The Experts below are selected from a list of 279 Experts worldwide ranked by ideXlab platform
Erol Gelenbe - One of the best experts on this subject based on the ideXlab platform.
-
Big Data for Autonomic Intercontinental Overlays
IEEE Journal on Selected Areas in Communications, 2016Co-Authors: Olivier Brun, Lan Wang, Erol GelenbeAbstract:This paper uses Big Data and Machine Learning for the real-time management of Internet scale Quality-of-Service Route Optimisation with an overlay Network. Based on the collection of data sampled each 2 minutes over a large number of source-destinations pairs, we show that intercontinental Internet Protocol (IP) paths are far from optimal with respect to Quality of Service (QoS) metrics such as end-to-end round-trip delay. We therefore develop a machine learning based scheme that exploits large scale data collected from communicating node pairs in a multi-hop overlay Network that uses IP between the overlay nodes, and selects paths that provide substantially better QoS than IP. Inspired from Cognitive Packet Network protocol, it uses Random Neural Networks with Reinforcement Learning based on the massive data that is collected, to select intermediate overlay hops. The routing scheme is illustrated on a 20-node intercontinental overlay Network that collects some 2 × 10^6 measurements per week, and makes scalable distributed routing decisions. Experimental results show that this approach improves QoS significantly and efficiently.
-
EPEW - Approximate Analysis of a Round Robin Scheduling Scheme for Network Coding
Computer Performance Engineering, 2009Co-Authors: Omer H. Abdelrahman, Erol GelenbeAbstract:Network coding (NC) has been proposed as a way to compress flows of Packets by combining different Packets, provided that there is sufficient redundancy through multiple transmission paths so that the receiving nodes can then decode the flows to reconstruct all of the individual Packets. However NC does introduce additional computational overhead, and also creates different additional delays which are the subject of this paper. In particular, we evaluate the queueing and delay performance of NC at intermediate nodes of a store and forward Packet Network when the cross-encoding of Packets is restricted within disjoint subsets of the traffic streams so that Packets from different subsets cannot be encoded together. We propose a round robin scheduling scheme for serving these disjoint subsets, and present a queueing model for a single encoding node that captures the effect of NC. The model is analyzed approximately using a decoupling approach, and can be used to predict the additional delays incurred by Packets in nodes that use NC. The accuracy of the analytical solution is validated via simulations.
-
Admission of QoS aware users in a smart Network
ACM Transactions on Autonomous and Adaptive Systems, 2008Co-Authors: Erol Gelenbe, Georgia Sakellari, Maurizio D'arienzoAbstract:Smart Networks have grown out of the need for stable, reliable, and predictable Networks that will guarantee Packet delivery under Quality of Service (QoS) constraints. In this article we present a measurement-based admission control algorithm that helps control traffic congestion and guarantee QoS throughout the lifetime of a connection. When a new user requests to enter the Network, probe Packets are sent from the source to the destination to estimate the impact that the new connection will have on the QoS of both the new and the existing users. The algorithm uses a novel algebra of QoS metrics, inspired by Warshall's algorithm, to look for a path with acceptable QoS values to accommodate the new flow. We describe the underlying mathematical principles and present experimental results obtained by evaluating the method in a large laboratory test-bed operating the Cognitive Packet Network (CPN) protocol.
-
ICONIP - Cognitive Routing in Packet Networks
Neural Information Processing, 2004Co-Authors: Erol GelenbeAbstract:While the Internet has beome a resounding success for human communication and information gathering and dissemination, Quality of Service (QoS) remains a major subject of concern for users of Packet Networks. In this paper we show how distributed intelligence and on-line adaptation can be used to enhance QoS for users of Packet Networks, and present our Cognitive Packet Network approach as a manner to address this important issue. Our presentation is based on ongoing experimental reserach that uses several “Cognitive Packet Network” (CPN) test-beds that we have developed.
-
Design and performance of cognitive Packet Networks
Performance Evaluation, 2001Co-Authors: Erol Gelenbe, Ricardo LentAbstract:Abstract We discuss a Packet Network architecture called a cognitive Packet Network (CPN), in which intelligent capabilities for routing and flow control are moved towards the Packets, rather than being concentrated in the nodes and protocols. Our architecture contains “smart” and “dumb” Packets, as well as acknowledgement Packets. Smart CPN Packets route themselves, and learn to avoid congestion and losses from their own observations about the Network and from the experience of other Packets. They use a reinforcement learning algorithm to route themselves based on a goal function which has been assigned to them for each connection. Dumb CPN Packets of a specific quality of service (QoS) class use routes which have been selected by the smart Packets (SPs) of that class. Acknowledgement (ACK) Packets are generated by the destination when an SP arrives there; the ACK heads back to the source of the SP along the inverse route and is used to update mailboxes in CPN routers, as well as to provide source routing information for dumb Packets. We first summarize the basic concepts behind CPN, and present simulations illustrating their performance for different QoS goals, and analytical results for best and worst case performance. We then describe a test-bed Network we have designed and implemented in order to demonstrate these ideas. We provide measurement data on the test-bed to illustrate the capacity of the Network to adapt to changes in traffic load and to failures of links. Finally, we use measurements to evaluate the impact of the ratio of smart to dumb Packets on the end-to-end delay experienced by all of the Packets.
