Network slicing is an important characteristic of 5G/6G networks that increases flexibility and enables different applications over a single infrastructure. The physical resources are partitioned to create virtualized networks, each dedicated to services with specific requirements. Several entities participate in network slicing, including Mobile Network Operators (MNOs), Mobile Virtual Network Operators (MVNOs), and users. An MNO owns the physical network infrastructure and the resources. MVNOs lease resources from the MNO and operate as service providers towards their subscribers. The goal of this work is to optimize the end-to-end network slicing process to provide services to users with a fair sharing of resources. We model this problem as a hierarchical combinatorial auction with a modified Vickrey-Clarke-Groves pricing mechanism. In the upper-level auction, an MNO is the seller supplying Network Slice to several MVNOs, who act as the bidders. In the lower-level auction, each MVNO holds an auction as a seller delivering services to their subscribed end-users, who play the role of bidders. We formulate and solve the Winner Determination Problem using mathematical programming and heuristic algorithms. The simulations show that the model can achieve fair sharing of resources, and it enables improving the MNO and MVNO revenue.

Network virtualization facilitates the deployment of diversified services and flexible resource management in elastic optical networks (EONs). However, due to the explosive growth of traffic, considerable energy consumption and spectrum usage have restricted the sustainable development of cloud services. This paper addresses joint energy and spectrum efficient problem for virtual optical network embedding (VONE) over EONs. We propose a heuristic algorithm to improve energy and spectrum efficiency while keeping a high acceptance rate. With consideration of factors influencing energy and spectrum efficiency, a feasible shortest path is preferred; meanwhile, an appropriate modulation format is dynamically selected according to transmission distance and the trade-off between energy and spectrum consumption. To improve the acceptance rate of virtual network requests, a dual mapping is employed to reinforce the embedding process by multi-dimensional resources integrated mapping. The simulation results show that the proposed algorithm can achieve a joint energy and spectrum efficiency with a much lower blocking probability compared with the baseline approach.

Traffic demand in fixed and mobile networks is increasing rapidly, driven especially by the growing adoption of Video-on-Demand (VoD) services, which are responsible for roughly 70% of today's Internet's traffic. Network operators must continuously explore new architectural solutions to satisfy increasing traffic at minimum cost. A promising solution consists in deploying caches at the network edge such that VoD requests can be terminated locally. The dimensioning of edge network nodes in terms of storage capacity as well as their placement in the network must be optimized, to reduce costs, improve quality of service, and utilize network resources efficiently. In this paper, we aim to find the optimal deployment of caches, which minimizes overall network resource occupation for VoD service, across the various levels of a hierarchical optical metro network, in terms of the number of caches, their location and dimension (i.e., storage capacity). We develop a discrete-event simulator for dynamic VoD provisioning to measure the performance of different cache deployment strategies in terms of overall network resource occupation and blocking probability. We prove that deploying all the available storage capacity in nearest cache locations does not guarantee the minimal resource occupation. In fact, to minimize resource occupation given a fixed budget in terms of storage capacity, storage capacity must be distributed strategically among caches at different layers of the metro network based on the characteristics of the service, e.g., VoD content catalog popularity distribution.

Highly fault-tolerant optical switch is becoming a critical network element for large wavelength division multiplexing mesh networks. Optical nodes built on replicated banyan network (RBN) structures are attractive for serving as optical switching systems due to the nice properties of banyan network structures. Furthermore it is expected that an optical RBN structure also has a good fault-tolerant capability because it consists of multiple identical banyan networks. Previous analytical models have been developed to understand the blocking behaviors of optical RBN networks without taking into account the link failure in optical RBN networks. In this paper, we conduct the link failure-tolerant analysis of optical RBN networks and present an analytical model for the blocking probability of optical RBN networks when link failures are taken into account. The model is significant because it provides network developers a quantitative guidance to determine the effects of network failure on the overall blocking behaviors of optical RBN networks and enables a desirable tradeoff to be made between blocking probability and hardware cost in a faulty optical RBN network.

For the incoming 5G mobile communications, enhanced cell coordination is deemed as crucial to maximize network throughput. In this paper we focus on the impact that the placement of coordination controllers in the metro optical network has on the performance of cell coordination, and we aim at verifying how throughput maximization can be achieved through optimized controller placement. Specifically, we refer to Coordinated Multipoint (CoMP), a protocol that allows to coordinate cell clusters through a Radio Controller Coordinator (RCC) to enhance the mobile-network throughput by reducing interference. Several CoMP techniques have been designed, whose throughput gain is affected differently by various factors. E.g., gains increases with the size of the cluster, while it decreases for larger latencies between RCC and cells. Current metro/aggregation networks are characterized by a hierarchical architecture with different levels of central offices, so the decision of where to place RCC in order to maximize CoMP throughput gain is not trivial and depends on several factors, such as geographical dimension of the network, cluster size and latency experienced before reaching the RCC. Our results show that under certain conditions optimized placement provides benefits with respect to fixed placement of controllers.

Telecom operators worldwide are witnessing squeezed profit margins mainly due to hyper‐competition. Hence, new business models/strategies are needed to help operators reduce Operational and Capital Expenditures. In this context, the Network Function Virtualization (NFV) paradigm, which consists of running Virtual Instances of Network Functions (NFs) in Commercial‐Off‐The‐Shelf (COTS) hardware, represents a solid alternative. Virtual Network Functions (VNFs) are then concatenated together in a sequential order to form service chains (SCs) that provide specific Internet services. In this article, we study different approaches to provision SCs with resiliency against single‐link and single‐node failures. We propose three Integer Linear Programming (ILP) models to jointly solve the problem of VNF placement and traffic routing, while guaranteeing resiliency against single‐link and/or single‐node failures. Specifically, we focus on the trade‐off between the conflicting objectives of meeting SCs latency requirements and consolidating as many as possible VNFs in NFV‐capable nodes. We show that providing resiliency against both single‐link and single‐node failures comes at twice the amount of resources in terms of NFV‐capable nodes, and that for latency‐critical services providing resiliency against single‐node failures comes at the same cost with respect to resiliency against single‐link and single‐node failures. Finally, we discuss important insights about the deployment of bandwidth‐intensive SCs. © 2017 Wiley Periodicals, Inc. NETWORKS, Vol. 70(4), 373–387 2017

Banyan networks serve as a class of important switching network architecture, whose multicast capability is critical for future high-performance switches to support multicast-intensive applications. The available literature indicates that, in order to construct optical multicast banyan networks based on the directional coupler (DC) technology, a high-hardware cost is usually involved to ensure the nonblocking property. In this paper, we conduct blocking probability analysis for such networks to explore the inherent tradeoff between blocking probability and hardware cost. In particular, we focus on a class of DC-based optical networks built on the replicated banyan network (RBN) architecture, and develop a theoretical upper bound on blocking probability for such networks under the general f-cast traffic, which covers the unicast and multicast as special cases. This bound captures the overall blocking behavior of f-cast optical RBN networks and agrees with the conditions of strictly nonblocking f-cast optical RBN networks. The proposed bound is significant because it provides a fundamental guideline to achieve the desirable tradeoff between blocking probability and hardware cost. This paper shows that the hardware cost of an f-cast optical RBN network can be dramatically reduced if a small blocking probability is allowed.

Data backup in data center networks (DCNs) is critical to minimize the data loss under disaster. This paper considers the cost-efficient data backup for DCNs against a disaster with ε early warning time. Given geo-distributed DCNs and such a ε-time early warning disaster, we investigate the issue of how to back up the data in DCN nodes under risk to other safe DCN nodes within the ε early warning time constraint, which is significant because it is an emergency scheme for data protection against a predictable disaster and also help DCN operators to build a complete backup scheme, i.e., regular backup and emergency backup. Specifically, an Integer Linear Program (ILP)-based theoretical framework is proposed to identify the optimal selections of backup DCN nodes and data transmission paths, such that the overall data backup cost is minimized. Extensive numerical results are also provided to illustrate the proposed framework for DCN data backup.

The high predictability in human mobility patterns is a valuable information to optimize resource allocation and increase energy efficiency by effectively adapting to expected traffic load variations. In this work, we addressed the impact of human mobility patterns of mobile users in a metro area over the aggregated traffic offered to the mobile metro-core networks. Exploiting predictable trends recognized in aggregated traffic, we propose two optimization methods to reduce the energy consumption in the optical layer of metro-core networks. Our results display energy savings of more than 20% by introducing load adaptive network operation.

Data center network (DCN) placement with the consideration of potential large-scale region failures is critical to ensure its functionality under such catastrophic scenarios. This paper considers the optimal DCN placement for DCN failure risk minimization against such region failures. We first apply a general probabilistic region failure model to determine the failure probability of a node/link, and then propose a general grid partition-based scheme to flexibly define the global distribution of a region failure in terms of its occurring probability and intensity. Such grid partition scheme also helps us to evaluate the vulnerability under a region failure. With the help of the vulnerability information, we further propose an Integer Linear Program-based theoretical framework to identify the optimal DCN placement to lead to the minimum DCN failure risk under a region failure. An example and extensive numerical results are also provided to illustrate the proposed framework for DCN placement.

