Conor McArdle

Evolutionary trends in intelligent networks

A number of groups are currently developing technologies aimed at evolving and enhancing the capabilities of intelligent networks. In this article we discuss three of these initiatives: PINT, Parlay, and IN/CORBA interworking. The IETF PINT work addresses how Internet applications can request and enrich telecommunications services. The Parlay consortium is specifying an object-oriented service control API that facilitates the access, control, and configuration of IN services by enterprise IT systems. The OMGs IN/CORBA interworking specification enables CORBA-based systems to interwork with existing IN infrastructure, thereby promoting the adoption of CORBA for the realization of IN functional entities. We address how all three of these technologies could be deployed together in order to provide a basis for a more flexible and open IN architecture. We also identify a number of common trends and potential pitfalls highlighted by current work on the evolution of IN.

Distributed management of energy-efficient lightpaths for computational grids

Information and Communication Technologies (ICTs) are contributing to a large amount of the global electricity consumption. Due to tremendous increase in the bandwidth demands and utilisation of non-renewable energy resources Greenhouse Gas Emissions are increasing proportionally with the increasing demand. Despite their advantages in terms of computing performance, distributed applications such as computational grids are major factors that increase the traffic volume in the Internet. In this paper, we propose a distributed framework to ensure energy savings in the optical WDM backbone which transport the traffic between nodes and several computational grids based on anycast routing. According to the proposed framework, the backbone nodes go to sleep mode and resume active mode in a distributed manner with the objective of maximum energy savings in the backbone. Each node maintains two thresholds which are adaptively adjusted based on the network performance, and these thresholds play the key role in determining the decision of a node whether to sleep or resume. Numerical results confirm that the proposed framework can ensure significant energy savings in the network when compared to the conventional energy-unaware operation mode. We further show that the adoption of the proposed network framework does not degrade significantly the network performance in terms of average blocking probability and end-to-end delay.

Simplified overflow analysis of an optical burst switch with fibre delay lines

We develop an approximate analytic model of an Optical Burst Switch with share-per-node fibre delay lines and tuneable wavelength converters by employing Equivalent Random Theory, an approach from circuit-switching analysis. Our model is formulated in terms of virtual traffic flows within the switch from which we derive expressions for burst blocking probability, fibre delay line occupancy and mean delay, which we then resolve numerically. Emphasis is on simplicity of the model to achieve good numerical efficiency so that the method can be useful for formulating dimensioning problems for large-scale networks. Solution values from the analysis are compared with discrete-event simulation results.

Analysis of a Buffered Optical Switch with General Interarrival Times

We propose an approximate analytic model of an optical switch with fibre delay lines and wavelength converters by employing Equivalent Random Theory. General arrival traffic is modelled by means of Gamma-distributed interarrival times. The analysis is formulated in terms of virtual traffic flows within the optical switch from which we derive expressions for burst blocking probability, fibre delay line occupancy and mean delay. Emphasis is on approximations that give good numerical efficiency so that the method can be useful for formulating dimensioning problems for large-scale networks. Numerical solution values from the proposed analysis method compare well with results from a discrete-event simulation of an optical burst switch.

A two-moment performance analysis of optical burst switched networks with shared fibre delay lines in a feedback configuration

Abstract Fibre delay lines (FDLs) can substantially reduce the burst loss in Optical Burst Switching (OBS) networks and share-per-node FDL configurations can provide a more cost-efficient solution compared to architectures where delay lines are shared per port. Nevertheless, mathematical performance analysis of this configuration is more difficult due to traffic correlations arising from the shared resource. In this paper, an approximate two-moment traffic model is developed for quantifying end-to-end burst blocking probability in networks of OBS switches with share-per-node FDLs. The two-moment approach can improve model accuracy over more usual Poisson network analysis methods and additionally allows the characteristics of offered load to be taken into account. The accuracy of the proposed method is shown to be favourable, when compared to discrete-event simulations of an OBS network.

