Abstract This article considers resilience of service networks that are composed of service and control nodes to node‐targeted attacks. Two complementary problems of selecting attacked nodes and placing control nodes reflect the interaction between the network operator and the network attacker. This interaction can be analyzed within the framework of game theory. Considering the limited performance of the previously introduced iterative solution algorithms based on non‐compact problem models, new compact integer programming formulations of the node attack optimization problem are proposed, which are based on the notion of pseudo‐components and on a bilevel model. The efficiency of the new formulations is illustrated by the numerical study that uses two reference networks (medium‐size and large‐size), and a wide range of the sizes of attacks and controllers placements.

Abstract In software defined networks (SDN) packet data switches are configured by a limited number of SDN controllers, which respond to queries for packet forwarding decisions from the switches. To enable optimal control of switches in real time the placement of controllers at network nodes must guarantee that the controller‐to‐controller and switch‐to‐controller communications delays are bounded. Apart from the primary controllers that control the switches in the nominal state, separate backup controllers can be introduced that take over when the primary controllers are unavailable, and whose delay bounds are relaxed. In this paper, we present optimization models to jointly optimize the placement of primary and backup controllers in long‐distance SDN networks, aimed at maximizing the network's resilience to node‐targeted attacks. Applying the models to two well‐known network topologies and running a broad numerical study we show that, when compared with the standard approach of using only primary controllers, the use of backup controllers provides significant resilience gains, in particular in case of tight delay bounds.

As the increasing traffic patterns show asymmetric feature, directed pre-configured-cycle (p-cycle) has indicated the ability of better protection in elastic optical networks (EONs). In this paper, we investigate three different integer linear program (ILP) models of directed p-cycle without candidate cycle enumeration leveraging flow conservation. Three directed p-cycle designs are based on the same directed p-cycle strategy but differ from each other in how the flows can construct the directed p-cycles, namely individual link flow (ILF) directed p-cycle, aggregated link flows (ALF) directed p-cycle, and loop-eliminating flow (LEF) directed p-cycle, respectively. These ILPs aim to jointly minimize power consumption and spectrum usage of all directed p-cycles configured in EONs. The problem formulation involves directed p-cycle generation, modulation format (MF) selection, power consumption optimization, and spectrum allocation. Furthermore, the proposed directed p-cycle strategy is designed with a compact and novel MF adaptation relying on accurate protection path lengths. Simulations are conducted to compare the proposed ILPs with the conventional method which uses a rough upper bound on MF adaptation. Numerical results demonstrate that all of the three proposed ILPs have better performances on the joint objective, in which the improvement is up to 24.31%. Although the proposed ILPs are with the same performance on the objective due to the same directed p-cycle strategy, the LEF directed p-cycle shows the best efficiency.

In Software Defined Networks (SDNs), a number of controllers are placed in a given data plane network. In a standard logically centralized control plane, each controller acts simultaneously as a primary controller for some switches and as a backup controller for other switches, and the controller placements must meet given switch-controller (SC) and controller-controller (CC) delay bounds. Then, the SDN should be resilient to network disruptions such as node-targeted attacks. To improve the SDN resilience to this kind of disruptions, we assume that some controllers are deployed only as backup controllers so that they take over the functions of primary controllers only in case of disruption. We propose an optimization model that solves a relevant primary and backup controller placement problem, where a minimum number of primary controllers minimizing the maximum SC delay is first established, and then a joint primary and backup controller placement maximizing the resilience of the SDN against a list of the most dangerous node-targeted attacks is determined. A numerical study illustrating the merits of the proposed optimization methodology is presented.

Abstract Ideally, the network should be dynamically reconfigured as traffic evolves. Yet, even within the software defined network paradigm, network reconfigurations cannot be too frequent due to a number of reasons related to route consistency, forwarding rules instantiation, individual flows dynamics, traffic monitoring overhead, and so on. In this paper, we focus on the fundamental issue of deciding whether, when, and how to reconfigure the network while traffic evolves. We consider a problem of optimizing semi‐stable routing in the capacitated multicommodity flow network when one may use at most a given maximum number of routing configurations (called routing clusters) and when each routing configuration must be used for at least a given minimum amount of time. We propose an efficient solution approach based on routing cluster generation that provides a tight lower bound on the minimum of a selected objective function (like maximum link delay or a sum of link delays) and suboptimal solutions very close to the calculated bound. The approach scales well with the size of the network.