Vertical stacking of banyan networks has been an attractive architecture to construct optical switching networks due to its small depth, absolute signal loss uniformity and good fault tolerance property. Recently, F.K. Hwang extended the study of banyan-based networks to the general f-cast case, which covers the unicast (f = 1) and multicast (f = N) as special cases. In this paper, we study the blocking probability of f-cast optical banyan networks under crosstalk-free constraint. It is expected that the proposed probability model can be used to dimension such an f-cast network and achieve a graceful tradeoff between hardware cost and blocking probability.

Energy efficiency is expected to be a key design parameter for next-generation access/aggregation networks. Using a single network infrastructure to aggregate/backhaul both mobile and fixed network traffic, typically referred to as fixed/mobile convergence (FMC), seems a promising strategy to pursue energy efficiency. WDM networks are a prominent candidate to support next-generation FMC network architectures, as they provide huge capacity at relatively low costs and energy consumption. We consider a FMC WDM aggregation network in which the novel concept of “hotelling” of mobile baseband units (BBUs) is employed. The so-called BBU hotelling consists in separating BBUs from their cell sites and consolidating them in single locations, called hotels. As a result, the aggregation network will transport both IP (fixed and mobile) traffic and CPRI (fronthaul) traffic, the latter exchanged between each BBU and its cell site. In this paper, we propose an energy-efficient WDM aggregation network, and we formally define the BBU placement optimization problem, whose objective is to minimize the defined aggregation infrastructure power (AIP). We consider three different network architectures: Bypass, Opaque, and No-Hotel, which feature different placement of BBUs and routing of traffic. By modeling the power contributions of each active device, we study how and how much BBU consolidation, optical bypass, and active traffic aggregation influence the AIP, for all the three architectures. Our results show that, in our case study, the proposed architectures enable savings up to about 60%-65% in dense-urban/urban and about 40% in rural scenarios.

Cloud Computing (CC) services are rapidly catching on as an alternative to conventional office-based computing. As cloud computing adoption increases, the energy consumption of the network and of the computing resources that underpin the cloud is growing and causing the emission of enormous quantities of CO2. Research is now focusing on novel "low-carbon" cloud-computing solutions. Renewable energy sources are emerging as a promising solution both to achieve drastic reduction in CO2emissions and to cope with the growing power requirements of data centers. These infrastructures can be located near renewable energy plants and data can be effectively transferred to these locations via reconfigurable optical networks, based on the principle that data can be moved more efficiently than electricity. This paper focuses on how to dynamically route on-demand optical circuits that are established to transfer energy-intensive data processing towards data centers powered with renewable energy. Our main contribution consists in devising two routing algorithms for connections supporting CC services, aimed at minimizing the CO2emissions of data centers by following the current availability of renewable energies (e.g., coming from sun and wind). The trade-off with energy consumption for the transport equipments is considered. We also compare three different IP-over-WDM network architectures. The results show that relevant reductions, up to about 30% in CO2emissions can be achieved using our approaches compared to baseline shortest-path-based routing strategies, paying off only a marginal increase in terms of network blocking probability.

This paper defines a procedure for designing the architecture of multistage switching networks. Our proposal is the result of a systematic study of the properties of multistage networks on one side and of optical switching systems on the other side. Modularity of the implementation is the primary concern, allowing the construction of a generic-size fabric by the simple cascading of multiple-stage modules. In this paper, we show in detail the application of the approach to a family of banyan networks, although the method can be extended to other types of networks. The designed architecture can be exploited for various implementation technologies, for instance, integrated optics and free-space optics. The method can even be applied to electronic VLSI/ULSI circuits. After presenting the design procedure, we discuss practical implementations in different technology scenarios. The main contribution of this paper is to simplify the building of large switches by cascading a number of switching modules that combine the internal stage architecture with the interconnection between stages.

Data center networks (DCNs) generally adopt Clos network with crossbar middle switches to achieve non-blocking data switching among the servers, and the number of middle switches is proportional to the number of ports of the aggregation switches in a fixed manner. Besides, reconfiguration overhead of the switches is generally ignored, which may contradict the engineering practice. In this paper, we consider batch scheduling based packet switching in DCNs with reconfiguration overhead at each middle switch, which inevitably leads to packet delay. With existing state-of-the-art traffic matrix decomposition algorithms, we can generate a set of permutations, each of which stands for the configuration of a middle switch. By reconfiguring each middle switch to fulfill multiple configurations in parallel with others, we reveal that a tradeoff exists between packet delay and switch cost (denoted by the number of middle switches), while performance guaranteed switching with bounded packet delay can be achieved without any packet loss. Based on the tradeoff, we can minimize the number of middle switches (under a given packet delay bound) and an overall cost metric (by translating delay into a comparable cost factor), as well as formulating criterions for choosing a matrix decomposition algorithm. This provides a flexible way to reduce the number of middle switches by slightly enlarging the packet delay bound.

Interconnection systems inside switching equipment, high-performance computers and data-centers are nowadays facing more and more demanding requirements. Optical interconnections based on multistage switching networks provide more bandwidth and less energy consumption compared to electronic counterparts. In this work a procedure to design the architecture of optical multistage switching networks is proposed which exploits the properties of multistage networks, on one side, and of optical switching systems on the other side. Thanks to the modularity of the architecture, a generic-size fabric can be implemented by simply cascading multiple stage-modules. In this paper we show in details the application of the approach to the Extended Generalized Shuffle (EGS) networks, though the method can be extended to other types of networks. The proposed procedure supports various implementation technologies, as, for example, integrated optics with micro-ring resonators, free- space optics with 2-D MEMS, networks on chip.

Future switching and interconnection fabrics inside switching equipment, high-performance computers and datacenters will require more throughput and more energy efficiency. Optical technology provides many opportunities of improvement of both features compared to electronic counterparts. This work defines a procedure to design the architecture of optical multistage switching networks. Modularity of the implementation is the primary concern, allowing for the construction of a genericsize fabric by the simple cascading of multiple stage-modules. In this paper we show in details the application of the approach to a family of banyan networks. The designed architecture can be exploited for various implementation technologies, as, for instance, integrated optics with micro-ring resonators, free-space optics with 2-D MEMS, networks on chip.

The continuous traffic growth keeps challenging operators' backbone networks, which nowadays are typically implemented through Wavelength Division Multiplexing (WDM) technology. Operators are always looking for new technical solutions to efficiently exploit the bandwidth provided by WDM fiber links, while avoiding undesired increases in network cost and energy consumption. The Time Driven Switching (TDS) technique promises to effectively addresses this issue by enabling the switching of “fractions” of wavelengths thanks to the time-coordination of all network elements. By doing so, connection requests can be efficiently “packed” into WDM links without necessarily converting optical signals into the electronic domain, i.e., traffic grooming can be accomplished directly in the optical domain. In this paper we propose new dynamic routing and scheduling strategies for TDS networks and investigate the performance of a TDS network in terms of blocking probability Pb. Due to the complexity of the joint routing and scheduling, we solve the problem by a two-step approach. Three different routing strategies are evaluated and we observe that Pbis substantially reduced (up to five orders of magnitude lower) when allowing connections to be routed over different wavelengths or different physical paths. We also note that using optical buffers at most reduces Pb10 times. Moreover we study the variation of the time-frame size switched by TDS nodes, which influences Pbup to two orders of magnitude.

The communication network is now one of the critical infrastructures in our society. However, the current communication networks are facing more and more large-scale region failure threats, such as natural disasters (e.g. earthquake, tornado) and physical attacks (e.g. dragging anchors or EMP attack). Therefore, a deep understanding of network behaviors under region failure is essential for the design and maintenance of future highly survivable networks. In this paper, we focus on the network vulnerability assessment under the geographically correlated region failure(s) caused by a random “line-segment” cut, an important region failure model that can efficiently capture the behaviors of some region failures like earthquake, tornado and anchor cutting. To facilitate such vulnerability assessment, we apply the geometrical probability theory to design a grid partition-based estimation scheme for Disrupted Link Capacity, Pairwise Traffic Reduction and Pairwise Disconnection Probability, three commonly used metrics for statistical vulnerability assessment. A theoretical framework is also established to determine a suitable grid partition such that a specified estimation error requirement is satisfied.

Energy consumption in telecommunication networks keeps growing rapidly, mainly due to emergence of new Cloud Computing (CC) services that need to be supported by large data centers that consume a huge amount of energy and, in turn, cause the emission of enormous quantity of CO2. Given the decreasing availability of fossil fuels and the raising concern about global warming, research is now focusing on novel “low-carbon” telecom solutions. E.g., based on today telecom technologies, data centers can be located near renewable energy plants and data can then be effectively transferred to these locations via reconfigurable optical networks, based on the principle that data can be moved more efficiently than electricity. This paper focuses on how to dynamically route on-demand optical circuits that are established to transfer energy-intensive data processing towards data centers powered with renewable energy. Our main contribution consists in devising two routing algorithms for connections supporting CC services, aimed at minimizing the CO2emissions of data centers by following the current availability of renewable energy (Sun and Wind). The trade-off with energy consumption for the transport equipments is also considered. The results show that relevant reductions, up to about 30% in CO2emissions can be achieved using our approaches compared to baseline shortest-path-based routing strategies, paying off only a marginal increase in terms of network blocking probability.

