The Experts below are selected from a list of 60843 Experts worldwide ranked by ideXlab platform
Ewa Kusmierek - One of the best experts on this subject based on the ideXlab platform.
-
streaming video delivery over internet with adaptive end to end qos
Journal of Systems and Software, 2005Co-Authors: Ewa KusmierekAbstract:Many multimedia applications rely on video streaming techniques. However, video delivery over a Best Effort Network such as today's Internet is a challenge. Traffic load in the Network changes dynamically and in an unpredictable way causing the resource availability to vary. Providing an application level end-to-end quality of service in such an environment requires Network-Awareness and ability to adapt. We address the issue of mapping between application-level quality of service for streaming video and Network-level quality of service. We show that continuous playback requires a limit on the delay jitter. We tackle the problem of providing end-to-end video quality given that the Network does not guarantee limited delay variability. Our approach is unique in a way we do not model Network as a black box but investigate what information about the Network status is necessary for an application to make adaptation decisions. We rely on simple multi-level ECN-based mechanism to obtain a feedback from the Network as well as on end-point observations to determine available bandwidth. Such an approach allows to obtain better user-perceived video quality by providing additional information to properly interpret the arrival rate observed at the end-point. We propose a 3-rate adaptation mechanism for video streaming to illustrate the philosophy of adaptivity based on Network Awareness, where the Network Awareness is not limited to observing Network reaction to a set of stimuli.
Ayalvadi Ganesh - One of the best experts on this subject based on the ideXlab platform.
-
Network Awareness and failure resilience in self organizing overlay Networks
Symposium on Reliable Distributed Systems, 2003Co-Authors: Laurent Massoulie, Ayalvadi GaneshAbstract:The growth of peer-to-peer applications on the Internet motivates interest in general purpose overlay Networks. The construction of overlays connecting a large population of transient nodes poses several challenges. First, connections in the overlays should reflect the underlying Network topology, in order to avoid overloading the Network and to allow god application performance. Second, connectivity among active nodes of the overlay should be maintained, even in the presence of high failure rates or when a large proportion of nodes are not active. Finally, the cost of using the overlay should be spread evenly among peer nodes for fairness reasons as well as for the sake of application performance. To preserve scalability, we seek solutions to these issues that can be implemented in a fully decentralized manner and rely on local knowledge from each node. In this paper, we propose an algorithm called the localizer which addresses these three key challenges. The localizer refines the overlay in a way that reflects geographic locality so as to reduce Network overload. Simultaneously, it helps to evenly balance the number of neighbors of each node in the overlay, thereby sharing the load evenly as well as improving the resilience to random node failures or disconnections. The proposed algorithm is presented and evaluated in the context of an unstructured peer-to-peer overlay Network produced using the Scamp protocol. We provide a theoretical analysis of the various aspects of the algorithm. Simulation results based on a realistic Network topology model confirm the analysis and demonstrate the localizer efficiency.
Laurent Massoulie - One of the best experts on this subject based on the ideXlab platform.
-
epidemic information dissemination in distributed systems
IEEE Computer, 2004Co-Authors: Patrick Eugster, Rachid Guerraoui, Annemarie Kermarrec, Laurent MassoulieAbstract:Easy to deploy, robust, and highly resilient to failures, epidemic algorithms are a potentially effective mechanism for propagating information in large peer-to-peer systems deployed on Internet or ad hoc Networks. It is possible to adjust the parameters of epidemic algorithm to achieve high reliability despite process crashes and disconnections, packet losses, and a dynamic Network topology. Although researchers have used epidemic algorithms in applications such as failure detection, data aggregation, resource discovery and monitoring, and database replication, their general applicability to practical, Internet-wide systems remains open to question. We describe four key problems: membership maintenance, Network Awareness, buffer management, and message filtering, and suggest some preliminary approaches to address them.
-
Network Awareness and failure resilience in self organizing overlay Networks
Symposium on Reliable Distributed Systems, 2003Co-Authors: Laurent Massoulie, Ayalvadi GaneshAbstract:The growth of peer-to-peer applications on the Internet motivates interest in general purpose overlay Networks. The construction of overlays connecting a large population of transient nodes poses several challenges. First, connections in the overlays should reflect the underlying Network topology, in order to avoid overloading the Network and to allow god application performance. Second, connectivity among active nodes of the overlay should be maintained, even in the presence of high failure rates or when a large proportion of nodes are not active. Finally, the cost of using the overlay should be spread evenly among peer nodes for fairness reasons as well as for the sake of application performance. To preserve scalability, we seek solutions to these issues that can be implemented in a fully decentralized manner and rely on local knowledge from each node. In this paper, we propose an algorithm called the localizer which addresses these three key challenges. The localizer refines the overlay in a way that reflects geographic locality so as to reduce Network overload. Simultaneously, it helps to evenly balance the number of neighbors of each node in the overlay, thereby sharing the load evenly as well as improving the resilience to random node failures or disconnections. The proposed algorithm is presented and evaluated in the context of an unstructured peer-to-peer overlay Network produced using the Scamp protocol. We provide a theoretical analysis of the various aspects of the algorithm. Simulation results based on a realistic Network topology model confirm the analysis and demonstrate the localizer efficiency.