We consider a network security and configuration management problem of locating service controllers so as to maximize availability of services in case of targeted attacks on the network infrastructure. Assuming that the attacker has full knowledge of the network topology but can only try to predict controller locations, we model the attacker's behavior introducing a set of probabilistic network availability measures and formulating an optimization problem model that determines the potentially most dangerous attacks the attacker might launch. We also formulate a counter-part optimization model that allows the network operator to derive the optimal placement of controllers, which maximizes availability of services with respect to a given set of network attacks. We explain the models and illustrate our considerations using a running small, intuitive network example. And we also perform extensive numerical experiments with a realistic network data to evaluate and compare the potential effectiveness of different attack strategies, and the effectiveness of the counter-measures that the network operator can adopt.

Abstract Flow thinning (FT) is a traffic protection mechanism for communication networks with variable link capacities, for example wireless networks. With FT, end‐to‐end traffic demands use dedicated logical tunnels, for example MPLS tunnels, whose nominal capacity is subject to thinning in order to follow fluctuations in link capacities availability. Moreover, instantaneous traffic of each demand is throttled at its originating node accordingly to the current total capacity available on the demand's dedicated tunnels so that the network is always capable of carrying the admitted traffic. In this paper, we deal with efficient, implementable versions of FT, referred to as affine FT (AFT) and quadratic FT (QFT). By deriving appropriate link availability state and path generation algorithms, we show how real‐life network dimensioning problems for AFT/QFT can be efficiently treated using a proper characterization of the network link availability states. Results of a numerical study illustrate tractability of the cost minimization problems, and assess efficiency of AFT/QFT as compared with other protection mechanisms.

Ideally, the network should be dynamically reconfigured as traffic evolves. Unfortunately, even in SDN paradigm, network reconfigurations cannot be too frequent due to a number of reasons related to route stability, forwarding rules instantiation, individual flows dynamics, traffic monitoring overhead, etc. In this paper, we focus on the fundamental problem of deciding whether, when, and how to reconfigure the network during traffic evolution. We consider a problem of optimizing semi-stable routing in the capacitated multicommodity flow network when one may use at most a given maximum number of routing configurations (called clusters) and when each routing configuration must be used for at least a given minimum amount of time. We propose a solution method based on cluster generation that provides a good lower bound on the minimum network delay (i.e., the total of link delays) and scales well with the size of the network.

The presented paper deals with wireless networks composed of FSO (free space optics) links supported by terrestrial optical fiber connections in order to increase resilience to adverse weather conditions. An FSO link realizes, at low cost, a broadband optical transmission system established by means of two parallel light beams connecting a pair of (remote) transceivers placed in the line of sight. However, a major disadvantage of FSO links (with respect to fiber links) is their sensitivity to weather conditions: bad weather affects optical channel, resulting in substantial degradation of the transmission power at the receivers, which in general leads to capacity degradation on multiple FSO links. When the transmission power received at the end node of an FSO link is decreased, the signal modulation and coding scheme (MCS) at the transmitter should be adjusted accordingly; in effect, an FSO link can be kept operational, but with decreased capacity, even if the channel is affected. Yet, since in severe weather conditions some FSO links may not be able to realize any reliable transmission at all (whichever MSC is used), the network may become disconnected and a portion of traffic lost. Therefore, it is reasonable to consider using fibers instead of free space light beams since weather insensitive (and high capacity) fibers installed on the key links will ensure network connectivity in all weather states. Certainly, the so obtained connectivity does not come for free because of the high cost of the optical fiber connections as compared to the FSO systems and hence the number of fibers should be kept at minimum. Thus, for designing hybrid FSO/fiber networks we need an optimization model for finding cheapest configurations of links (and their capacities) that will be able to carry the demanded traffic on acceptable level in all weather states foreseen in network operation. This is not a simple task as it requires not only robust network optimization methods but also a tractable way of characterizing possible weather states (which are not known in advance) and their influence on FSO link capacity. In the paper we present an approach to the so described task and illustrate its effectiveness by means of a numerical study.

