Showing papers by "Piet Demeester published in 2004"

Benchmarking and viability assessment of optical packet switching for metro networks

Abstract: Optical packet switching (OPS) has been proposed as a strong candidate for future metro networks. This paper assesses the viability of an OPS-based ring architecture as proposed within the research project DAVID (Data And Voice Integration on DWDM), funded by the European Commission through the Information Society Technologies (IST) framework. Its feasibility is discussed from a physical-layer point of view, and its limitations in size are explored. Through dimensioning studies, we show that the proposed OPS architecture is competitive with respect to alternative metropolitan area network (MAN) approaches, including synchronous digital hierarchy, resilient packet rings (RPR), and star-based Ethernet. Finally, the proposed OPS architectures are discussed from a logical performance point of view, and a high-quality scheduling algorithm to control the packet-switching operations in the rings is explained.

An OBS-based grid architecture

Abstract: Grids offer a uniform interface to a distributed collection of heterogeneous computational, storage and network resources. Most current operational grids are dedicated to a limited set of computationally and/or data intensive scientific problems. The de facto modus operandi is one where users submit job requests to a grid portal, acting as an interface to the grid's management system, which in turn negotiates with 'dumb' resources (computing and storage elements, network links) and arranges for the jobs to be executed. In this paper, we present a new grid architecture featuring generic application support, direct user access (grid-to-the-home) and decentralized scheduling intelligence in the network. We show how optical burst switching (OBS) enables these features while offering the necessary network flexibility demanded by future grid applications.

Network aware scheduling in grids

Abstract: Computational Grids consist of an aggregation of data and computing resources, which can be co-allocated to execute large data-intensive tasks which cannot be handled by an individual resource. The vast amounts of (possibly remote) data these tasks process give rise to the need for a high-capacity network interconnecting the various resources. In this paper, it is argued that, in order to obtain quality job schedules resulting in low throughput times while making efficient use of available resources, networkand service-aware algorithms need to be used. We present several heuristic algorithms, which we simulate extensively in order to gain insight in the quality of the generated schedules. In particular, in a Grid with heterogeneous Computational Resources and link capacities, we show how the average job response time can be improved by distinguishing between data-intensive and compute-intensive jobs, and scheduling these jobs based upon both Computational Resource and network load. We show that the heuristics presented perform well when compared to an Integer Linear Programming approach modelling a periodic on-line scheduler.

Scheduling Algorithms for a Slotted Packet Switch with either Fixed or Variable Length Packets

Abstract: We address the problem of congestion resolution in optical packet switching (OPS). We consider a fairly generic all-optical packet switch architecture with a feedback optical buffer constituted of fiber delay lines (FDL). Two alternatives of switching granularity are addressed for a switch operating in a slotted transfer mode: switching at the slot level (i.e., fixed length packets of a single slot) or at the burst level (variable length packets that are integer multiples of the slot length). For both cases, we show that in spite of the limited queuing resources, acceptable performance in terms of packet loss can be achieved for reasonable hardware resources with an appropriate design of the time/wavelength scheduling algorithms. Depending on the switching units (slots or bursts), an adapted scheduling algorithm needs to be deployed to exploit the bandwidth and buffer resources most efficiently.

On the Construction of QoS Enabled Overlay Networks

Abstract: The lack of QoS support in the Internet makes it difficult for service providers to give guarantees regarding the timely delivery and quality of their services. For multimedia services like video on demand and video conferencing however, the delay should be minimal. One of the main causes of the absence of QoS is the inter- and intradomain routing scheme in the Internet, minimizing the hop count instead of optimizing the QoS. In this paper we will discuss the construction of an overlay network that is able to meet the requirements of time-critical services. An optimal algorithm, for the placement of overlay servers in the Internet, will be described together with a number of heuristics. These server placement algorithms will be evaluated by comparing the resulting overlay network to the standard Internet in terms of number of connections accepted and number of overlay servers required.

Capex comparison between link-by-link and end-to-end grooming in a European backbone network

Abstract: This paper compares the traditional link-by-link traffic grooming approach to the end-to-end grooming solution where OXCs are present in all network nodes. The comparison is made based on the difference in expected capital expenditures between both approaches. We indicate that, as traffic grows, there comes a point where the savings in IP layer expenses realized by end-to-end grooming compensate the extra expenses of introducing the needed optical cross connects. This cost intersection point is determined in a realistic case, considering a European backbone network with fast growing traffic. The impact of different component cost ratios and cost evolutions on this intersection point is studied. This point is shown to be most sensitive to a changing SDH card cost. Using realistic cost figures for the considered network components, which are decreasing over time, we expect the end-to-end grooming approach to become more economical than link-by-link grooming for the studied European backbone by the end of 2005.

