Today's Data Centers networks depend on optical switching to overcome the scalability limitations of traditional architectures. All optical networks most often use slotted Time Division Multiple Access (TDMA) operation; their buffers are located at the optical network edges and their organization relies on effective scheduling of the TDMA frames to achieve efficient sharing of the network resources and a collision-free network operation. Scheduling decisions have to be taken in real time, a process that becomes computationally demanding as the network size increases. Accelerators provide a solution and the present paper proposes a scheduler accelerator to accommodate a data center network divided into points of delivery (pods) of racks and exploiting hybrid electro-optical top-of-rack (ToR) switches that access an all-optical inter-rack network. The scheduler accelerator is a parallel scalable architecture with application specific processing engines. Case studies of 2, 4, 8, 16 processors configuration are presented for the processing of all the transfer TDMA time slot requests for the cases of 512 and 1024 ToR network nodes. The architecture is realized on a Xilinx VC707 board to validate the results.

In the Internet of Things (IoT) era, a vast number of (smart) end devices forward their traffic to the Internet, either by direct communication to LTE networks, or by multihop transmissions to a specific gateway. Acquiring both types of communication capabilities for IoT end devices would be unnecessarily costly. Instead, for a cost effective IoT infrastructure, only devices performing as gateways could be fully equipped with such capabilities, while the remaining devices could have simple low cost wireless transceivers, of differing transmission specifications, to forward/relay the traffic toward a gateway. Furthermore, the IoT devices network should comply with specific quality of service (QoS) requirements, specified for each IoT device. In this context, the IoT network planning problem, where we have to select the number of gateways and their locations along with respective transceivers for IoT devices, is key to provide a low cost and QoS aware IoT infrastructure. We formulate the planning problem as an integer linear program (ILP) that minimizes the total cost of the devices deployed in the network, while achieving the mandatory QoS requirements. We also present a heuristic algorithm of lower complexity that was observed to provide solutions near the optimal ones, in scenarios that we tracked optimal solutions with the ILP. A variety of performance evaluation results exhibits the effectiveness of the proposed algorithms in terms of network cost and efficiency.

We propose a scheme to reduce monitoring uncertainties in optical networks. The proposed scheme uses monitoring data of optical connections (lightpaths) which can be obtained from coherent optical receivers that can also function as optical performance monitors (OPM). We exploit both space and time correlation of the monitoring data in order to reduce the monitoring uncertainties. The improved accuracy can result in various benefits, the most common one is that the Quality of Transmission (QoT) can be estimated with higher accuracy which can in turn lead to more optimized decisions and lower provisioning costs. In this paper we present another application, we show how to use the obtained accurate monitoring data to localize soft failures (also referred to as QoT problems) on a per link level.

In the Internet of Things (IoT) era, a large number of data packets are exchanged among devices over the Internet. IoT (smart) end devices forward their traffic to the Internet, either by direct communication to LTE networks, or by multi-hop transmissions to a specific gateway. Acquiring both types of communication capabilities would be unnecessarily costly for IoT end devices. Instead, to lower the overall cost, only devices performing as gateways could be fully equipped with such capabilities, while the rest of the devices could have simple low cost wireless transmitters for forwarding the traffic towards a gateway. Towards this end, the decision on gateway placement is critical to provide low cost and Quality of Service (QoS). To address this design problem, we present an Integer Linear Programming (ILP) that minimizes the total cost of the network with respect to the deployed devices, while achieving mandatory QoS requirements. Using this formulation we obtain solutions for random topologies that exhibit the effectiveness of our approach in terms of network cost and efficiency.