M.j. Karol - One of the best experts on this subject based on the ideXlab platform.
-
Large capacity multiaccess optical Packet Network
IEEE Photonics Technology Letters, 1994Co-Authors: Bernard Glance, M.j. KarolAbstract:A multiaccess optical Packet Network capable of providing multigigabit-per-second transmission per user is presented. The proposed system is based on wavelength division multiplexing and frequency routing. It is optically transparent, has no internal blocking, and is optically self-routing. Users freely transmit their Packets. Collision-free reception is achieved by means of a novel fast-tunable optical filter that controls the Packet flow. Reception acknowledgments are passively generated without the use of optical sources and are transported on the same Packet Network. Assuming a single transmitter per user, and either one or two receivers (each with its own fiber) per user, the system throughput can approach 60% and 90% (per wavelength), respectively. >
Bernard Glance - One of the best experts on this subject based on the ideXlab platform.
-
Large capacity multiaccess optical Packet Network
IEEE Photonics Technology Letters, 1994Co-Authors: Bernard Glance, M.j. KarolAbstract:A multiaccess optical Packet Network capable of providing multigigabit-per-second transmission per user is presented. The proposed system is based on wavelength division multiplexing and frequency routing. It is optically transparent, has no internal blocking, and is optically self-routing. Users freely transmit their Packets. Collision-free reception is achieved by means of a novel fast-tunable optical filter that controls the Packet flow. Reception acknowledgments are passively generated without the use of optical sources and are transported on the same Packet Network. Assuming a single transmitter per user, and either one or two receivers (each with its own fiber) per user, the system throughput can approach 60% and 90% (per wavelength), respectively. >
-
Performance of the PAC optical Packet Network
IEEE Global Telecommunications Conference GLOBECOM '91: Countdown to the New Millennium. Conference Record, 1Co-Authors: Mark J. Karol, Bernard GlanceAbstract:The performance aspects of the PAC (protection-against-collision) optical Packet Network are discussed. An analysis is made of the delay-throughput performance of this Network for uniform traffic patterns. Results show that in geographically distributed applications the maximum achievable throughput (normalized to the transmission rate) is typically between 0.4 and 0.5 per channel. For example, the maximum achievable throughput is between 400 and 500 Gb/s for a 1000-channel system with a 1-Gb/s transmission rate. In a centralized switch, the (normalized) maximum achievable throughput can approach 0.8 per channel. These values are higher than previously published optical systems. >
R.v. Richard V. Penty - One of the best experts on this subject based on the ideXlab platform.
-
WASPNET: A Wavelength Switched Packet Network
IEEE Communications Magazine, 1999Co-Authors: David K. Hunter, Meow C. Chia, John D. Bainbridge, Marc F C Stephens, Michael Joseph Mahony, Mohd Hairul Nizam, Ken M. Guild, Ivan Andonovic, Anna Tzanakaki, R.v. Richard V. PentyAbstract:WASPNET is an EPSRC-funded collaboration between three British Universities: the University of Strathclyde, Essex University, and Bristol University, supported by a number of industrial institutions. The project which is investigating a novel Packet-based optical WDM transport Network-involves determining the management, systems, and devices ramifications of a new Network control scheme, SCWP, which is flexible and simplifies optical hardware requirements. The principal objective of the project is to understand the advantages and potential of optical Packet switching compared to the conventional electronic approach. Several schemes for Packet header implementation are described, using subcarrier multiplexing, separate wave lengths, and serial transmission. A novel node design is introduced, based on wavelength router devices, which reduce loss, hence reducing booster amplifier gain and concomitant ASE noise. The fabrication of these devices, and also wavelength converters, are described. A photonic Packet switching testbed is detailed which will allow the ideas developed within WASPNET to be tested in practice, permitting the practical problems of their implementation to be determined
Claudio Sansoe - One of the best experts on this subject based on the ideXlab platform.
-
FPL - An FPGA-based Node Controller for a High Capacity WDM Optical Packet Network
Lecture Notes in Computer Science, 2002Co-Authors: Roberto Gaudino, Vito De Feo, Marcello Chiaberge, Claudio SansoeAbstract:We present architecture and experimental results for an highspeed digital node controller developed in the "RINGO" project, a demonstrator of a next-generation ultra-high capacityn etwork based on WDM optical fiber ring. The MAC protocol and other controller functions have been developed on a high capacityF PGA, which supports efficient input queuing and multicast capabilities byh andling several concurrent real time processes.
