching retransmission scheme without any loss of bursts for star topology networks that exploits the entire network capacity. The basic idea is to allow the core node to resolve contentions by itself; balancing the intelligence of the network between edge and core nodes, losing neither the advantages of optical burst switching nor the simplicity of the core switch architecture. To achieve this objective, two retransmissions schemes are defined. The former confines the average number of retransmissions per burst below 2. The later exploits the whole network capacity.