Today's “green thrust” has been completely steering the evolution of our society. Therefore, since enhancements on telecommunication networks are expected to cope with the growth of traffic demand, energy consumption is becoming a crucial design metric. In this paper we focus on IP over Wavelength Division Multiplexing (IPoWDM) optical network architectures, where optical circuits (or lightpaths) interconnect IP routers. Various implementations of IPoWDM networks are possible, e.g., transparent, translucent or opaque each accounting different routing and switching constraints, in order to evolve from the classic IPoWDM approach, where only signal transmission is performed optically. Although it is a common belief that “more optical” architectures can generally guarantee a much lower total power consumption, we show that identifying the most energy-efficient IPoWDM architecture is not a trivial task. In fact we demonstrate that the most energy-efficient architecture depends on some network parameters such as number of nodes, link length and average distance between nodes. This work provide an analytical framework by which we draw guidelines to (a) assess under which conditions optical switching actually becomes more desirable than electronic switching and (b) evaluate which architecture, among those employing optical switching, is more energy-efficient under different parameter settings.

In this letter we focus on the construction of optical buffers exploiting a static priority scheduling policy. Optical Switches and fiber Delay Lines (SDLs) are used to build the proposed multiple input/output ports Optical Priority Multiplexers and Priority Switching Nodes, used for contention-resolution in Optical Packet Switched networks.

All-optical router has been considered as a natural choice to keep pace with growing fiber link capacity. One main research issue of all-optical router is the design of optical queues with the same flexibility as their electronic counterparts, and some recent works have proved the feasibility of using optical switches and fiber delay lines (SDL) to emulate the electronic queues. In this paper, we focus on the SDL-based construction of N -to- N shared optical queue, a more efficient queue structure in comparison with the dedicated input and output queues. The construction we consider consists of a crossbar switch of size (N + M) × (N + M), where N inputs (outputs) are reserved for external arrivals (departures), and M fiber delay lines are connected from the remaining M outputs back to the remaining M inputs. We first show that by setting the length riof fiber delay line i as ri= 1 + [(1 - 1) mod N], i = 1, . . . M, and scheduling packets properly among these delay lines, such a construction can work as a non- idling first in first out (FIFO) shared queue of size B = Σi=1Mri. We further extend our work to the design of more general shared buffer, where the packets can be stored for an arbitrary time and may depart in a non-FIFO order.

Future interconnection subsystems for switches and routers must overcome physical limitations of current electronic backplanes in order to achieve aggregate bandwidth much greater than today. Driven by this consideration we study the scalability of switching architectures implementing backplanes based on optical interconnection technology. In particular, this paper deals with the issue of interconnecting the line cards of a switch/router exploiting the wavelength agility of the transmitters and the switching/filtering capabilities of Micro Ring Resonators. To accomplish this task, the resonators deployed in the backplane are both tunable and fixed, thus defining a hybrid architecture. In our analysis we address feasibility and scalability issues in terms of maximum number of interconnected cards, taking various design parameters into account. Moreover, we perform a preliminary power consumption analysis of the proposed backplane architecture.

We present a novel architecture for strictly non-blocking multistage photonic switches implemented using tunable wavelength converters and arrayed waveguide gratings. Clos networks, studied mostly in electronic domain, are required when the physical switching needs exceed the capacity of the largest feasible single crossbar switch. Unfortunately, straightforward extensions of these networks to the photonic domain show that the switch size has to be severely limited by the coherent crosstalk in each of the AWGs. We describe how recursive Clos networks free of coherent crosstalk can be built taking into account the current physical limitations of arrayed waveguide gratings.

The lack of optical buffer is still one of the main problems that hinder the development of all optical networks. One approach to this problem is to emulate the behavior of optical buffers by using optical switches and fiber delay lines (SDL). Current works on this topic have demonstrated the feasibility of constructing SDL-based First In First Out (FIFO) buffer, Priority buffer, etc. The Last In First Out (LIFO) buffer is another important network component for congestion control and QoS guarantee, and parallel and cascade architectures have been peoposed for the efficient design of such optical buffer. The recent work in showed that it is possible to use M fiber delay lines (FDLs) to construct a LIFO buffer of size B = (3/2) · 2M/2- 1 and B = 2(M+1)/2- 1 when M is even and odd, respectively. In this paper, we improve the work in [3] by providing a more efficient construction of SDL-based optical LIFO buffer. We first show that with a single stage feedback structure consisted of one (M + 1) × (M + 1) crossbar switch and M FDLs connecting M outputs of the crossbar back to M its inputs, we are able to construct a LIFO buffer of size B = 2 · 2M/2- 2 and B = (3/2) · 2(M+1)/2- 2 when M is even and odd, respectively. This is achieved through adopting a properly delay length setting for each FDL and a careful packets scheduling among FDLs, as well as exploiting the nice function of simultaneous packet reading and witting a FDL can support. We further show that if we adopt a cascade of smaller switches rather than a single (M+1)×(M+1) big switch, the new LIFO design can be implemented with much lower complexity in terms of the total number of basic 2 × 2 switch elements.

The mission critical network infrastructures are facing potential large region threats, both intentional (like EMP attack, bomb explosion) and natural (like earthquake, flooding). The available research on region failure related vulnerability studies generally adopt a kind of simple “deterministic” region failure models, which can not capture some important features of real region failure scenarios, where a network component in the region only fails with certain probability, and more importantly, such failure probability tends to vary with both its dimension and its distance to failure center. In this paper, we provide a more general “probabilistic” region failure model to capture the key features of a region failure and apply it for the network vulnerability assessment. To facilitate such assessment, we adopt a grid partition-based scheme to estimate various statistical network metrics under a random region failure. A theoretical framework is also established to determine a suitable grid partition such that a specified estimation error requirement is satisfied. The grid partition technique is also useful for identifying the vulnerable zones of a network, which can guide network designers to initiate proper network protection against such failures. The work in this paper helps us more deeply understand the network vulnerability behavior under region failure and facilitates the design and maintenance of future highly survivable mission critical networks.

Reducing the Internet power consumption will become a challenging issue, since the Internet is expected to face a high growth in terms of traffic requirements. Simply scaling the network architecture, thus increasing its energy consumption, proportionally to this growth would not be a practical solution. Various energy-efficient approaches have been considered, typically consisting in a dual-layer network architecture through the introduction of an all-optical transport network layer, as relevant power savings can be reached thanks to optical technologies. Optical switching especially allows to significantly reduce the quantity of high power-requiring optical/electronic/optical conversions and electronic processing operations. A further improvement can be obtained with Time Driven Switching (TDS), a technique which allows to switch "fractions" of wavelengths directly in the optical domain exploiting the time-coordination of all network components. In this paper we show the benefits, in terms of power saving, provided by the TDS transport architecture, by performing a comprehensive overview of the different core network architectures and comparing the power consumption obtained in the different cases. With such an optical transport technology, power savings of more than 40% are demonstrated with respect to existing architectures.

This paper studies the optical shuffle-exchange interconnection network based on Arrayed Waveguide Gratings (AWGs). It proposes implementations of the network suitable to be exploited in an optical backplane that interconnects line-cards of a high-performance switch or router. A procedure to design AWG interconnection stages with zero coherent crosstalk is presented and applied to the perfect-shuffle pattern, the building-block of the shuffle-exchange. The scalability of the proposed architecture is analyzed with a simple approach.

The lack of optical buffer is still one of the main problems that hinder the development of all optical networks. Current works on this topic mainly focus on the emulation of optical buffers based on a combination of fiber delay line (FDL) and switch. These works have demonstrated the feasibility of FDL-based emulation for many kinds of optical buffers, like the First In First Out (FIFO) buffer, Priority buffer, etc. The Last In First Out (LIFO) buffer is another basic network component for congestion control and QoS guarantee. Recently, Huang et al. introduced a recursive construction for the LIFO buffer, which requires no less than 9 log2B FDLs to build such a buffer of size B. In this paper, we first show that by a proper FDL grouping and a suitable FDL length assignment for each FDL-group, we need approximately 3 log2B FDLs to emulate a LIFO buffer of size B. We then demonstrate that if a careful packet scheduling among FDL-groups is adopted, this number of FDLs can be further reduced to 2 log2B.