Evaluation of grid scheduling strategies through NSGrid: a network-aware grid simulator

Abstract: Due to the size of Computational Grids and the number and different types of resources involved, it is usually very hard to build Grid testbeds on a realistic scale, or to devise analytically tractable scheduling algorithms for distributing workloads on such a Grid. Therefore, simulation is an important tool in studying a Grid's behaviour under different management and scheduling policies. In this paper, we make the case for NSGrid, which is an NS2-based Grid Simulator capable of evaluating Grid Scheduling Strategies on different Grid topologies, providing for accurate simulation on the network packet level. Various Grid Resources have been modelled in NSGrid: Computational, Storage and Information Resources and VPN connections. These resources are managed by the Grid Scheduler, Information Service and VPN Management components and can be interconnected by any network that can be modelled in NS2. A generic job model allows the creation of different simulated Grid job loads. We show how this simulation framework can be used to study the behaviour of Grids under different scheduling algorithms, rescheduling strategies and scheduling architectures (distributed, hierarchical, centralized) and present results in this area. Specifically, we show the importance of taking into account network-related information in Grid scheduling algorithms.

Micromobility support for random access network topologies

Abstract: The combination of mobile IP with frequent handovers results in high handoff latencies and control overhead in the core network. Micromobility protocols like cellular IP. Hawaii and MIPv4 regional registration were developed to avoid these problems, supporting local mobility within one IP domain. Hereby, tree access topologies were considered to evaluate the protocol performance A. Campbell et al., (2000) (2002) although meshed topologies are desired for reasons of robustness against link failures and load balancing L. Peters et al., (2003). The mentioned micromobility protocols do not use the extra links or result in a bad performance in terms of end-to-end delay. This paper proposes a new micromobility protocol for random access topologies. During handoff a router in the access network detects if it has the function of cross-over node. Hereby, control traffic is concentrated near the involved access routers, while optimal paths in the access network are achieved. The basic concept of the protocol is presented and the handoff performance is compared to other micromobility protocols.

A heterogeneity based clustering heuristic for mobile ad hoc networks

Abstract: An ad hoc network is an autonomous system of heterogeneous, mobile nodes that communicate with each other over wireless links. Routing protocols for these networks are inherently based on broadcasting control information and are therefore very bandwidth consuming. In order to limit the amount of routing information that has to be stored and maintained by the individual nodes, the technique of clustering is used. The network is partitioned into nonoverlapping sub networks, referred to as clusters, and one cluster node, the clusterhead, will take a leading role in the dissemination of control information. In this paper we approach the problem of finding an optimal partition that explicitly takes into account the heterogeneity of the network, as an integer linear programming (ILP) problem. In a second phase we have developed a new heuristic that approximates our ILP solution that is used in our clustering algorithm. It is shown that this heuristic tends to be more stable than existing clustering techniques that are solely based on ID number and/or connectivity and that do not take into account the heterogeneity of the network.

Grid computing: the next network challenge!

Abstract: Contact details for authors: Bruno Volckaert Department of Information Technology (INTEC) Ghent University, Sint-Pietersnieuwstraat 41, B-9000 Gent (Belgium) Tel: +32 (0) 9 264 9964 Email: Bruno.Volckaert@intec.UGent.be For the past five years, grid computing has been a very hot and fruitful research theme resulting now in the deployment of the first operational grid systems. The main motivation for this new computing paradigm lies in the observation that the demand for computational and storage resources is ever growing while on the other hand vast resources remain underused. The grid paradigm aims at solving this mismatch by offering its users computational and storage resources transparently, making abstraction of the exact geographic location of the physical resource (this approach has appealing similarities to the power grid, hence the term grid computing). Despite the current deployment of operational grid systems, important challenges still lie ahead. New applications, opening the grid also for commercial exploitation, impose new requirements in terms of, for example, security, scaling behaviour, quality of service and robustness. In particular the geographic spread of grid users in combination with these new requirements will certainly have drastic consequences for the communication infrastructure. This paper discusses an overview of current grid systems, application taxonomy and emerging trends. The implications for the network are analysed, taking current grid deployments as a starting point. The importance of co-management of computational/ storage and network resources is illustrated.

An active networking based service for media transcoding in multicast sessions

Abstract: Active networking is one of the suggested technologies to introduce additional intelligence and programmability in the network and its services. In this paper, the use of active networking to support advanced multicast services providing media transcoding inside the network is investigated. In the multicast service different versions of the streamed data are made available and customers can select a specific version according to their wishes or their capabilities. Based on the active networking facilities of the underlying framework the different versions of the streamed data can be created inside the network, through transformations or transcodings of the original data. Both design and performance issues of the detailed service are discussed. A new multicast tree set-up protocol, taking into account the required transcodings, is introduced. A number of different strategies are discussed optimizing the location of the transcodings as well as the use of bandwidth in the network, while considering the availability of sufficient processing power in the nodes. The performance analysis is done for a voice stream multicast service, addressing the efficiency of the tree set-up strategies, the optimization of network resource utilization, the use of processing power for transcodings, and the resulting quality of streamed voice signals after multiple transcodings.