Massive multiple-input multiple-output (MIMO) is expected to be a vital component in future 5G systems. As such, there is a need for new modeling in order to investigate the performance of massive MIMO not only at the physical layer but also higher up the networking stack. In this paper, we present general optimization models for massive MIMO, based on mixed-integer programming and compatible sets, with both maximum ratio combining and zero-forcing precoding schemes. We then apply our models to the case of joint device scheduling and power control for heterogeneous devices and traffic demands, in contrast to the existing power control schemes that consider only homogeneous users and saturated scenarios. Our results show that substantial benefits, in terms of energy usage, can be achieved without sacrificing throughput and that both the signaling overhead and the complexity of end devices can be reduced by abrogating the need for uplink power control through efficient scheduling.

Established approaches to data aggregation in wireless sensor networks (WSNs) do not cover the variety of new use cases developing with the advent of the Internet of Things (IoT). In particular, the current push toward fog computing, in which control, computation, and storage are moved to nodes close to the network edge, induces a need to collect data at multiple sinks, rather than the single sink typically considered in WSN aggregation algorithms. Moreover, for machine-to-machine communication scenarios, actuators subscribing to sensor measurements may also be present, in which case data should be not only aggregated and processed in-network but also disseminated to actuator nodes. In this paper, we present mixed-integer programming formulations and algorithms for the problem of energy-optimal routing and multiple-sink aggregation, as well as joint aggregation and dissemination, of sensor measurement data in IoT edge networks. We consider optimization of the network for both minimal total energy usage, and min-max per-node energy usage. We also provide a formulation and algorithm for throughput-optimal scheduling of transmissions under the physical interference model in the pure aggregation case. We have conducted a numerical study to compare the energy required for the two use cases, as well as the time to solve them, in generated network scenarios with varying topologies and between 10 and 40 nodes. Although aggregation only accounts for less than 15% of total energy usage in all cases tested, it provides substantial energy savings. Our results show more than 13 times greater energy usage for 40-node networks using direct, shortest-path flows from sensors to actuators, compared with our aggregation and dissemination solutions.

We jointly optimize Discontinuous Reception (DRX) cycle length and LTE scheduling to minimize mobile devices' energy usage for video delivery, utilising the now well-established potential to predict future channel conditions in cellular networks. Employing in-network caching, we set a strict buffer constraint which provides zero buffer underflow to improve Quality of Experience. Our study provides insight into the energy saving potential sophisticated DRX schemes hold, compared with the currently used static method. To this end, two novel DRX approaches are proposed and studied. The results show that more sophisticated DRX schemes (with variable DRX cycle length) can potentially save 69 percent energy for mobile devices, encouraging further research in the field.

We have previously presented Intent, a medium access control scheme for WLANs based on cooperative, distributed, non-binding frame scheduling. The main idea behind Intent is “intention sharing”-a mechanism that allows a node to be aware of transmissions previously scheduled by its neighbors. We now extend this scheme to networks with hidden nodes. We give a formulation of the scheduling problem and solve for the optimal solution using mixed-integer programming in a range of scenarios covering both mesh and infrastructure networks of varying density. We also provide an algorithmic solution and perform simulations using four variants of this, comparing the results with both the optimal solution and standard 802.11. Our algorithmic solution gives significantly better performance than 802.11 and comes within 1.1 times optimal performance when full, two-hop scheduling information is available. In addition, we show that it is not necessary to exchange complete scheduling information in order to perform distributed scheduling effectively but rather a conflict resolution or avoidance mechanism combined with non-binding schedules allows for similar performance with only a single round of information exchange between neighbouring nodes.

With the densification of nodes in cellular networks, free space optic (FSO) connections are becoming an appealing low cost and high rate alternative to copper and fiber backhaul solutions for wireless communication systems. To ensure a reliable cellular backhaul, provisions for redundant disjoint paths between the nodes must be made in the design phase. This paper aims at finding a cost-effective solution to upgrade the cellular backhaul with pre-deployed optical fibers using FSO links and mirror components. Since the quality of the FSO links depends on several factors, such as transmission distance, power, and weather conditions, we adopt an elaborate formulation to calculate link reliability. We present a novel integer linear programming model to approach optimal FSO backhaul design, guaranteeing $K$-disjoint paths connecting each node pair. Next, we derive a column generation method to a path-oriented mathematical formulation. Applying the method in a sequential manner enables high computational scalability. We use realistic scenarios to demonstrate that our approaches efficiently provide optimal or near-optimal solutions, and thereby allow for accurately dealing with the trade-off between cost and reliability.