Abstract—It is generally admitted that interdomain peering links represent nowadays the main bottleneck of the Internet, particularly because of lack of coordination between providers, which use independent and “selfish ” routing policies. We are interested in identifying possible “light ” coordination strategies that would allow carriers to better control their peering links while preserving their independence and respective interests. We propose a robust multipath routing coordination framework for peering carriers, which relies on the multiple-exit discriminator (MED) attribute of Border Gateway Protocol (BGP) as signaling medium. Our scheme relies on a game theory modeling, with a non-cooperative potential game considering both routing and congestions costs. Peering equilibrium multipath (PEMP) coordination policies can be implemented by selecting Pareto-superior Nash equilibria at each carrier. We compare different PEMP policies to BGP Multipath schemes by emulating a realistic peering scenario. Our results show that the routing cost can be decreased by roughly 10 % with PEMP. We also show that the stability of routes can be significantly improved and that congestion can be practically avoided on the peering links. Finally, we discuss practical implementation aspects and extend the model to multiple players highlighting the possible incentives for the resulting extended peering framework. Index Terms—Border Gateway Protocol (BGP), game theory, interdomain routing, multiple-exit discriminator (MED), multipath, peering. I.

This paper analyzes the loss process distribution of a finite buffer queue. In contrast to the previous work that assumed the buffer can merely store finite number of packets, our model adopts the bounded delay policy where only the packet arrival finding its delay not exceeding a preset value is admitted into the buffer. The quantity of interest is the probability distribution of the number of lost packets within a block of n consecutive packet arrivals, which is an important measure for the design of communication networks, e.g., the forward error correction (FEC). We derive a set of recursive equations to compute the above quantity for various packet size distributions. We then focus on the influence of adding redundant packets on loss probability of message block and FEC efficiency. The impacts of bounded delay, packet size distribution and traffic load are also evaluated. We demonstrate a unique property of the finite queue with bounded delay, which is different from that of the conventional finite queue (e.g., M/G/1/K queue).

Segment Protection (SP) is an efficient scheme for protection in WDM optical networks. This letter provides algebraic formulations to evaluate SP availability both in the dedicated and shared backup case. The availability models are applied to numerical examples and significant relations between SP availability and relevant connection parameters are identified.

Protection techniques for optical networks mainly rely on preallocated backup bandwidth, which may not be able to provide full protection guarantee when multiple failures occur in a network. To protect against multiple concurrent potential failures and to utilize the available resources more efficiently, strategies such as backup reprovisioning (BR) rearrange backups of protected connections after one failure occurs or, more generally, whenever the network state changes, e.g., when a new request arrives or terminates. Recently, new solutions for automatized management in optical networks promise to allow customers to specify on-demand the terms of the service level agreement (SLA) to be guaranteed by the service provider. In this paper, we study different backup reprovisioning techniques able to further reduce the capacity requirements exploiting the knowledge, among the other service level specifications (SLSs), of the connection holding time.

In this paper we provide a mathematical ILP model for the Regeneration Placement Problem (RPP) which minimizes the total number of regeneration nodes allocated in a translucent optical network ensuring that all the node pairs can always reach one another via two link-disjoint lightpaths under physical-impairment constraints. Since RPP is NP-complete, large-site design problem can not be solved relying upon exact approaches. We then propose a game-theoretic approach to model RPP as a non-cooperative game and solve it applying the best response dynamic concept. Finally, we evaluate the performance of the proposed approach in terms of closeness of the obtained results to these provided by ILP: a MILP formulation is given in order to study the quality of the Nash equilibria by comparison to Price-of-Anarchy and Price-of-Stability bounds.

Recently, new solutions for automatized management in optical networks promise to allow customers to specify on-demand the terms of the service level agreement (SLA) to be guaranteed by the service provider. In this paper we show that is possible to design a highly efficient load balancing algorithm, called RABBIT, for the dynamic provisioning of connections exploiting the knowledge, among the other service level specifications (SLS), of the connections duration. The core idea of RABBIT consists in routing connections based on the transient probability of future-link congestion, that can be estimated with higher precision when the knowledge of connections durations is given. So, we introduce a time-dependent link-weight assignment that evaluates future link congestions probability based on the transient analysis of the Markovian model of the link, making it computationally feasible by means of an effective approximation technique. By means of an extensive set of simulative experiments, we compare our approach to other traditional holding-time agnostic, yet efficient, dynamic routing algorithms. We consider different performance metrics, among which the blocking probability (BP), in a wavelength-convertible WDM mesh network scenario. For a typical US nationwide network, RABBIT obtains savings on BP of up to 20% for practical scenarios.

It is generally admitted that Inter-domain peering links represent nowadays the main bottleneck of the Internet, particularly because of lack of coordination between providers, which use independent and "selfish" routing policies. We are interested in identifying possible "light" coordination strategies, that would allow carriers to better control their peering links, while preserving their independence and respective interests. We propose a robust multi-path routing coordination framework for peering carriers, which relies on the MED attribute of BGP as signalling medium. Our scheme relies on a game theoretic modelling, with a non-cooperative potential game considering both routing and congestions costs. Peering Equilibrium MultiPath (PEMP) coordination policies can be implemented by selecting Pareto-superior Nash equilibria at each carrier. We compare different PEMP policies to BGP Multipath schemes by emulating a realistic peering scenario. Our results show that the routing cost can be decreased by roughly 10% with PEMP. We also show that the stability of routes can be significantly improved and that congestion can be practically avoided on the peering links.

In this paper we devise an highly-efficient load balancing algorithm, called LB-HTA, for the dynamic provisioning of connections with known duration in WDM networks. We introduce a time-dependent link-weight assignment that captures future congestion of links, leveraging the knowledge of connection durations. By means of an extensive set of simulative experiments, we compare our approach to other traditional, yet holding-time agnostic, dynamic routing algorithms. For a typical US nationwide network, LB-HTA obtains significant saving in blocking probability for practical scenarios. Moreover, we address two key-questions regarding holding-time-aware dynamic routing. Three main traffic models are considered here, namely i) dynamic traffic, ii) dynamic traffic with known durations and iii) scheduled traffic: how much the knowledge of connection durations improves the performance with respect of a holding- time-agnostic solution? Is the obtained solution close to the most effective solution provided by scheduled traffic? In order to exhaustively evaluate the performance of LB-HTA, we consider as benchmark the solution obtained under the well-known Scheduled Traffic model (TI-ST) and also under an approximated, but more effective, approach for traffic scheduling, called Time- Variant Scheduled Traffic (TV-ST). For both TI-ST and TV-ST, Integer Linear Program (ILP) formulations are proposed and results compared with dynamic routing algorithms.

Most of the previous works on optical burst switching (OBS) assume in their analysis that signalling does not affect network performance. It is analysed here, under which conditions the effect of signalling is actually negligible, taking into account the effect of signalling in the evaluation of burst discard probability. First, analytical models for two different signalling approaches in an OBS network are presented: ‘out-of-band’ and ‘in-band’ techniques. The impact of these two signalling strategies in terms of the probability of burst discard are evaluated, identifying the component of bursts discarded as a consequence of control message losses or of excessive signalling delay. A new method is also discussed, based on the previous models, to assign the correct amount of resources to the control plane. To verify the accuracy of the analytical results, these are compared with results based on discrete-event simulationns: results are found to be in a highly satisfactory agreement with simulations.

Optical buffering is fundamental to contention resolution in optical networks. The current works on this line mainly focus on the emulation of dedicated input/output buffer queue by using switched fiber delay lines (SDL). It is notable that the shared buffer queue, where a common buffer pool is shared by all the input/output ports of a switch, has the potential to significantly reduce the overall buffer capacity requirement. As far as we know, however, no related work is available yet on the exact emulation of a shared optical buffer queue with SDLs. In this paper, we focus on the design of first in first out (FIFO) shared optical buffer queue based on the optical feedback SDL construction. The construction considered consists of an (M + 2) x (M + 2) switch fabric and M fiber delay lines FDL1, . . ., FDLM, where FDLiconnects the ithoutput of the switch fabric with its ithinput. We show that by setting the length of FDL1as min(M + 1 - i, i), i = 1, . . . , M, such a construction can actually work as an 1-to-2 shared buffer queue. We then extend this emulation to the more general N-to-2 case.

In a wavelength-division-multiplexing (WDM) optical network, backup reprovisioning (BR) provides a means to improve survivability (by protecting against multiple concurrent potential failures) and to utilize the available resources more efficiently. In particular, global backup reprovisioning (GBR), i.e. the reprovisioning involving the totality of the backup capacity has been demonstrated to be almost as capacity-effective as other re-optimization techniques involving primary paths. Unfortunately, existing ILP and heuristic solutions for the GBR problem are very time consuming and unsuitable to be adopted in time-sensitive network scenarios (e.g., subject to frequent network-status changes and/or connection-provisioning setup- time constraints). In this paper, we first analytically evaluate how the GBR computational time may impact on the connection provisioning process. Then, we propose a fast BR procedure based on simulated annealing (SA) which significantly decreases the computational time with respect to existing heuristics while still achieving very high resource efficiency.