Adaptive Multi-mode Routing in Mobile Ad Hoc Networks

Abstract: Mobile ad hoc networks are wireless multi-hop networks with a completely distributed organization. The dynamic nature of these networks imposes many challenges on mobile ad hoc routing protocols. Current routing protocols do not take into the account the network context and therefore their performance is only optimal under certain network conditions. This paper proposes a novel concept for routing in mobile ad hoc networks, adaptive multi-mode routing, and demonstrates its feasibility and effectiveness.

Performance Evaluation of a Probabilistic Packet Filter Optimization Algorithm for High-Speed Network Monitoring

Abstract: Because of the ever-increasing popularity of the Internet, network monitoring becomes very mission critical to guarantee the operation of IP networks, e.g. to detect network failures and stop intrusion attempts. A majority of these monitoring tasks require only a small subset of all passing packets, which share some common properties such as identical header fields or similar patterns in their data. Nowadays, next to the increasing network speed, much of these tasks become very complex. In order to capture only the useful packets, these applications need to evaluate a large set of expressions. In this paper, we present a platform independent filter and pattern matcher optimization algorithm, which reduces the required number of evaluated expressions. The performance of the algorithm will be validated both analytically and by means of a high-speed monitoring system.

On the Use of NSGrid for Accurate Grid Schedule Evaluation.

Abstract: Computational Grids consist of a multitude of geographically distributed resources. The co-allocation of several of those resources allows for the execution of highly computing-intensive and data-intensive jobs. In order to obtain quality schedules (in terms of job response time and resource utilization), different factors such as resource load (computational resource load and band-width usage) and data location need to be taken into account. In addition, because of the size of realistic Grids, relevant parameters w.r.t. schedule quality can often only be obtained through simulations. In this paper, we detail how a discrete-event simulator was extended to support the simulation of scheduling both cpu- and data-intensive jobs on a Grid using network-aware scheduling algorithms. These jobs can either pre-stage data, or access data remotely while executing. We examine the case in which resource-to-resource data connections with guaranteed bandwidth can be set up, both when capacitated Virtual Private Networks (for certain job classes) are defined upfront and when connection requests for all jobs are dealt with in a pure First Come First Serve way. Our results show that in order to generate quality grid schedules, it is necessary to take into account network status information in the scheduling algorithms in order to improve the average job response time. Furthermore, we show how making upfront bandwidth reservations induces fair sharing of network resources between different job classes and thus improves the weighted average response time.

Design and implementation of an application server load balancing architecture supporting the end-to-end provisioning of value-added services

Abstract: Service providers can offer a wide variety of value added services, such as managed home networking, home-automation and security management, multimedia multiparty conferencing and content delivery. Very often hard- and software from different vendors are used to implement these services. In most cases, multiple application servers need to be deployed for each service in order to balance the load of the service sessions. We describe an architecture that offers automatic load balancing of service sessions over the available application servers, abstracting from the actual type of the service and its implementation by translating the service and vendor specific request formats into an internal XML format. The load balancing of the service sessions is done taking into account the capabilities of each application server (which service(s) does it implement), its load, and the location of running sessions (a request to join an active session has to be forwarded to the application server running the session). Furthermore, scalability is ensured by organizing the application servers into pools. New application servers can easily be added.

Optical networking research: a European perspective

Abstract: This work presents a review of optical network research activities in Europe, based on projects funded by the EU and national governments. Research networks and new initiatives are discussed.

A detailed look on orion: traffic simulations on a node level

Abstract: Previously we introduced a novel network paradigm: Overspill Routing In Optical Networks (ORION), which combined the advantages of circuit and packet switching. It has a low packet processing load on IP-routers and an efficient wavelength usage in the transport network. We have reported on extensive network dimensioning studies. We now zoom in on a node level, where packet streams can be studied in some more detail. We present a case study showing the effectiveness of ORION.

Integrating policy-based access management and adaptive traffic engineering for QoS deployment

Abstract: In this paper, we describe our experiences of designing a platform that integrates distributed policy-based access control and trust management technologies with a developed QoS-aware network management platform. The main goal of this study is to demonstrate how the application of policy-based decision processes can help to deploy the platform in a scalable way offering adaptive services with enriching features such as flexible rule-based service definitions and roaming.