Optical packet switch with energy-efficient hybrid optical/electronic buffering for data center and HPC networks

Advanced optical switching architectures, capable of scaling to thousands of ports while achieving low communication latency and reduced power consumption, are becoming a dominant theme for interconnection networks in next-generation data centers and high-performance computing systems. The arrayed waveguide grating (AWG) device, with its inherent ability to perform wavelength routing of many wavelengths in parallel, has been recognized as a promising core component for fast optical switching. Although the AWG is energy efficient (as essentially a passive optical device), has high-bandwidth switching capabilities and has relative simplicity and low cost, an inherent characteristic of switching schemes based on the AWG is potential wavelength oversubscription at switch output ports, which can lead to high packet blocking probabilities. To resolve this traffic congestion, this paper proposes a hybrid optical/electronic buffering scheme and a method for efficiently integrating fiber delay line buffer capacity into the AWG wavelength assignment scheme. The dimensioning of the optical and electronic buffer resources is then carried out using simulations. The results indicate that with the proper dimensioning, the hybrid-buffered AWG switch achieves significantly increased overall energy efficiency, compared to electronic-only buffering, while maintaining low latency and non-blocking performance. We also investigate the computational complexity of the required scheduling algorithm in the hybrid-buffered switch, which in turn allows us to estimate the required processing power of the switch controller.

Load Balancing for a Distributed CORBA-Based SCP

This paper examines load balancing issues relating to a distributed CORBA-based Service Control Point. Two types of load balancing strategies are explored through simulation studies: (i) A novel ant-based load balancing algorithm, which has been devised specifically for this type of system. This algorithm is compared to more traditional algorithms, (ii) A method for optimal distribution of the computational objects composing the service programs. This is based on mathematically minimising the expected communication flows between network nodes and message-level processing costs. The simulation model has been based on the recently adopted OMG IN/CORBA Interworking specification and the TINA Service Session computational object model.

Large-scale hybrid electronic/optical switching networks for datacenters and HPC systems

We propose a novel, hybrid optical-electronic switched network architecture with flexible bandwidth provisioning for scaling the capacity of the upper tiers of next generation datacenter and high-performance computing (HPC) networks. The network combines transparent optical packet switches, based on Arrayed Waveguide Grating (AWG) routers, and a deflection routing scheme to alleviate packet buffering requirements in the optical domain. Through simulation studies we show that, coupled with a dynamic switch-port bandwidth provisioning and load balancing scheme, the network can maintain high performance under changing traffic loading.

Energy-efficient lightpaths for computational grids

Optical networks have been pointed out as strong candidates to ensure energy-efficiency in the Internet backbone, as well as the distributed applications such as computational grids. In this paper, we propose a distributed framework to address energy-efficient lightpath establishment problem for computational grids over the optical WDM backbone. Each backbone node maintains two thresholds to determine its sleep/wakeup cycle. A node in the sleep mode saves energy by rejecting transient traffic. Each demand is routed based on anycast routing and with the objective of minimum power consumption. Through numerical results, we show that significant energy savings are attained by the proposed framework without requiring centralized information and control. We further show that coordinated sleep and wakeup management in the optical backbone can address the trade-off between propagation delay and energy savings.

Energy-efficient optical packet switch with recirculating fiber delay line buffers for data center interconnects

Potentially, optical packet switching architectures can alleviate the power consumption, bandwidth and connectivity constraints that existing electronically-switched networks are encountering as interconnects in Data centers and High Performance Computing (HPC) platforms scale out to many thousands of processing nodes. In this paper, we propose a novel optical packet switch architecture, based on the Arrayed Waveguide Grating Router (AWGR), which requires many less tunable wavelength converters than wavelength channels and where contention is efficiently resolved using a low-power re-circulating optical delay module. Compared to previous schemes which propose using an electronic buffer to resolve packet contentions in this type of architecture, we show through simulation that our scheme can improve scalability, by reducing switch power consumption by up to 50% while maintaining low packet latencies. By proper integration of the buffer into the switch and correct dimensioning of the optical buffers, packet loss rates of <; 10-10 are achievable.