In this paper, we study the problem of inter-domain AS tree selection for multipoint tunnel set-up within an alliance of ASs. We first describe the framework of our work, based on the introduction of a service plane for automatic multi-domain service provisioning. We introduce an abstract representation of domain relationship by means of directional metrics which are applied to a triplet (ingress point, transit AS, egress point) where the ingress and egress points can be ASs or routers. Then, we focus on the multipoint AS Selection problem that arises in such an architecture. The corresponding constrained Steiner problem is known to be a hard problem, and the introduction of directional metrics increases its complexity. We propose an original approach that allows one to reach almost optimal solutions with tractable computation times. Besides its performance, one contribution of this paper is that some steps of the proposed heuristic can be precomputed, independently of the tunnel demands. By extensive tests on random topologies derived from the Internet, we show that our heuristic is often equal or a few percent close to the optimal, and that, in the case of precomputation, its time consumption can be much lower than other well-known algorithms.

This paper analyzes possible Banyan network solutions as a multistage interconnection network exploiting time and space domains. The well-known approach to replicate a Banyan in space provides non- blocking property, meanwhile increasing the cost index to a great extent. In order to keep the cost function under control we investigate how a Banyan network can be built that maintains the blocking probability below acceptable thresholds. To this extent we first discuss the non-blocking conditions for multistage networks in the space domain with arbitrary size of the basic switching element. Then we evaluate how exploiting time division multiplexing in such a network can help relieving the cost issue while keeping the blocking probability under control. The study compares all-optical vs. electronic implementations and analyzes the cost index. The traffic performance results of Banyan with different sizes of the switching element, with replications and TDM are presented.

Protection techniques for optical networks mainly rely on pre-allocated backup bandwidth, which may not be able to provide full protection guarantee when multiple failures occur in a network. To protect against multiple concurrent potential failures and to utilize the available resources more efficiently, strategies such as backup reprovisioning rearrange backups of protected connections after one failure occurs or, more generally, whenever the network state changes, e.g., when a new request arrives or terminates. Recently, new solutions for automatized management in Optical networks promise to allow customers to specify on-demand the terms of the service level agreement (SLA) to be guaranteed by the service provider. In this paper we show that is possible to further reduce the capacity requirements of backup reprovisioning techniques exploiting the knowledge, among the other service level specifications (SLS), of the connection holding-time. Our main contributions are as follows. First, we prove that the problem of backup reprovisioning for all the lightpaths requiring shared-path protection under a current network state is NP-complete. Second, we provide a mathematical ILP model of the reprovisioning problem considering the additional holding-time information. Third, since the problem is NP-complete and we can not efficiently rely on exact approaches, a new global reprovisioning algorithm called Ph-GBR is proposed which can significantly reduce the resource overbuild exploiting the information about connection durations. By means of simulative experiments, we compare capacity requirement and computational complexity of Ph-GBR to those of another holding-time unaware, yet efficient algorithm, called GBR, considering a dynamic traffic in a wavelength-convertible WDM mesh network scenario.

Progress in network technologies and protocols is paving the road towards flexible optical transport networks, in which dynamic leasable circuits could be set up and released on a short-term basis according to customers requirements. Recently, new solutions for automated network management promise to allow customers to dinamically specify the terms of the Service Level Agreement (SLA) to be guaranteed by the service provider. Since this new information is made available, we propose to exploit the knowledge of connection holding time, among the other Service Level Specifications (SLS), to improve the routing efficiency. In this work, we consider that a typical electronic-layer (e.g., SDH or MPLS) demand requires only a fraction of the capacity of the single wavelength bandwidth and we investigate a new algorithm for traffic grooming of sub-wavelength connections in an optical mesh network. We rely on the knowledge of the holding time of connection requests to exploit lightpath capacity and hence to achieve significant reduction in blocking probability for the traffic grooming problem. Our new methodology is applied on a typical US nation-wide network and results are compared with those given by previous known approaches.

Multi-log2N networks (or vertically stacked banyan networks) have been an attractive class of switching networks due to their small depth O(log N), absolute signal loss uniformity and good fault tolerance property. Recently, F.K.Hwang extended the study of multi-log2N networks to the general f-cast case, which covers the unicast case (f = 1) and multicast case (f = N) as special cases, and determined the conditions for these networks to be f-cast strictly nonblocking when the fan-out capability is available at both the input stage and middle banyan stage. In this paper, we study the rearrangeable f-cast multilog2N networks under both node-blocking scenario (relevant to photonic switches) and link-blocking scenario (relevant to electronic switches). In particular, we consider the following three fan-out cases in our study: (1) no restriction on fan-out capability; (2) input stage has no fan-out capability; (3) middle banyan stage has no fan-out capability. We determine the necessary conditions for the first two cases while obtaining the necessary and also sufficient condition for the third one.

The fast deployment of broadband networks and the evolution of communication services has revealed the importance of a network-wide availability of multicast communications, that is from one source to multiple destinations. This paper considers three-stage switching networks able to support multicast traffic, i.e. connections in which one inlet of a node is connected to more than one outlet of the node at the same time. An analytical model is developed here that evaluates the blocking probability of multicast connections taking into account statistically the correlation between occupancy events in links belonging to different interstage patterns. Such new model proves to be more accurate than any previous model addressing the same issue. Interestingly enough, unlike other models the new formulation is consistent with the theoretical conditions of a non-blocking three-stage switching network.

In this paper we propose an edge-to-edge closed-loop scheme to provide Quality-of-Service (QoS) in Optical Burst Switching (OBS) networks. Performance parameters for each burst flow are exchanged by a feedback message called Service Control Message (SCM) which is processed only at the edge/ingress nodes. According to the information contained in the SCM the edge nodes could adapt the values of some flow parameters (typically burst assembly parameters and offset time) to apply differentiated quality of service to the various flows. Simulations results are presented to investigate the feasibility of this new proposal.

In this paper the new control algorithm for log2(N, 1, p) switching fabrics is proposed. This algorithm is the extension of the algorithm we propose for log2(N, 1, p) switching fabrics. The idea of the algorithm is to choose such a plane for the new connection in which this new connection will block the fewest number of possible future connections. Wide-sense nonblocking conditions are derived and proved when this algorithm is used for connection set up in such kind of switching fabrics with odd or even number of stages. The number of planes needed is the same as for the rearrangeable switching fabrics when the number of stages is even, and greater than for the rearrangeable switching fabric but lower than for the strict-sense nonblocking fabric, when the number of stages is odd.

The differentiated services (DiffServ) architecture is a scalable solution for supporting quality of service (QoS) in the Internet. The scheduling discipline plays a fundamental role in QoS capable network nodes. The weighted fair queuing (WFQ), has shown many advantages and widely spread in the last years. In this work we present a mechanism for a dynamic tuning of the WFQ scheduler weights based on measurements of the incoming traffic. The purpose is to obtain proportional maximum delay guarantees for the traffic classes. Our proposal is reliable and efficient in supporting a proportional relative model of QoS. Collected results show how a WFQ-based router can effectively guarantee the proportionality between the maximum delays of DiffServ classes at a given interface in realistic scenarios.

Progress in network technologies and protocols is paving the road towards flexible optical transport networks, in which leasable circuits could be set up and released on a short-term basis. Thus we consider it reasonable that, at least for some types of services, the holding time of connection requests can be known in advance. In this paper, we propose to exploit the knowledge of connection-holding time to improve the performance of an algorithm for shared-segment protection (SSP). For a typical US nationwide network, we compared our approach to an holding-time-unaware, but otherwise shared segmented efficient, approach, obtaining savings on resource overbuild of up to 7% for practical scenarios.

The WONDER project aims at designing and building a WDM packet-switched slotted ring network, with tunable transmitters and fixed receivers. An experimental prototype is under development at the laboratories of Politecnico di Torino. In this paper, we study two different strategies of slot synchronization for such networks: in the former, a slot synchronization signal is transmitted by the master station on a dedicated control wavelength; in the latter, slave nodes achieve slot synchronization aligning on data packets received from the master. Emphasis has been given to the architecture of the WONDER network. The performance of the two synchronization strategies, in terms of packet collision probability, has been evaluated by simulation. The technique based on transmitting a timing signal on a dedicated control wavelength resulted in achieving a better performance, although being more expensive due to need of an additional wavelength. However, the technique based on aligning on data packets received from the master should be not discarded, although attaining lower timing stability, especially if limiting the number of wavelengths and receivers is a strong requirement.

Since the demand for high-capacity in optical WDM networks keeps increasing, the realization of cheap and simple Optical Cross-Connects (OXCs) becomes more and more important. In this paper we propose new solutions for the construction of WDM optical switches employing only arrayed waveguide gratings and wavelength converters. Our approach defines single-stage OXC architectures and derives their nonblocking conditions, by considering two different connection request models for the lightpaths to be set through the switch. Depending on the degree of lightpath addressability, the OXC blocking is referred to either the wavelength-to-wavelength model or the fiber-to-fiber model. Further architectural considerations are added to identify different non-blocking solutions employing multiple switching/conversion stages.

This paper deals with the design of an optical packet switch architecture adopting an arrayed waveguide grating device for packet routing. Due to the high cost of optical components, the node structure proposed here is equipped with a pool of tunable wavelength converters fully shared among all input channels. An analytical model is developed that allows us to evaluate the packet loss performance of a bufferless version of such structure. The model accuracy is assessed by means of computer simulation.