Short distance optical interconnections, on-printed circuit boards, on-backplanes, and even on-chip, are a promising solution for replacing copper interconnections in future Data Center and HPC systems. Since photonic technology introduces new network building blocks, topology design for all the packaging levels should be reconsidered. This paper focuses on the on-board level of the packaging hierarchy, and proposes lay-out strategies for optical interconnection networks on optical printed circuit boards (OPCBs), based on direct topology families (tori, meshes and fully connected networks). We also describe a methodology for designing OPCBs given a set of input parameters, including building blocks specifications as well as traffic demands. The on-board topology design methodology generates all the feasible designs within the topology families examined, following our proposed OPCB lay-out approach, and selects the optimal designs based on specific optimization criteria.

SUMMARY We consider a hybrid electronic packet switched and optical circuit switched interconnection network for future high performance computing and datacenter systems. Given the logical task‐to‐task communication graph of an application, our objective is to cluster the logical parallel tasks to compute resources and configure the (reconfigurable) optical part of the hybrid interconnect to efficiently serve application communication requirements. We formulate the clustering and topology configuration problem in such a network, prove that it is NP‐complete, and provide an optimal algorithm to solve it based on an integer linear programming formulation. The integer linear programming algorithm is used to optimally solve small‐scale instances of the problem for the purpose of obtaining performance bounds. Aiming at large‐scale, we also present a heuristic based on simulated annealing that trades‐off performance for responsiveness. We measure the performance of a hybrid interconnect employing the proposed algorithm using real workloads, as well as extrapolated traffic, and compare it against application mapping on conventional fixed, electronic‐only interconnects based on toroidal topologies. Copyright © 2013 John Wiley & Sons, Ltd.

We consider the problem of serving traffic in a spectrum-flexible optical network, where the spectrum allocated to an end-to-end connection can change so as to adapt to the time-varying required transmission rate. In the proposed framework, each connection is assigned a route and is allocated a reference frequency over that route, using an appropriate Routing and Spectrum Allocation (RSA) algorithm, but the spectrum it utilizes around the reference frequency is allowed to expand and contract to match source rate fluctuations. We propose and analyze three spectrum expansion/contraction (SEC) policies for modifying the spectrum allocated to each connection. The first policy, named the Constant Spectrum Allocation (CSA) policy, allocates a number of spectrum slots for exclusive use by each connection. We also present two policies that enable the dynamic sharing of spectrum slots among connections, named the Dynamic High Expansion-Low Contraction (DHL) and the Dynamic Alternate Direction (DAD) policy. We give exact formulas for calculating the blocking probability for a connection and for the whole network under the CSA policy and provide corresponding approximate analyses under the DHL and DAD policies. We also present a simple iterative RSA algorithm that uses the developed blocking models so as to minimize the average blocking of the network.

We consider the online impairment-aware routing and wavelength assignment (IA-RWA) problem in transparent WDM networks. To serve a new connection, the online algorithm, in addition to finding a route and a free wavelength (a lightpath), has to guarantee its transmission quality, which is affected by physical-layer impairments. Due to interference effects, the establishment of the new lightpath affects and is affected by the other lightpaths. We present two multicost algorithms that account for the actual current interference among lightpaths, as well as for other physical effects, performing a cross-layer optimization between the network and physical layers. In multicost routing, a vector of cost parameters is assigned to each link, from which the cost vectors of the paths are calculated. The first algorithm utilizes cost vectors consisting of impairment-generating source parameters, so as to be generic and applicable to different physical settings. These parameters are combined into a scalar cost that indirectly evaluates the quality of candidate lightpaths. The second algorithm uses specific physical-layer models to define noise variance-related cost parameters, so as to directly calculate theQ-factor of candidate lightpaths. The algorithms find a set of so-called nondominated paths to serve the connection in the sense that no path is better in the set with respect to all cost parameters. To select the lightpath, we propose various optimization functions that correspond to different IA-RWA algorithms. The proposed algorithms combine the strength of multicost optimization with low execution times, making them appropriate for serving online connections.