The present study deals with Elastic Flow Rerouting (EFR)—an original traffic restoration strategy for protecting traffic flows in communication networks (including wireless networks) against multiple link failures. EFR aims at alleviating the trade‐off between practicability of traffic restoration and the cost of network resources observed in existing networking solutions. We present an extension of EFR capable of managing multiple partial link failures. We describe EFR and its extension, formulate the EFR related optimization problems, and discuss approaches for their resolution. We also discuss numerical results illustrating effectiveness of EFR in terms of the link capacity cost. © 2015 Wiley Periodicals, Inc. NETWORKS, Vol. 66(4), 267–281 2015

Multipath routing gives traffic demands an opportunity to use multiple paths through a network. In a single-demand situation, its benefits are easy to see. In a multi-commodity case, when potentially all node-pairs (demands) generate traffic, they compete for the same network resources. In this work, we consider multipath routing in communication networks in a multi-commodity setting from a traffic engineering perspective. Based on a result from linear programming, we show that at an optimal solution, the number of demands that can have multiple paths with nonzero flows is of the order of the number of network links for three commonly used traffic engineering objectives. We introduce a multipath measure (MPM) and show that under certain traffic conditions and topological structures, the MPM is zero or close to zero, i.e., Multipath routing provides little or limited gain compared to single-path routing. For the all-pair traffic case, multipath routing is observed to be advantageous for small networks. When the number of nodes is about 25 or higher and all node pairs have traffic, this advantage drops as the number of nodes in a network increases. For the fat-tree data center topology, the benefit of multipath routing also drops as the number of pods increases. Our findings are somewhat against a common belief (expressed by the term "load sharing") that multipath routing is significantly better in effective distribution of traffic over the network resources.

For some time, directional antennas have been considered to solve connectivity and interference issues in wireless networks. Several scenarios have been presented and often the conclusions drawn are positive, showing increase in capacity. However, to date there has been no effort to assess a holistic picture of the benefit/cost tradeoff and previous work mainly concerns either link scheduling or antenna placement but not the two combined. Such consideration will become increasingly important in the near future with the advent of heterogeneous networks and other possible combinations of mesh and public access networks.

Abstract In this article, we revisit a classical optimization problem occurring in designing survivable multicommodity flow networks. The problem, referred to as FR, assumes flow restoration that takes advantage of the so‐called stub release. As no compact linear programming (LP) formulation of FR is known and at the same time all known noncompact LP formulations of FR exhibit \documentclass{article}\usepackage{mathrsfs}\usepackage{amsmath}\pagestyle{empty}\begin{document}\begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard dual separation, the problem itself is believed to be \documentclass{article}\usepackage{mathrsfs}\usepackage{amsmath}\pagestyle{empty}\begin{document}\begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard, although without a proof. In this article, we study a restriction of FR (RFR) that assumes only elementary (cycle‐free) admissible paths—an important case virtually not considered in the literature. The two problems have the same noncompact LP formulations as they differ only in the definition of admissible paths: all paths (also those including cycles) are allowed in FR, while only elementary paths are allowed in RFR. Because of that, RFR is in general computationally more complex than FR. The purpose of this article, is three‐fold. First, the article reveals an interesting special case of RFR—the case with only one failing link—for which a natural noncompact LP formulation obtained by reducing the general RFR formulation still exhibits \documentclass{article}\usepackage{mathrsfs}\usepackage{amsmath}\pagestyle{empty}\begin{document}\begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard dual separation, but nevertheless this special case of RFR is polynomial. The constructed example of a polynomial multicommodity flow problem with difficult dual separation is of interest since, to our knowledge, no example of this kind has been known. In this article, we also examine a second special case of RFR, this time assuming two failing links instead of one, which turns out to be \documentclass{article}\usepackage{mathrsfs}\usepackage{amsmath}\pagestyle{empty}\begin{document}\begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard. This implies that problem RFR is \documentclass{article}\usepackage{mathrsfs}\usepackage{amsmath}\pagestyle{empty}\begin{document}\begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard in general (more precisely, for two or more failure states). This new result is the second contribution of the article. Finally, we discuss the complexity of FR in the light of our new findings, emphasizing the differences between RFR and FR. © 2013 Wiley Periodicals, Inc. NETWORKS, 2013