In order to support the continuous growth of transmission capacity demand, optical packet switching technology is emerging as a strong candidate, promising to allow fast dynamic allocation of WDM channels, combined with a high degree of statistical resource sharing. This paper investigates the performance of optical packet-switching architectures employing input and/or shared buffer, based on an arrayed waveguide gratings device and tunable wavelength converters to route packets. Due to the high cost of optical components we analyze the switch performance when a pool of wavelength converters is fully shared among all input channels

This paper deals with optical packet switching in a full-IP transport network scenario. Given the technological limits of accomplishing packet buffering in the optical domain, deflection routing is here explored as an alternative technique for resolving packet contentions without buffering packets. Two different network topologies have been considered here, that is a regular six-node network with different connectivity factors and the classical NSF network. The performance improvements that can be obtained by deflection routing have been evaluated considering different methods for choosing the alternative paths where to deflect packets that cannot be transmitted onto the shortest: path to the addressed destination.

A design technique for reliable optical transport networks is presented. The network is first dimensioned in order to carry a given set of static protected optical connections, each one routed maximizing its availability. The network can be further optimized by minimizing the number of fibers to be installed, while keeping a control on connection availability, which can remain the same or decrease by a prefixed margin factor. Design and optimization algorithms are provided for networks adopting dedicated and shared path-protection. The optimization approach is heuristic. Results obtained by applying the proposed technique to two case-study networks are shown and discussed. These two case-study experiments are carried out exploiting a realistic model to evaluate terrestrial and submarine optical link availability.

This article considers three-stage switching networks able to support multicast traffic, that is, connections in which one inlet is connected to more than one output at the same time. The nonblocking conditions for this network are studied under the assumption of absence of any optimized routing of the connections inside the structure (the so-called strict-sense nonblocking networks). The theoretical nonblocking condition for such network under point-to-point traffic is the well known Clos condition. We give here the necessary and sufficient conditions for such network to be strict-sense nonblocking under multicast traffic. © 2005 Wiley Periodicals, Inc. NETWORKS, Vol. 46(4), 163–170 2005

In paper deals with optical packet switching in a full-IP transport network scenario. For the switching of IP packet flows different node architectures are considered that are based on current optical routing devices. The traffic performance of a mesh network is evaluated with the various node structures, assuming that nodes employ either shortest path routing or deflection routing to forward packets to the addressed destinations. The paper shows how the different node structures behave in terms of packet loss probability with different network configurations when the node parameters are varied.

The paper presents and compares various unicast nonblocking architectures to be used into space-domain photonic switching networks. All the analyzed architectures have been evaluated and compared considering a possible physical implementation based on guided-wave structures realized with integrated optics technology. Some properties including number of switching elements required, blocking performance, number of waveguide crossovers, system attenuation, and signal-to-noise ratio are evaluated and analyzed. The main purpose of this work is to review the state-of-the-art of optical guided-wave space-switching architectures and to provide a relevant set of technical elements useful in the selection of architectures to be used in all-optical cross-connect implementation.

This work further studies blocking issues that are related to time driven switching (TDS). TDS is a novel and promising technology for the realization of low complexity and high scalability switches in the electrical and optical switching domains. TDS is unique in enabling the implementation of all-optical fractional wavelength switching with current state-of-the-art components. Furthermore, TDS provides deterministic QoS guarantees for all streaming media applications. Being a new technology, TDS requires investigation, especially regarding its performance and efficiency. The performance in this paper considers multiple scenarios with variety of optical transmission and switch parameters. The performance results show that blocking does not compromise the efficiency in most cases, even with elevated offered traffic load with the restrictive scheduling of immediate forwarding.

The deep changes undergone by network traffic have strongly impacted its statistical characterization. The traditional assumptions on traffic characteristics have been proven to be obsolete and a new formalism has been brought about to properly take into account traffic new properties. Accordingly, newly designed analysis tools have been proposed. Our work is focused on the presentation and the test of these tools of analysis, believing that we have first of all to be confident on the methods used to properly characterize network traffic.

A new approach is introduced in this paper to make possible a flexible utilization of WDM networks using current technology. It is shown that the bandwidth made available end-to-end by a single wavelength can be simply broken up into smaller fractions by relying on worldwide free synchronization systems, such as GPS, previously deployed for different applications. It is shown how this new approach, called time driven switching, behaves in terms of call blocking when the basic parameters of the scheme are varied.

This work is based on the optical switch architecture proposed in [M.M Renaud, et al., April 1997] [SBregni, G Guerra, and A Pattavina, Nov. 2002] to handle variable-length packets such as IP datagrams. This optical switch is based on an AWG device to route packets and is equipped with a fiber delay-line stage as optical input buffer. Unfortunately, this switch would require considerable buffering capability to achieve acceptable performance. A possible solution, studied in this paper, is to implement efficient packet deflection inside the optical network as a mean for solving packet contentions on outputs of optical switches. Thus, optical transport networks have been simulated to assess deflection effectiveness, based on a traffic model adherent to real IP traffic measurement. Full-mesh and wheel network topologies have been considered, comparing results to assess deflection effectiveness.

This paper considers three-stage switching networks able to support multicast traffic, i.e. connections in which one inlet is connected to more than one output at the same time. The nonblocking conditions for this network are studied under the assumption of absence of any optimized routing of the connections inside the structure (the so-called strict-sense nonblocking networks). An analytical model is developed here that provides not only the nonblocking conditions of three-stage multicast networks, but also the evaluation of the blocking probability when such conditions are not satisfied. Unlike previous well-known approaches, our model takes into account the correlation between occupancy events in links belonging to different interstage patterns. The results being found also provide a more stringent condition of network nonblocking for multicast traffic which disproves some of the claimed results recently published in the technical literature.

In order to support the continuous growth of transmission capacity demand, optical packet switching technology is emerging as a strong candidate, promising to allow fast dynamic allocation of wavelength-division multiplexing (channels, combined with a high degree of statistical resource sharing). This work addresses the design of optical switch architectures, based on previous proposals available in the technical literature that use an arrayed waveguide grating (AWG) device to route packets. Since the port number of currently available AWGs is a limiting factor, we propose two new modified structures which better exploit the switching capability of this component in the wavelength domain. Since a limited hardware complexity is a key requirement for all-optical switches, due to the high cost of optical components, these different node configurations are compared in terms of complexity. Traffic performance of these new structures in a full optical packet switching scenario is also examined.

In order to support efficiently the fast-growing demand for transmission capacity, optical packet-switched systems seem to be strong candidates as they allow fast dynamic allocation of WDM channels combined with a high degree of statistical resource sharing. We propose the architecture of an input/shared-buffered optical packet switching node, for all-optical switching of IP traffic flows. Since a limited hardware complexity is a key requirement for all-optical switches, due to the high cost of optical components, different node configurations are compared in order to evaluate the effect of a hardware-complexity variation on the performance of the proposed architecture.

This paper presents a workable solution for transport of IP packets over optical networks. The protocol heavily relies on the data link and physical layer of the proposed standard 10 Gbit/s Ethernet. The architecture is able to optically multiplex and forward asynchronous packets like those generated by classical Ethernet devices.

We present a heuristic method to optimize the capacity of a WDM mesh network assigning working and protection capacity to a set of static optical-connection requests under the link-disjoint constraint. The method can be applied to networks supporting either dedicated or shared path-protection. Routing and fiber and wavelength assignment (RFWA) are jointly performed for all the lightpaths with the aim of minimizing the number of fibers in the network. The employment of multiple prioritized heuristic criteria and an efficient link disjoint paths search algorithm are proposed. With less computational effort, the proposed heuristic approach allows one to obtain good suboptimal results compared to the exact integer-linear-programming optimization. After introducing our design approach, we discuss the optimization of particular case-study networks under various conditions and compare the results.

Efficient planning and optimization of wavelength division multiplexing networks is an important issue today. Integer linear programming (ILP) is the most used exact method to perform this task. We propose a new ILP formulation that allows to solve optimization with less computational effort compared to other ILP approaches. This formulation applies to multifiber mesh networks with or without wavelength conversion, when either the total fiber number or the total fiber length is the cost function to be minimized. After presenting the formulation we discuss the results we obtained by exploiting it in the optimization of two case-study networks.

This paper considers three-stage switching networks for which nonblocking conditions with point-to-point traffic are given by the well known Clos (1953) theorem, under the assumption of absence of any optimized routing of the connections inside the network. We give the conditions for such a network to be strict-sense nonblocking under multicast traffic, by showing also that previously published papers, although claiming the same result, only provided sufficient conditions.

In this paper, four heuristic algorithms for the optimal allocation of limited optical-layer, resources in WDM networks (virtual-topology optimization by routing and wavelength assignment) under static traffic demand are compared. While some previous papers assumed that all connection requests must be necessarily satisfied, in this work algorithms are studied aiming at comparing their ability in minimizing both the usage of optical-layer resources and the number of connection requests that cannot be allocated. Thus, this paper reports the results obtained on a 13-nodes simplified topology of the pan-European optical transport network designed in the COST 239 Project. The results of a few hundreds simulations with different traffic matrices have been gathered, by considering separately the cost of resources allocated and the percentage of rejected connections. One algorithm resulted best performing, because in most tests it allocated the lowest cost of optical-layer resources. On the other hand, the four algorithms exhibited comparable performance with respect to the percentage of rejected connections.

