Lúcia Martins

A survey of strategies for communication networks to protect against large-scale natural disasters

Recent natural disasters have revealed that emergency networks presently cannot disseminate the necessary disaster information, making it difficult to deploy and coordinate relief operations. These disasters have reinforced the knowledge that telecommunication networks constitute a critical infrastructure of our society, and the urgency in establishing protection mechanisms against disaster-based disruptions. Hence, it is important to have emergency networks able to maintain sustainable communication in disaster areas. Moreover, the network architecture should be designed so that network connectivity is maintained among nodes outside of the impacted area, while ensuring that services for costumers not in the affected area suffer minimal impact. As a first step towards achieving disaster resilience, the RECODIS project was formed, and its Working Group 1 members conducted a comprehensive literature survey on “strategies for communication networks to protect against large-scale natural disasters,” which is summarized in this article.

A New Multiobjective Dynamic Routing Method for Multiservice Networks: Modelling and Performance

There are potential advantages in formulating the routing problems in modern multiservice networks as multiple objective problems. This paper presents a novel hierarchical bi-level multiobjective dynamic routing model for multiservice networks. It is based on a bi-objective shortest path algorithm, with dynamically adapted soft-constraints, to compute alternative paths for each node pair and on a heuristic to synchronously select alternative routing plans for the network in a dynamic alternative routing context. It is a routing method which periodically changes alternative paths as a function of periodic updates of certain QoS related parameters obtained from real-time measurements. The performance of the proposed routing method is compared with two reference dynamic routing methods namely RTNR and DAR by means of a discrete-event simulator.

An efficient algorithm for sequential generation of failure states in a network with multi-mode components

Abstract In this work a new algorithm for the sequential generation of failure states in a network with multi-mode components is proposed. The algorithm presented in the paper transforms the state enumeration problem into a K -shortest paths problem. Taking advantage of the inherent efficiency of an algorithm for shortest paths enumeration and also of the characteristics of the reliability problem in which it will be used, an algorithm with lower complexity than the best algorithm in the literature for solving this problem, was obtained. Computational results will be presented for comparing the efficiency of both algorithms in terms of CPU time and for problems of different size.

On a bi-dimensional dynamic alternative routing method

Abstract The analysis of a bi-dimensional dynamic routing model for alternative routing telecommunication networks led to the identification of an instability problem in the synchronous path selection associated with the complex interdependencies among the coefficients of the objective functions and the computed paths for every node pair. In this paper an analytical model enabling to make explicit this problem and evaluate its effects in terms of two global network criteria, is presented. Also a heuristic procedure dedicated to overcome this instability problem and select “good” compromise solutions in terms of network performance is developed. Finally the performance of the proposed routing method using the heuristic is compared by recurring to discrete-event simulation with a reference dynamic routing method (Real Time Network Routing) for some test networks.

Protected shortest path visiting specified nodes

In this paper heuristics are proposed for finding the shortest loopless path, from a source node to a target node, that visits a given set of nodes in a directed graph, such that it can be protected using a node-disjoint path. This type of problem may arise due to network management constraints. The problem of calculating the shortest path that visits a given set of nodes is at least as difficult as the traveling salesman problem, and it has not received much attention. Nevertheless an efficient integer linear programming (ILP) formulation has been recently proposed for this problem. Here, the ILP formulation is adapted to include the constraint that the obtained path will be able to be protected by a node-disjoint path. Computational experiments show that this approach, namely in large networks, may fail to obtain a solution in a reasonable amount of time. Therefore three versions of a heuristic are proposed, for which extensive computational results show that they are able to find a solution in most cases, and that the calculated solutions present an acceptable relative error regarding the cost of the optimal active path. Further the CPU time required by the heuristics is significantly smaller than the required by the used ILP solver.