Abstract This survey deals with computational complexity of column generation problems arising in the design of survivable communication networks. Such problems are often modeled as linear programs based on noncompact multicommodity flow network formulations. These formulations involve an exponential number of path‐flow variables, and therefore require column generation to be solved to optimality. We consider several path‐based protection and restoration mechanisms and present results, both known and new, on the complexity of the corresponding column generation (also called pricing) problems. We discuss results for the case of single link or single node failures scenarios, and extend the considerations to multiple link failures. Further, we classify the design problems corresponding to different survivability mechanisms according to the structure of their pricing problem. Eventually, we show that almost all the encountered pricing problems are hard to solve for scenarios admitting multiple failures, while a great deal of them are \documentclass{article} \usepackage{mathrsfs} \usepackage{amsmath, amssymb} \pagestyle{empty} \begin{document} \begin{align*}\mathcal{NP}\end{align*} \end{document} ‐hard already for single failure scenarios. © 2011 Wiley Periodicals, Inc. NETWORKS, 2012

Abstract This article describes the Survivable Network Design Library (SNDlib), a data library for fixed telecommunication network design available at http://sndlib.zib.de . In the current version 1.0, the library contains data related to 22 networks which, combined with a set of selected planning parameters, leads to 830 network design problem instances. In this article, we discuss the data concepts of SNDlib and describe a mathematical model for each design problem considered in the library. We also provide information on characteristic features and the origin of the SNDlib problem instances. © 2009 Wiley Periodicals, Inc. NETWORKS, 2010

Abstract In this article we prove 𝒩𝒫‐hardness of two well‐known optimization problems related to the design of multicommodity flow networks with two different methods for providing network resiliency against failures: path diversity and flow restoration. Path diversity is a static mechanism that consists of using, for each demand, a number of paths and oversizing the flows assigned to these paths so that for any failure the total surviving flow is not less than the volume of the demand. By contrast, flow restoration is a dynamic mechanism that consists of reassigning the failed flows to backup paths when a failure occurs. Both mechanisms are of practical interest because although flow restoration is in general superior to path diversity in terms of the required amount of resource capacity, it might be too complicated to implement. By providing an appropriate reduction from the fractional graph coloring problem, we show that both problems are 𝒩𝒫‐hard in the general case of failure scenarios that admit simultaneous failures of multiple links. Finally, we discuss how to efficiently solve the two problems using path generation techniques. © 2009 Wiley Periodicals, Inc. NETWORKS, 2010

The paper introduces a peer-to-peer system called P2PRIV (peer-to-peer direct and anonymous distribution overlay). Basic novel features of P2PRIV are: (i) a peer-to-peer parallel content exchange architecture, and (ii) separation of the anonymization process from the transport function. These features allow a considerable saving of service time while preserving high degree of anonymity. In the paper we evaluate anonymity measures of P2PRIV (using a normalized entropy measurement model) as well as its traffic measures (including service time and network dynamics), and compare anonymity and traffic performance of P2PRIV with a well known system called CROWDS.

Our goal is to present a mathematical model useful for distributed optimization of traffic routing in multi-domain Internet networks. In such environments each domain is operated autonomously and has access to limited information about the rest of the network. Our model reflects this through an appropriate problem decomposition with respect to individual domains. The decomposition aims at supporting a distributed process of routing optimization that could be run in the control plane of the network using existing EGP routing protocols. The usefulness of the decomposition is investigated and illustrated with numerical examples.

For label switched networks, such as MPLS and GMPLS, most existing traffic engineering (TE) solutions work in a single routing domain. These solutions do not work when a route from the ingress node to the egress node leaves the routing area or the autonomous system (AS) of the ingress node. In such cases, the path computation problem becomes complicated because of the unavailability of the complete routing information throughout the network. We present CWS (computation while switching), a new inter-domain path computation scheme which tries to compute a near-optimal path without assuming the availability of complete topology information. We provide a detailed comparison of the CWS scheme with another per-domain path computation scheme given in J.-P. Vasseur et al. (2006). Unlike the standard per-domain path computation scheme (Vasseur et al., 2006), the CWS scheme continues the quest for a better path instead of terminating the search at the first available path, resulting a significant improvement in terms of path optimality. In particular, CWS guarantees that, for a given network state, a computed inter-domain path will traverse a minimum number of domains. This improvement in path computation directly impacts the amount of traffic that can be allowed on the network. For example, for the COST266 topology with 28 domains and 37 bidirectional inter-domain links, CWS places 960 of the requested 2000 paths as compared to 683 paths placed by existing schemes. Finally, the path setup latency of the CWS scheme remains comparable to that of existing schemes, by allowing the data flow as soon as the first feasible path is found.