Routing Strategies to Minimize Packet Loss in an Optical Packet Switched Network with Recirculating FDL Buffers

Abstract: The major goal of optical packet switching (OPS) is to match switching technology to the huge capacities provided by (D)WDM. We study optical packet switches with recirculating fiber delay line (FDL) buffers. Through simulation, we have assessed the logical performance of a single optical packet router (OPR), focusing on packet loss rate (PLR). By verifying that our scheduling algorithm does not alter the traffic profile characteristics from in- to output, we illustrate how the single node results can be used to assess network-wide performance. We use the capability of assessing end-to-end PLRs to develop network-wide routing algorithms designed to minimize the maximal PLR occurring in the network. In case studies on pan-European networks, we first compare two algorithm variants and thereafter we compare the PLR-based routing algorithm with both load balancing and shortest path routing. While load balancing achieves PLRs that are multiple orders of magnitude lower than shortest path routing, the PLR-based algorithm can reach PLRs up to two orders of magnitude better. The improvement in PLR comes at the price of only a small increase in used bandwidth (a few percent). Subsequently we show that the discussed PLR-based routing algorithm can be easily extended to multiple priorities. By introducing multiple priorities we can keep the loss rates for high priority traffic very low. However, it may lead to an increase of the obtained minimal max-PLR value for low priority traffic. But as we prove this increase to be limited, the cost of introducing multiple priorities is small.

A performance oriented Grid monitoring architecture

Abstract: Resource state monitoring is a critical component of any Grid Management Architecture, providing Grid scheduler, job/execution manager and state estimation components with accurate information about network, computational and storage resource status. Without up-to-date monitoring information, intelligent scheduling decision making would be a near-impossible task. In this paper we describe a scalable, portable and nonintrusive Grid Monitoring Architecture whose implementation decisions were made with performance in mind. We compare it to well-known Grid monitoring systems, and measure our platform’s performance to the Globus MDS2.2 and the Globus 3.2 web services Information Service (WS-IS) performance.

Analytical MMAP-based bounds for packet loss in optical packet switching with recirculating FDL buffers

Abstract: The major goal of optical packet switching (OPS) is to match switching technology to the huge capacities provided by (D)WDM. A crucial issue in packet switched networks is the avoidance of packet losses stemming from contention. In OPS, contention can be solved using a combination of exploitation of the wavelength domain (through wavelength conversion) and buffering. To provide optical buffering, fiber delay lines (FDLs) are used. In this paper, we focus on an optical packet switch with recirculating FDL buffers and wavelength converters. We introduce the Markovian arrival process with marked transitions (MMAP), which has very desirable properties as a traffic model for OPS performance assessment. Using this model, we determine lower and upper bounds for the packet loss rate (PLR) achieved by the aforementioned switch. The calculation of the PLR bounds through matrix analytical methods is repeated for a wide range of traffic conditions, including highly non-uniform traffic, both in space (i.e., packet destinations) and time (bursty traffic). The quality of these bounds is verified through comparison with simulation results.

SONAR: a platform for flexible DiffServ/MPLS traffic engineering

Abstract: A major opportunity for IP network operators to increase the revenue stream generated by their network is a closer and faster integration of services into the IP broadband network. However, with this integration come responsibilities concerning the quality, availability and continuity of these services, which are much harder to fulfil than the requirements for the current IP data traffic. SONAR (service-oriented networking using adaptive resource control) is a network management and operations platform designed to drive a DiffServ/MPLS network so that these responsibilities can be met. In this paper we discuss and evaluate the architecture and algorithms that are the basis for this platform. The architecture decomposes the network management problem in two components: a centralised path computation and a distributed tunnel management. Finally this hybrid on line/off line adaptive traffic engineering approach is evaluated in an modified NS-2 simulator environment.

Evaluation of a job admission algorithm for bandwidth constrained grids

Abstract: One of the goals envisioned by Grid computing is to make the execution of both computational and data-intensive problems possible. A key problem is finding the optimal set of resources to use for each submitted job, especially when we take network usage into consideration. To address this problem, we introduce an integer linear programming model, which incorporates the functions of job admission control and resource assignation. Scheduling decisions are based on a cost model for resources. We analyse our algorithm in terms of job acceptance rate, resource utilization and average job queueing time. The results show the correctness of the model, and prove the importance of network-awareness in grid scheduling algorithms.

Application-Specific Hints in Reconfigurable Grid Scheduling Algorithms

Abstract: In this paper, we investigate the use of application-specific hints when scheduling jobs on a Computational Grid, as these jobs can expose widely differing characteristics regarding CPU and I/O requirements. Specifically, we consider hints that specify the relative importance of network and computational resources w.r.t. their influence on the associated application’s performance. Using our ns-2 based Grid Simulator (NSGrid), we compare schedules that were produced by taking application-specific hints into account to schedules produced by applying the same strategy for all jobs. The results show that better schedules can be obtained when using these scheduling hints intelligently.