N. Merayo

A Cognitive Quality of Transmission Estimator for Core Optical Networks

We propose a cognitive Quality of Transmission (QoT) estimator for classifying lightpaths into high or low quality categories in impairment-aware wavelength-routed optical networks. The technique is based on Case-Based Reasoning (CBR), an artificial intelligence technique which solves new problems by exploiting previous experiences, which are stored on a knowledge base. We also show that by including learning and forgetting techniques, the underlying knowledge base can be optimized, thus leading to a significant reduction on the computing time for on-line operation. The performance of the cognitive estimator is evaluated in a long haul and in an ultra-long haul network, and we demonstrate that it achieves more than 98% successful classifications, and that it is up to four orders of magnitude faster when compared with a non-cognitive QoT estimator, the Q-Tool.

Implementation of a PID controller for the bandwidth assignment in long-reach PONs

In this paper a proportional-integral-derivative (PID) controller to efficiently allocate bandwidth in high coverage passive optical networks (PONs) is presented. The novelty of this proposal relies on the fact that this is the first proposal to use a PID to control a network parameter in PONs. As the PID takes into account present, past and possible future errors in the bandwidth adjustment, this self-adapting technique results in a very robust process. Simulation results exhibit that not only is it faster and more stable than other algorithms, but also it auto-adapts the resources very efficiently when facing real time changes in the network parameters or in the bandwidth conditions. In fact, the standard deviation of the difference between the allocated and the guaranteed bandwidth is reduced by up to 50% and the convergence speed is up to four times quicker than other proposals.

A bandwidth assignment polling algorithm to enhance the efficiency in QoS long‐reach EPONs

A novel polling algorithm is proposed to provide subscriber differentiation at the upstream channel in long-reach EPONs. As operators and service providers prefer to operate at loads with no congestion and losses, the developed algorithm has been designed to improve the efficiency for those loads at which the network is likely to work. This new scheme permits to anticipate the transmission of some packets in order to take advantage of the wasted bandwidth between consecutive transmissions of ONUs. As a result, not only does this algorithm behave alike other polling algorithms for high loads, but it also improves the efficiency at low and medium loads, leading to a reduction in the mean packet delay. The new algorithm has been tested in different scenarios, motivated by the increasing interest in enlarging the long-reach EPON distances. Simulation results show that as the distance increases, the algorithm achieves a significant reduction of the mean packet delay for an extended range of loads. Copyright

Cognition to design energetically efficient and impairment aware virtual topologies for optical networks

“Greening the Internet” is an important research topic in the last years. The Internet capacity and energy consumption have increased, and the utilization of design and operation techniques to reduce this consumption are a must. In this paper, we present a multiobjective genetic algorithm to design virtual topologies for reconfigurable wavelength-routed optical networks with the aim of reducing both the energy consumption and the network congestion while ensuring that the lightpaths of the virtual topologies fulfill quality of transmission requirements. Moreover, we also present another version of that method enhanced with cognitive techniques, and we show, by means of simulation, the performance advantages brought when introducing these cognitive techniques.

EPON bandwidth allocation algorithm based on automatic weight adaptation to provide client and service differentiation

In this article, a new dynamic bandwidth allocation algorithm for the upstream channel of Ethernet Passive Optical Networks, called DySLa, is proposed not only to provide service differentiation but also to offer subscriber differentiation. In contrast to previous methods in which the performance of each class of service is not insured, DySLa is continuously evaluating the mean packet delay and guarantees that the highest priority services fulfil the packet delay requirements in the access network for every type of client. Simulation results show that DySLa can maintain both the mean packet delay and packet loss ratio below the maximum upper bounds permitted for the most sensitive services of every class of customer. Moreover, DySLa outperforms other dynamic bandwidth allocation algorithms which provide service and client differentiation and it makes a fairer bandwidth distribution than those methods.

Multiobjective Genetic Algorithm to Design Cost-Efficient Wavelength-Routed Optical Networks

A new algorithm, GAPDELT, is presented to jointly design the resource provisioning and the logical topology of wavelength-routed optical networks. GAPDELT is based on genetic algorithms and uses Pareto optimality to reduce both the network congestion and the number of resources employed. By means of a simulation study, we show that GAPDELT designs a set of good alternatives that achieve up to 50% of reduction in terms of congestion or 75% of reduction in the number of transmitters and receivers required when compared with another well-known

An auto-tuning PID control system based on genetic algorithms to provide delay guarantees in Passive Optical Networks

Abstract Passive Optical Networks (PONs) are the most important access architectures since their deployment is massive all around the world. However, the QoS (Quality of Service) and the efficient management of the network resources have become the key point, especially with the new emerging services and applications. In particular, the delay and the bandwidth are becoming important limiting factors for the user experience. As a consequence, in a previous research we proposed the implementation of PID (Proportional-Integral-Derivative) control strategies to manage these networks parameters in PONs, demonstrating higher efficiency and more robustness than other previous existing algorithms. It is worth emphasizing that this is the first time to apply this control strategy in PONs access networks. However, in this paper we improve the PID control strategy by automating the tuning process with a genetic algorithm. Indeed, we have developed a novel automatic tuning technique based on genetic algorithms to tune a P controller that provides delay guarantees. Simulation results show that the control strategy reduces the tuning time up to 64% in comparison with the Ziegler–Nichols manual technique (ZN). On the other hand, it is demonstrated that our proposal is more accurate and robust that ZN since the genetic algorithm automatically evolves to the best solutions of the tuning parameters in contrast to the manual experiments required for the ZN method. Furthermore, we have complemented the use of the P controller with a new dynamic Admission Control (AC) module. This module implements a policy to selectively transmit or drop packets and leads a better delay control. The simulation analysis reveals that the real time evolution of the delay with the dynamic AC is more stable when compare to a conventional and simple fixed AC, reaching differences near one order of magnitude in the delay fluctuations.

A cognitive system for fast Quality of Transmission estimation in core optical networks

A cognitive system is proposed to quickly estimate the Quality of Transmission (QoT) of lightpaths in optical networks. It achieves more than 99% successful classifications of lightpaths into high or low QoT categories.

Performance comparison of methods to solve the Routing and Spectrum Allocation problem

Future optical networks may increase their efficiency and flexibility by the utilization of Orthogonal Frequency Division Multiplexing (OFDM). When such a technique is employed, the spectrum assigned to a lightpath can be tailored to suit its requirements. Thus, when addressing the establishment of optical connections (lightpaths) in flexible optical networks, rather than solving the well-known Routing and Wavelength Assignment (RWA) problem, the Routing and Spectrum Assignment (RSA) problem should be solved. In this paper, we propose and compare a number of techniques to solve the RSA problem.

Genetic algorithm to design logical topologies in reconfigurable WDM networks

A new method to design logical topologies based on genetic algorithms is presented. Not only does the algorithm determine which nodes should be connected by means of lightpaths with the aim of minimizing congestion, but it also solves the routing and wavelength assignment problem. In this way, the algorithm guarantees that the logical topology obtained can be embedded in the optical network subject to the available set of resources. The algorithm is effective in terms of both congestion and fairness. For instance, when compared with other work, the congestion is significantly reduced (from 20% to 75% depending on the matrix of traffic considered), and the fairness, when evaluated in terms of the Jain index, is generally higher than 0.94. Moreover, the algorithm brings advantages when employed in dynamic scenarios where the logical topology is frequently reconfigured, as it is fast and, in contrast to other algorithms previously proposed, the calculation process can be stopped at any time (if required) in order to give the best virtual topology found up to the moment.