Despite an observable trend in reducing the number of layers in core and metropolitan area networks, still the optimal design of multi-layer networks (like IP over WDM) remains an important issue as a means to reduce CAPEX and OPEX. In multi-layer networks, usually a connection oriented transport network (like WDM) is used to deploy a layout for optimal transport of traffic. In today's networks, the layouts are based on point to point LSPs. We will show here that by using a novel concept named bus-LSPs, we can achieve some significant savings for the operator, both in terms of CAPEX and OPEX. We also introduce a heuristic algorithm to design optimal layouts, and provide a quantitative evaluation

An important traffic engineering problem for OSPF networks is the determination of optimal link weights. Certainly, this depends on the traffic engineering objective. Regardless, often a variety of performance measures may be of interest to a network provider due to their impact on the network. In this paper, we consider different objectives and discuss how they impact the determination of the link weights and different performance measures. In particular, we propose a composite objective function; furthermore, we present a Lagrangian relaxation-based dual approach to determine the link weight system. We then consider different performance measures and discuss the effectiveness of different objectives through computational studies of a variety of network topologies. We find that our proposed composite objective function with Lagrangian relaxation-based dual approach is very effective in meeting different performance measures and is computationally very fast.

In the paper, we consider the problem of routing and bandwidth allocation in networks that support elastic traffic. We assume that the bandwidth demand between each source-destination (S-D) pair is specified in terms of a minimum and maximum value, and a set of flows between each S-D pair is allowed to realize these demands. (We say that a set of flows realizes the demand associated with an S-D pair, if the sum of the bandwidths allocated to these flows is greater than the minimum value assumed for the demand of that S-D pair). In this setting, we show that routing and bandwidth allocation can be formulated as an optimization problem, where network utilization is to be maximized under capacity and the widely used max-min fairness constraints. We describe three different algorithms to solve variants of this problem. The most important one, an efficient, original algorithm assuming multipath routing is studied in detail and illustrated with a numerical example.

We consider the problem of link capacity dimensioning and routing optimization in networks that support elastic flows and maintain proportional fairness among these flows. We assume that each demand between the origin-destination (O-D) pairs is associated with a minimum and a maximum bandwidth requirement and that a certain allocated bandwidth to a user demand (which must be between these minimum and maximum values) generates revenue for the network operator. On the other hand, the operator is incurred a capacity dependent cost for each link in the network. We then formulate the problem of bandwidth allocation, routing optimization and link capacity dimensioning as an optimization problem where the operator's objective is to maximize profit (under the fairness constraint). We propose computationally efficient algorithms to solve some important variants of this problem.

The paper addresses the IP/MPLS network cost optimization problem of selecting the localizations of nodes and links, combined with links' dimensioning. As MPLS is flexible in routing demands' flows through the network, the considered problem is similar to the classical NP-hard topological design problems dealt with in the past, mostly in the context of the road traffic. We assume that the network nodes are composed of two disjoint sets: the set of access nodes (label edge routers) and the set of transit nodes (label switching routers). The access (end) nodes only generate demands and do not transit end-to-end demands' flows, while the transit nodes do not generate demands but do transit the flows. The selection of the actually installed transit nodes and the links interconnecting (all) network nodes is the major subject of optimization. As the considered problem is hard, we discuss and propose branch-and-bound based solution methods. The effectiveness of the methods is illustrated by means of a numerical study.

The Warsaw metropolitan network is undergoing a radical architectural and technological transformation. The transformation, forced by the large unrealized demand for telephone stations, and structural traffic bottlenecks, is based on the installation of eight digital transit switches of high capacity. The so-formed upper transit layer gives an opportunity for the deployment of an efficient traffic routing system, for carrying the growing traffic well beyond the year 2000 without additional investments. A step-wise strategy is presented far the deployment of a modern traffic routing system in the Warsaw network, adapted to, and taking advantage of the fast transformation of the network. The choice of the strategy is based on a case study of the traffic handling efficiency of various routing systems in a model of the Warsaw metropolitan network.>