Optical cross-connects (OXCs) are gaining an increasing importance in high-capacity wavelength-division multiplexing networks. In this work we propose a theoretical approach for the analysis and design of these optical-switching systems. Our method is based on a space-switching equivalent representation which allows one to easily classify a general OXC architecture, identify its blocking properties and understand the trade-off between space and wavelength switching-domains. We explain this method and demonstrate its application to some of the best known OXC implementations presented so far in the literature.

The development of optical cross-connect architectures is a very important topic today. We consider here in particular the class of optical space-division switching fabrics configured as multistage structures built with 2/spl times/2 optical switching elements (SEs) and derived from a combination of vertical replication and horizontal expansion of Banyan networks. We determine the necessary and sufficient conditions for these matrices to be rearrangeably nonblocking and free of first-order crosstalk in SEs. This impairment is one of the major limitations in optical cross-connect performance. We focus on rearrangeable matrices since they have lower complexity than their strict-sense nonblocking counterparts. Given the current high cost of optical SEs, the rearrangeable solution looks attractive today.

The future growth in the demand of transmission bandwidth makes the development of all-optical transport networks and optical cross-connects very important. We consider rearrangeable non-blocking switching fabrics configured in a multistage structure with very simple switching elements. One of the problems of these networks is the crosstalk: a wavelength approach is used here showing that two wavelengths suitably assigned are sufficient to solve any blocking and crosstalk problems. We compare this type of networks, called rearrangeable networks with zero non-filterable crosstalk, with other networks, showing that they give the best performance when employed to implement optical-cross-connects.

The advantages of employing passive optical architectures in the access network have been largely recognized. Particularly, developments in optical technologies have made the realization of wavelength division multiplexing passive optical networks (WDM-PONs) feasible and cost-effective. These networks are more future-proof than conventional PONs, thanks to their intrinsic optical transparency and their extremely high transmission capacity. A very useful optical routing device, called a waveguide grating router, is the basic building-block of new PON architectures capable of connecting a large number of users or to improve the use of the optical bandwidth. In this paper the feasibility of these architectures is discussed by considering the costs and the technological limitations on the optical components.

Dynamic wavelength assignment in a WDM all-optical network with sparse wavelength conversion is discussed in this paper. The performance of unidirectional and bidirectional ring networks in terms of blocking probability, conversion gain and fairness are evaluated under various types of offered traffic patterns. Smooth and peaked patterns have been considered in addition to the classical Poisson traffic type. It is shown how the use of wavelength converters affects the blocking performance of connections and, more importantly in terms of network fairness, its dependence on the source-destination distance.

The substantial growth expected in the near future in the demand of transmission bandwidth to be used in a dynamic and flexible transport network makes the development of all-optical digital crossconnect architectures very important. We consider here in particular the class of the rearrangeable non-blocking space-division switching fabrics configured as multistage structures built with very simple optical switching elements. Rearrangeable networks look today more attractive than strict-sense non-blocking networks since the former have a lower complexity and are compatible with the data loss rate required in a circuit-switched network even with current technology of optical switching devices. We face one the most important problems which arises in a multiple space-channel optical system such as a photonic switching fabric, that is the build-up of crosstalk noise on a certain channel due to the interference with other signals inside the system. We find here the design rules to configure a minimum-cost rearrangeable photonic switching network where the most serious cause of crosstalk (interference inside the switching elements) is eliminated.

The support to ATM services in wireless mobile networks is discussed. A reference model for the network architecture is outlined and a quantitative analysis is presented, in which the efficiency of resource management mechanisms and quality of service support in a wireless ATM network is compared to that of a circuit switched network. The analysis involves mobile users belonging to two different service classes (real-time and non-real-time traffic). This study has been carried out by considering the ATM layer to receive an ideal service from the underlying radio-channel access layer. The results show that ATM resource management improves the blocking performance of the network whenever a consistent portion of the total offered traffic is represented by variable bit rate applications; the QoS is largely improved for non-real-time variable bit rate traffic and is slightly worsened for real-time constant bit rate traffic.

The strict-sense non-blocking conditions of an interconnection network built by vertical replication or a Banyan network with additional stages are known only it the expanded Banyan network has a recursive construction. We consider a shuffle/exchange network as the basic building block to be replicated vertically. Since this network does not have a recursive construction, the previous analysis cannot be applied. By relying on the topological properties of the channel graph in a shuffle/exchange network, we find that the non-blocking conditions for this network an numerically the same needed for a recursive Banyan network.

It is well known that a multistage banyan network, which is a single-path blocking structure, becomes rearrangeable nonblocking in a circuit-switching environment if the number of its stages is increased so as to obtain a Benes network. Banyan networks, provided with a shared queue in each switching element, have often been proposed as the core of an interconnection network for an ATM packet switching environment. In this scenario, if the classical interstage backpressure protocols are used, adding stages to a banyan network can even degrade the banyan network performance, in spite of the multipath capability given by the additional stages. A class of new simple interstage protocols is here defined to operate in the added stages of the banyan network so that a sort of sharing of the queueing capability in each added stage is accomplished. Large improvements in the traffic performance of these extended banyan networks are obtained, especially in the region of offered loads providing a low packet loss probability.

We study a scheme that allows a smooth increase of the capacity of a cellular system for circuit-switching by applying cell-partitioning and dynamic channel allocation techniques. A bound is computed for this reuse partitioning scheme that gives the maximum theoretical gain accomplished in the system bandwidth. The performance of the proposed scheme, in terms of blocking probability, is evaluated both when the position of the mobiles remains unchanged and when mobility is taken into account. The numerical results show that the capacity of the proposed scheme is higher than that of a fixed allocation scheme.

Shuffleout is a blocking multistage asynchronous transfer mode (ATM) switch using shortest path routing with deflection, in which output queues are connected to all the stages. This paper describes a model for the performance evaluation of the shuffleout switch under arbitrary nonuniform traffic patterns. The analytical model that has been developed computes the load distribution on each interstage link by properly taking into account the switch inlet on which the packet has been received and the switch outlet the packet is addressing. Such a model allows the computation not only of the average load per stage but also its distribution over the different links belonging to the interstage pattern for each switch input/output pair. Different classes of nonuniform traffic patterns have been identified and for each of them the traffic performance of the switch is evaluated by thus emphasizing the evaluation of the network unfairness.

ATM switch architectures based on deflection routing are examined and compared with regards to their internal operations. Their common feature is the availability of multiple I/O paths through a multistage unbuffered interconnection network where conflicts for the same internal link are dealt with, stage by stage, by deflecting the packets onto the wrong path. The main engineering parameter of the architecture, that is the number of network stages that provides a given packet loss performance, is studied. In particular it is found that basically all the examined architectures have a complexity on the order of Nlog/sub 2/N in the range of switch sizes of usual interest. Furthermore it has been possible to rank the architectures with comparable complexity based on the loss performance they provide.

It is well known that a multistage banyan network, which is a single-path blocking structure, becomes rearrangeable non-blocking in a circuit switching environment if the number of its stages is increased so as to obtain a Benes network. Banyan networks, provided with a shared queue in each switching element, have often been proposed as the core of an interconnection network for an ATM packet switching environment. In this scenario, if the classical interstage back-pressure protocols are used, adding stages to a banyan network can even degrade the banyan network performance, in spite of the multipath capability given by the additional stages. A class of new simple interstage protocols is defined to operate in the added stages of the banyan network so that a sort of sharing of the queueing capability in each added stage is accomplished. Large improvements in the traffic performance of these extended banyan networks are obtained, especially in the region of offered loads providing a low packet loss probability.>

A new class of switching architectures for broadband packet networks, called shuffleout, is described and analyzed in the paper. Shuffleout is basically an output-queued architecture with a multistage interconnection network built out of unbuffered b/spl times/2b switching elements. Its structure is such that the number of cells that can be concurrently switched from the inlets to each output queue equals the number of stages in the interconnection network. The switching element operates the cell self-routing adopting a shortest path algorithm which, in case of conflict for interstage links, is coupled with deflection routing. The paper presents the basic shuffleout architecture, called open-loop shuffleout, in which the cells that cross the whole interconnection network without entering the addressed output queues are lost. The key target of the proposed architecture is coupling the implementation feasibility of a self-routing switch with the desirable traffic performance typical of output queueing.>