Orthogonal Frequency Division Multiplexing (OFDM) has been recently proposed as a modulation technique for optical networks, due to its good spectral efficiency and impairment tolerance. Optical OFDM is much more flexible compared to traditional WDM systems, enabling elastic bandwidth transmissions. We consider the planning problem of an OFDM-based optical network where we are given a traffic matrix that includes the requested transmission rates of the connections to be served. Connections are provisioned for their requested rate by elastically allocating spectrum using a variable number of OFDM subcarriers. We introduce the Routing and Spectrum Allocation (RSA) problem, as opposed to the typical Routing and Wavelength Assignment (RWA) problem of traditional WDM networks, and present various algorithms to solve the RSA. We start by presenting an optimal ILP RSA algorithm that minimizes the spectrum used to serve the traffic matrix, and also present a decomposition method that breaks RSA into two substituent subproblems, namely, (i) routing and (ii) spectrum allocation (R+SA) and solves them sequentially. We also propose a heuristic algorithm that serves connections one-by-one and use it to solve the planning problem by sequentially serving all traffic matrix connections. To feed the sequential algorithm, two ordering policies are proposed; a simulated annealing meta-heuristic is also used to obtain even better orderings. Our results indicate that the proposed sequential heuristic with appropriate ordering yields close to optimal solutions in low running times.

We consider the offline version of the routing and wavelength assignment (RWA) problem in transparent all-optical networks. In such networks and in the absence of regenerators, the signal quality of transmission degrades due to physical layer impairments. Because of certain physical effects, routing choices made for one lightpath affect and are affected by the choices made for the other lightpaths. This interference among the lightpaths is particularly difficult to formulate in an offline algorithm since, in this version of the problem, we start without any established connections and the utilization of lightpaths are the variables of the problem. We initially present an algorithm for solving the pure (without impairments) RWA problem based on a LP-relaxation formulation that tends to yield integer solutions. Then, we extend this algorithm and present two impairment-aware (IA) RWA algorithms that account for the interference among lightpaths in their formulation. The first algorithm takes the physical layer indirectly into account by limiting the impairment-generating sources. The second algorithm uses noise variance-related parameters to directly account for the most important physical impairments. The objective of the resulting cross-layer optimization problem is not only to serve the connections using a small number of wavelengths (network layer objective), but also to select lightpaths that have acceptable quality of transmission (physical layer objective). Simulations experiments using realistic network, physical layer, and traffic parameters indicate that the proposed algorithms can solve real problems within acceptable time.

We propose and evaluate an impairment-aware multi-parametric routing and wavelength assignment algorithm for online traffic in transparent optical networks. In such networks the signal quality of transmission degrades due to physical layer impairments. In the multi-parametric approach, a vector of cost parameters is assigned to each link, from which the cost vectors of candidate lightpaths are calculated. In the proposed scheme the cost vector includes impairment generating source parameters, such as the path length, the number of hops, the number of crosstalk sources and other inter-lightpath interfering parameters, so as to indirectly account for the physical layer effects. For a requested connection the algorithm calculates a set of candidate lightpaths, whose quality of transmission is validated using a function that combines the impairment generating parameters. For selecting the lightpath we propose and evaluate various optimization functions that correspond to different IA-RWA algorithms. Our performance results indicate that the proposed algorithms utilize efficiently the available resources and minimize the total accumulated signal degradation on the selected lightpaths, while having low execution times.

We design and implement a multicost impairment- aware routing and wavelength assignment algorithm for online traffic. In transparent optical networks the quality of a transmission degrades due to physical layer impairments. To serve a connection, the proposed algorithm finds a path and a free wavelength (a lightpath) that has acceptable signal quality performance by estimating a quality of transmission measure, called the Q factor. We take into account channel utilization in the network, which changes as new connections are established or released, in order to calculate the noise variances that correspond to physical impairments on the links. These, along with the time invariant eye impairment penalties of all candidate network paths, form the inputs to the algorithm. The multicost algorithm finds a set of so called non-dominated Q paths from the given source to the given destination. Various objective functions are then evaluated in order to choose the optimal lightpath to serve the connection. The proposed algorithm combines the strength of multicost optimization with low execution time, making it appropriate for serving online connections.