Algorithms for determining a node-disjoint path pair visiting specified nodes☆

Abstract The problem of calculating the shortest path that visits a given set of nodes is at least as difficult as the traveling salesman problem, and it has not received much attention. Nevertheless an efficient integer linear programming (ILP) formulation has been recently proposed for this problem. That ILP formulation is firstly adapted to include the constraint that the obtained path can be protected by a node-disjoint path, and secondly to obtain a pair of node disjoint paths, of minimal total additive cost, each having to visit a given set of specified nodes. Computational experiments show that these approaches, namely in large networks, may fail to obtain solutions in a reasonable amount of time. Therefore heuristics are proposed for solving those problems, that may arise due to network management constraints. Extensive computational results show that they are capable of finding a solution in most cases, and that the calculated solutions present an acceptable relative error regarding the cost of the obtained path or pair of paths. Further the CPU time required by the heuristics is significantly smaller than the required by the used ILP solver.

An effective algorithm for computing all-terminal reliability bounds

The exact calculation of all-terminal reliability is not feasible in large networks. Hence estimation techniques and lower and upper bounds for all-terminal reliability have been utilized. Here, we propose using an ordered subset of the mincuts and an ordered subset of the minpaths to calculate an all-terminal reliability upper and lower bound, respectively. The advantage of the proposed new approach results from the fact that it does not require the enumeration of all mincuts or all minpaths as required by other bounds. The proposed algorithm uses maximally disjoint minpaths, prior to their sequential generation, and also uses a binary decision diagram for the calculation of their union probability. The numerical results show that the proposed approach is computationally feasible, reasonably accurate and much faster than the previous version of the algorithm. This allows one to obtain tight bounds when it not possible to enumerate all mincuts or all minpaths as revealed by extensive tests on real-world networks.

A Hierarchical Multiobjective Routing Model for MPLS Networks with Two Service Classes

This work presents a model for multiobjective routing in MPLS networks formulated within a hierarchical network-wide optimization framework, with two classes of services, namely QoS and Best Effort (BE) services. The routing model uses alternative routing and hierarchical optimization with two optimization levels, including fairness objectives. Another feature of the model is the use of an approximate stochastic representation of the traffic flows in the network, based on the concept of effective bandwidth. The theoretical foundations of a heuristic strategy for finding “good” compromise solutions to the very complex bi-level routing optimization problem, based on a conjecture concerning the definition of marginal implied costs for QoS flows and BE flows, will be described. The main features of a first version of this heuristic based on a bi-objective shortest path model and some preliminary results for a benchmark network will also be revealed.

Evaluation of a Multiobjective Alternative Routing Method in Carrier IP/MPLS Networks

We present a first simplified version of the MultiObjective Dynamic Routing (MODR) method, more suitable for a realistic network environment as the computational effort is very much reduced while good results can still be reached. The simplified version presented herein is based on the results obtained from a discrete event simulation study which shows that, in case of overload, more important than the alternative routing algorithm itself is to control the excess of alternative routing traffic. Moreover, in a multiservice network in the case of lightly loaded traffic conditions, when alternative routing starts to be effective, network performance can still be improved if we can avoid alternative routing for specific traffic flows. Classical dynamic alternative routing methods for traditional ISDN networks have a trunk reservation mechanism with a similar purpose but apparently without the same performance. Our method applies to MPLS strongly meshed networks which are typical of core networks.

An algorithm for computing all-terminal reliability bounds

The exact calculation of all-terminal reliability is not feasible in large networks. Hence estimation techniques and lower and upper bounds for all-terminal reliability have been utilized. We propose using an ordered subset of the mincuts and an ordered subset of minpaths to calculate an all-terminal reliability upper and lower bound, respectively. The advantage of the proposed approach results from the fact that it does not require the enumeration of all mincuts or all minpaths as required by other bounds. The performance of the algorithm is compared with the first two Bonferroni bounds, for networks where all mincuts could be calculated. The results show that the proposed approach is computationally feasible and reasonably accurate. Thus allowing one to obtain bounds when it not possible to enumerate all mincuts or all minpaths.