Samee Ullah Khan - One of the best experts on this subject based on the ideXlab platform.
-
DENS: Data center energy-efficient Network-aware scheduling
Cluster Computing, 2013Co-Authors: Dzmitry Kliazovich, Pascal Bouvry, Samee Ullah KhanAbstract:In modern data centers, energy consumption accounts for a considerably large slice of operational expenses. The state of the art in data center energy optimization is focusing only on job distribution between computing servers based on workload or thermal profiles. This paper underlines the role of communication fabric in data center energy consumption and presents a scheduling approach that combines energy efficiency and Network Awareness, termed DENS. The DENS methodology balances the energy consumption of a data center, individual job performance, and traffic demands. The proposed approach optimizes the tradeoff between job consolidation (to minimize the amount of computing servers) and distribution of traffic patterns (to avoid hotspots in the data center Network).
Konstantin Pussep - One of the best experts on this subject based on the ideXlab platform.
-
Peer-assisted Video-on-Demand: Cost Reduction and Performance Enhancement for Users, Overlay Providers, and Network Operators
Arbeit, 2011Co-Authors: Konstantin PussepAbstract:Peer-assisted content delivery is an attractive way to distribute video content through the Internet at low costs. This approach combines the scalability of the peer-to-peer paradigm, in which users contribute their resources, and the service level guarantees of server-based systems. Thus, peer-assistance enables a content provider to reduce its server hosting costs, which is crucial in a commercial scenario. However, to be successful, such systems must take into account the interests of all three stakeholders involved: (1) users that demand high streaming quality with low fees and limited resource contribution, (2) content providers that aim to decrease server hosting costs, and (3) Network operators that aim to avoid inefficient use of their infrastructure due to the Network-oblivious behavior of peer-assisted overlays. In this thesis, we address these requirements and develop adaptive mechanisms to achieve a benefit for all three stakeholders, resulting in the so-called triple-win situation. Our main scenario is video-on-demand streaming, in which users can request pre-stored video content at any time and watch the video while downloading it. Thereby, video-on-demand streaming imposes stricter requirements compared to other systems that utilize peer resources, such as BitTorrent-like file sharing. Ideally, the video playback should start within few seconds and there should be no playback stalling. First, we focus on dedicated servers or caches that are essential resources in peerassisted streaming systems. Their provision is necessary to guarantee a satisfying quality of experience to consumers, yet they cause significant and largely avoidable costs for the content provider, which can be minimized. The high dynamics of uncontrolled peers, however, result in unpredictable changes of the resource demand. Since peers additionally offer services to other peers, the supply of resources is also dynamic. This behavior makes the management of peer-assisted systems and the proper allocation of resources challenging. This thesis proposes adaptive allocation policies, a new approach to address this issue. The policies estimate the capacity situation and service demand of the system to adaptively optimize allocated resources. Extensive simulations, verified by testbed measurements, prove the efficiency of our approach, which achieves a more competitive performance than well-dimensioned static systems. In the next step, we examine content delivery overlays from the Network operators’ perspective, since such overlays are responsible for a large amount of consumer traffic, including the costly inter-domain traffic. The existing approaches often fail to satisfy the requirements of all involved stakeholders.We propose a novel incentive-based traffic management mechanism where a Network operator offers additional free resources to selected users. The mechanism assigns resources to users that behave compliant with the Network and overlay policies. Our evaluation shows that this approach satisfies the requirements of Network operators and overlay participants (provider and users). To this end, the proposed mechanism is able to reduce the inter-domain traffic while improving the overlay performance.We also show that even a single Network operator can successfully apply the proposed mechanism. Finally, we consider the availability of peer resources in peer-assisted content distribution. Besides contributing upload bandwidth, it is important that peers stay online after finishing their downloads to serve new download requests. The longer a peer stays online the more it can help to offload the servers. However, too extensive online time often results in high energy consumption paid by users without an adequate benefit to the system. In upcoming decentralized architectures based on set-top boxes that act as tiny servers, their energy consumption can even dominate distribution costs. Therefore, we propose advanced standby policies that reduce the energy consumption of set-top boxes while still offloading servers significantly. We evaluate the standby polices in a lifelike scenario. The results show that a near-optimal behavior can be realized by utilizing common features of set-top boxes such as the wake-up timer. We further extend a standby policy with Network Awareness to address the needs of Network operators. In this regard, the resulting policy takes into account the interests of all three stakeholders: users, content providers, and Network operators.