We consider the offline version of the routing and wavelength assignment (RWA) problem in transparent all-optical networks. In such networks and in the absence of regenerators, the signal quality of transmission degrades due to physical layer impairments. We initially present an algorithm for solving the static RWA problem based on an LP relaxation formulation that tends to yield integer solutions. To account for signal degradation due to physical impairments, we model the effects of the path length, the path hop count, and the interference among ligthpaths by imposing additional (soft) constraints on RWA. The objective of the resulting optimization problem is not only to serve the connection requests using the available wavelengths, but also to minimize the total accumulated signal degradation on the selected lightpaths. Our simulation studies indicate that the proposed RWA algorithms select the lightpaths for the requested connections so as to avoid impairment generating sources, thus dramatically reducing the overall physical-layer blocking when compared to RWA algorithms that do not account for impairments.

We consider the offline version of the impairment-aware routing and wavelength assignment (IA-RWA) problem in transparent all-optical networks as a cross layer optimization problem. In optical networks and in the absence of regenerators, optical signal quality degrades due to physical layer impairments. We initially present an algorithm for solving the RWA problem based on an LP relaxation formulation that has acceptable integrality performance. To account for signal degradation due to physical layer impairments we extend our RWA formulation and constrain the interference among lightpaths using noise variance related parameters. The objective of the resulting optimization problem is not only to serve the connection requests by minimizing the number of utilized wavelengths, but also to select lightpaths that have acceptable physical layer performance.

A key problem in networks that support advance reservations is the routing and time scheduling of connections with flexible starting time. In this paper we present a multicost routing and scheduling algorithm for selecting the path to be followed by such a connection and the time the data should start so as to minimize the reception time at the destination, or some other QoS requirement. The utilization profiles of the network links, the link propagation delays, and the parameters of the connection to be scheduled form the inputs to the algorithm. We initially present a scheme of non-polynomial complexity to compute a set of so-called non-dominated candidate paths, from which the optimal path can be found. By appropriately pruning the set of candidate paths using path pseudo-domination relationships, we also find multicost routing and scheduling algorithms of polynomial complexity. We examine the performance of the algorithms in the special case of an Optical Burst Switched network. Our results indicate that the proposed polynomial time algorithms have performance that it is very close to that of the optimal algorithm.

In this paper we present a multicost algorithm for the joint time scheduling of the communication and computation resources that will be used by a task. The proposed algorithm selects the computation resource to execute the task, determines the path to route the input data, and finds the starting times for the data transmission and the task execution, performing advance reservations. We initially present an optimal scheme of non-polynomial complexity and by appropriately pruning the set of candidate paths we also give a heuristic algorithm of polynomial complexity. We evaluate the performance of our algorithm and compare it to that of algorithms that handle only the computation or communication part of the problem separately. We show that in a Grid network where the tasks are CPU- and data- intensive important performance benefits can be obtained by jointly optimizing the use of the communication and computation resources.

In this work we examine a task scheduling and data migration problem for grid networks, which we refer to as the data consolidation (DC) problem. DC arises when a task needs for its execution two or more pieces of data, possibly scattered throughout the grid network. In such a case, the scheduler and the data manager must select the data replicas to be used and the site where these will accumulate for the task to be executed. The policies for selecting the data replicas and the data consolidating site comprise the data consolidation problem. We propose and experimentally evaluate a number of DC techniques. Our simulation results brace our belief that DC is an important technique for data grids since it can substantially improve task delay, network load and other performance related parameters.