A new class of switching architectures for broadband packet networks, called shuffleout, is described and analyzed. Shuffleout is basically an output-queued architecture with a multistage interconnection network built out of unbuffered b/spl times/2b switching elements. Its structure is such that the number of cells that can be concurrently switched from the inlets to each output queue equals the number of stages in the interconnection network. The switching element operates the cell self-routing adopting a shortest path algorithm which, in case of conflict for interstage links, is coupled with deflection routing. The basic version of this architecture is called open-loop shuffleout. This paper describes the closed-loop shuffleout architecture with 2/spl times/4 switching elements in which cells that have crossed the whole interconnection network re-enter the network as long as they are not successfully routed to the addressed switch outlet. This result is accomplished by adding to the basic open-loop structure recirculation paths so that each packet can cross several times the interconnection network. Two different solutions are proposed to implement such functionality, the buffered closed-loop shuffleout and the expanded closed-loop shuffleout architecture. Both these solutions aim at reducing the number of stages in the network, compared to the open-loop structure, so as to reduce the complexity of the switch internal wiring and to simplify the output queue interface.>

The paper develops the analysis of multistage banyan interconnection networks in which the switching elements are provided with a buffer shared among all the inlets and outlets of the element. The packet transfer within the network takes place according to absence or presence of backpressure signals between adjacent stages. In this latter case four different modes for operating backpressure have been studied: local and global backpressure with acknowledgment or grant backward signaling. The paper describes three models for the analysis of these networks when loaded by a random traffic. These models are based on an increasing degree of accuracy (and hence of complexity) in the representation of the state of the generic switching elements. The accuracy of these models in evaluating the network performance is assessed in the paper also in comparison with the results given by previously proposed models.>

For pt.I see ibid., vol.2, no.4, p.398-410 (1994). The paper develops the analysis of multistage banyan interconnection networks in which the switching elements are provided with a buffer shared among all the inlets and outlets of the element. Two different internal protocols are considered for the transfer of packets from stage to stage based on the presence or absence of interstage backpressure to signal the occurrence of buffer saturation conditions. The analytical model is based on the representation of the switching element state by means of only two variables. Two kinds of offered traffic patterns are considered, a bursty balanced pattern and an unbalanced uncorrelated pattern. In this latter case, the load above the average is supposed to be addressed either to a single network outlet, or to a given group of network outlets.>

The performance of multistage banyan networks with a shared buffer in each switching element is studied. Two different internal protocols are considered for the transfer of packets from stage to stage based on the presence or absence of interstage backpressure signals to signal the occurrence of buffer saturation conditions. As far as the offered traffic is concerned, two kinds of traffic patterns have been studied, a burst and an unbalanced traffic pattern. For both of them an analytical model is developed based on the iterative solution of state equations in order to compute the main traffic performance of the network.>

A switch model for ATM networks is analyzed. Its interconnection network is internally nonblocking and is provided with dedicated input and output queues, one per switch inlet and one per switch outlet. The switch operates with an internal speed-up: more than one packet per slot can be transferred from the head-of-line positions of the input queues to each output queue by the interconnection network. Two different operation modes are considered for the interaction between input and output queues: backpressure mode and queue loss mode. The analytical model developed for the evaluation of the switch performance under random traffic assumes an infinite size for the switch, arbitrary values for input and output queue size, as well as for the speed-up factor. Switch throughput, packet delay and loss performance are evaluated and the analytical model accuracy is assessed using computer simulation results.>

The authors describe a model for the performance evaluation of the Shuffleout asynchronous transfer mode (ATM) switch under arbitrary nonuniform traffic patterns. Shuffleout is a blocking multistage structure using shortest path routing with deflection, in which output queues are connected to all the stages. The analytical model developed computes the load on each interstage link by tracing all the paths that a packet can follow to reach the addressed switch outlet, that is, by taking into account all the deflections from its shortest path it can receive. Such a model allows the computation not only of the average load per stage but also its distribution over the different links belonging to the interstage pattern. Thus fairness in packet switching by Shuffleout is provided by the model.>

The performance evaluation of multistage self-routing interconnection networks for asynchronous transfer mode (ATM) switching is carried out by means of an analytical approach. The switching elements in the interconnection network are provided with a buffer shared among all the inlets and outlets of the element and no backpressure signals are exchanged between adjacent stages. Two analytical models are considered: the scalar model and the vectorial model. In the former case the addresses of the packets in the same buffer are assumed to be mutually independent, whereas in the latter case the model keeps memory of the addresses of the packets stored in the buffer across consecutive time slots. The vectorial model provides performance measures much more accurate than the scalar model.>

A new class of asynchronous transfer mode (ATM) switching architectures for broadband packet networks, called Shuffleout, is introduced. Its interconnection network is a multistage structure built out of unbuffered 2*4 switching elements. Shuffleout is basically an output-queued architecture in which the number of cells that can be concurrently switched from the inlets to each output queue equals the number of stages in the interconnection network. The switching element operates the cell self-routing adopting a shortest path algorithm which, in case of conflict for interstage links, is coupled with deflection routing. The specific architecture presented is called Closed-Loop Shuffleout, since the cells that cross the whole interconnection network without entering the addressed output queues are lost. The key target of the proposed architecture is coupling the implementation feasibility of a self-routing switch with the desirable traffic performance typical of output queueing.>

A Batcher-banyan interconnection network that overcomes the limitation of switch throughput and unfair allocation of bandwidth is proposed, and its performance is evaluated. The throughput limitations of the interconnection network are substantially reduced by defining pools of outputs, in which groups of outputs that offer the same service behave each as a set of servers sharing a single waiting list. Through a careful system design and a suitable priority scheme, the servers in a set can be evenly allocated to the users requesting service from the inputs of the interconnection network. Three different solutions are proposed for the allocation of the output channels within a pool to those users requesting the same pool. Extensive computer simulation was used to evaluate the overall packet delay performance and the degree of fairness provided by each of these solutions.>

The author gives some qualitative performance targets to be fulfilled for the service classes proposed by CCITT for the future broadband-ISDN (B-ISDN) and proposes a nonblocking, self-routing asynchronous transfer mode (ATM) switching architecture that is able to fulfil the different performance figures of each class. To exploit the service integration accomplished by ATM switches, the switching bandwidth is allocated at call level and cell level. This allocation gives the flexibility of letting lower-priority services use the reserved bandwidth left temporarily unused by higher-priority services. The architecture adopts mixed input-output queuing. Input queuing is particularly suited to the definition of internal frame structures, making it possible to guarantee the absence of cell loss due to congestion for specific services (such as circuit emulation). Output queuing makes it possible to implement in hardware a switching speedup that practically removes the performance degradation due to the head-of-line blocking phenomenon typical of input queuing.>

A broadband packet switch is described whose interconnection network is based on a Batcher-banyan structure. The switch, which is able to transfer up to M packets per slot to a given switch outlet, is provided with packet buffers both at input and output ports. A probe-ack contention resolution scheme is adopted to select the conflict-free paths for the packets to be transferred through the interconnection network. This scheme at the same time limits the access to the interconnection network to a number of packets such that the buffer overflow at the output queues never takes place. Restricting the packet loss so that it occurs only at input queues simplifies the control procedures and makes it possible to jointly optimize the packet storage capability at input and output queues.>

An input queued packet switch for broadband networks that supports data link connections and whose capacity is not constrained by the capacity of a single packet channel is described. The switch bandwidth is allocated at two different times. At connection setup time, an output data link (a channel group) is selected on which some bandwidth is reserved for the new call. At transmission time, specific channels within the data link are optimally allocated, slot by slot, to the packets addressed to that link, solving the output channel contention at the same time. The interconnection network of the switching is basically a Batcher-banyan network including the additional networks required to handle channel groups and to allocate bandwidth fairly at transmission time. Packet sequence integrity can be guaranteed for jitter-sensitive services, as the emulation of a circuit connection. It is shown that the channel group capability of the switch greatly improves the performance of standard input queued packet switches, so that input queuing with suitable data link capacities and output queueing give comparable performances.>

The authors examine the optimal choice of the slot size in terms of bandwidth usage in slotted broadband communication networks based on self-routing switching fabrics and statistical multiplexing schemes. This is a difficult problem, because of the uncertainty of the kinds and of mix of services to be supported by the network. To make this problem less critical, a packet multiplexing scheme is proposed, allowing several packets to be multiplexed in the same slot on a digital link. This scheme aims at avoiding bandwidth inefficiency when the information unit is smaller than the user capacity of a slot. This can take place for transactional data traffic and for packet voice traffic when the voice packet size underutilizes the slot capacity. The performance of a concentrator implementing this feature loaded by a data traffic is evaluated, providing upper bounds on the throughput and evaluating packet delay and buffer occupancy for different input levels.>

The problem of bandwidth allocation in a packet switch supporting broadband services is addressed. To reduce the performance constraints imposed by limiting a data link to a single broadband packet channel, the author introduces the concept of channel group as a set of broadband packet channels that is viewed as a single data-link connection by routing entities. He uses a two-step bandwidth allocation scheme. At connection setup time, a call is allocated to a channel group. At transmission time, specific channels of a group are optimally allocated to the packets destined to the group. Because of the statistical smoothing of the large number of sources served by a channel group, the traffic performance of the switch is improved. This scheme also allows super-rate switching, i.e., the support of services with peak bandwidth exceeding the capacity of a single packet channel. The author shows the feasibility of this scheme in a Batcher-banyan switch, by implementing in hardware the bandwidth allocation at transmission time. Performance improvements obtained by this scheme are also provided in different traffic environments.>