We design and implement various algorithms for solving the static RWA problem with the objective of minimizing the maximum number of requested wavelengths based on LP relaxation formulations. We present a link formulation, a path formulation and a heuristic that breaks the problem in the two constituent subproblems and solves them individually and sequentially. The flow cost functions that are used in these formulations result in providing integer optimal solutions despite the absence of integrality constraints for a large subset of RWA input instances, while also minimizing the total number of used wavelengths. We present a random perturbation technique that is shown to increase the number of instances for which we find integer solutions, and we also present appropriate iterative fixing and rounding methods to be used when the algorithms do not yield integer solutions. We comment on the number of variables and constraints these formulations require and perform extensive simulations to compare their performance to that of a typical min-max congestion formulation.

The existence of good probabilistic models for the job arrival process and job characteristics is important for the improved understanding of grid systems and the prediction of their performance. In this study, we present a thorough analysis of the job inter-arrival times, the waiting times at the queues, the execution times, and the data sizes exchanged at the kallisto.hellasgrid.gr cluster, which is part of the EGEE Grid infrastructure. By computing the Hurst parameter of the inter-arrival times we find that the job arrival process exhibits self-similarity/long-range dependence. We also propose simple and intuitive models for the job arrival process and the job execution times. The models proposed were validated and were found to be in very good agreement with our empirical measurements.

In this paper we present a signaling protocol for QoS differentiation suitable for optical burst switching networks. The proposed protocol is a two-way reservation scheme that employs delayed and in-advance reservation of resources. In this scheme delayed reservations may be relaxed, introducing a reservation duration parameter that is negotiated during call setup phase. This feature allows bursts to reserve resources beyond their actual size to increase their successful forwarding probability and is used to provide QoS differentiation. The proposed signaling protocol offers a low blocking probability for bursts that can tolerate the round-trip delay required for the reservations. We present the main features of the protocol and describe in detail timing considerations regarding the call setup and the reservation process. We also describe several methods for choosing the protocol parameters so as to optimize performance and present corresponding evaluation results. Furthermore, we compare the performance of the proposed protocol against that of two other typical reservation protocols, a Tell-and-Wait and a Tell-and-Go protocol.

In this paper, we present a new hybrid switching architecture, termed as SLIP-IN, that combines electronic packet/burst with optical circuit switching. SLIP-IN architecture takes advantages of the pre-transmission idle periods of optical lightpaths and slips into them packets or bursts of packets. In optical circuit switching (wavelength-routing) networks, capacity is immediately hard-reserved upon the arrival of a setup message, but is only used after a round-trip time delay. This idle period is significant for optical multi-gigabit networks and can be used to transmit traffic of a lower class of service. In this paper, we present the main features and dependencies of the proposed hybrid switching architecture, and further we perform a detailed evaluation by conducting network wide simulation experiments on the NSFnet backbone topology. For this purpose, we have developed an extensive network simulator, where the basic features of the architecture were modeled. The extensive network study revealed that SLIP-IN architecture can achieve and sustain an adequate data rate with a finite worst case delay.

We present a detailed performance evaluation of a hybrid optical switching architecture called Overspill Routing in Optical Networks (ORION). The ORION architecture combines wavelength and (electronic) packet switching, so as to obtain the advantages of both switching paradigms. We have developed an extensive network simulator where the basic features of the ORION architecture were modeled, including suitable load-varying sources and edge/core node architectures. Various aspects of the ORION architecture were studied including the routing policies used (i.e. once ORION always ORION and lightpath reentry) and the various options available for the buffer architecture. The complete network study shows that ORION can absorb temporary traffic overloads, as intended, provided sufficient buffering is present.

We present a 40 Gb/s asynchronous self-routing network and node architecture that exploits bit and packet level optical signal processing to perform synchronization, forwarding and switching. Optical packets are self-routed on a hop-by-hop basis through the network by using stacked optical tags, each representing a specific optical node. Each tag contains control signals for configuring the switching matrix and forwarding each packet to the appropriate outgoing link and onto the next hop. Physical layer simulations are performed, modeling each optical sub-system of the node showing acceptable signal quality and bit error rates. Resource reservation-based signaling algorithms are theoretically modeled for the control plane capable of providing high performance in terms of blocking probability and holding time.