The growing demand for services and the rapid deployment of virtualized network functions (VNFs) pose significant challenges for achieving low-latency and energy-efficient orchestration in modern edge-core network infrastructures. To address these challenges, this study proposes a Digital Twin (DT)-empowered Deep Reinforcement Learning framework for intelligent VNF migration that jointly minimizes average end-to-end (E2E) delay and energy consumption. By formulating the VNF migration problem as a Markov Decision Process and utilizing the Advantage Actor-Critic model, the proposed framework enables adaptive and real-time migration decisions. A key innovation of the proposed framework is the integration of a DT module composed of a multi-task Variational Autoencoder and a multi-task Long Short-Term Memory network. This combination collectively simulates environment dynamics and generates high-quality synthetic experiences, significantly enhancing training efficiency and accelerating policy convergence. Simulation results demonstrate substantial performance gains, such as significant reductions in both average E2E delay and energy consumption, thereby establishing new benchmarks for intelligent VNF migration in edge-core networks.

Motivated by cooperative distributed networks in which users dynamically alternate between transmit and receive modes under half-duplex constraints, this paper studies the Age of Information (AoI) in a distributed multi-user network using an ALOHA-based protocol. We derive closed-form expressions for the average AoI and formulate an optimization problem over transmission probabilities. After proving the convexity of the problem, we leverage the derived optimality conditions to characterize optimal policies for general network graphs, obtain closed-form solutions for $d$-regular topologies, and derive tractable optimality conditions for star topologies. Numerical results confirm that the proposed mechanism can effectively and adaptively determine user-specific optimal transmission probabilities across varying network topologies. These findings contribute to the design of adaptive and efficient distributed networks with enhanced information freshness.

The next generation networks offers significant potential to advance Intelligent Transportation Systems (ITS), particularly through the integration of Digital Twins (DTs). However, ensuring the uninterrupted operation of DTs through efficient computing resource management remains an open challenge. This paper introduces a distributed computing architecture that integrates DTs and Mobile Edge Computing (MEC) within a software-defined vehicular networking framework to enable intelligent, low-latency transportation services. A network aware scalable collaborative task provisioning algorithm is developed to train an autonomous agent, which is evaluated using a realistic connected autonomous vehicle (CAV) traffic simulation. The proposed framework significantly enhances the robustness and scalability of DT operations by reducing synchronization errors to as low as 7% while achieving up to 99.5% utilization of edge computing resources.

In this paper, we examine a multi-sensor system where each sensor monitors multiple dynamic information processes and transmits updates over a shared communication channel. These updates may include correlated information across the various processes. In this type of system, we analyze the impact of preemption, where ongoing transmissions are replaced by newer updates, on minimizing the Age of Information (Aol). While preemption is optimal in some scenarios, its effectiveness in multisensor correlated systems remains an open question. To address this, we introduce a probabilistic preemption policy, where the source sensor preemption decision is stochastic. We derive closedform expressions for the Aol and frame its optimization as a sum of linear ratios problem, a well-known NP-hard problem. To navigate this complexity, we establish an upper bound on the iterations using a branch-and-bound algorithm by leveraging a reformulation of the problem. This analysis reveals linear scalability with the number of processes and a logarithmic dependency on the reciprocal of the error that shows the optimal solution can be efficiently found. Building on these findings, we show how different correlation matrices can lead to distinct optimal preemption strategies. Interestingly, we demonstrate that the diversity of processes within the sensors' packets, as captured by the correlation matrix, plays a more significant role in preemption priority than the number of updates.

Six-generation mobile systems aim to support stringent end-to-end service-level agreements for diverse user applications simultaneously. This paper introduces novel multi-domain computation-aware resource slicing orchestration that jointly manages in-network communications, computation, and caching storage resources for multi-domain networking. It automatically programs wireless access, edge cloud, and regional/central cloud infrastructure to enable wireless-optical network virtualization. Specifically, a mobile virtual network operator's long-term profit maximization problem and two subproblems are formulated to slice wired and wireless infrastructure resources and assign user requests and contents to slices. Accordingly, a reinforcement learning-based slicing with greedy pre-caching is proposed, which automatically allocates in-network resources for dynamic wireless connectivity and supports real-time inferring with minimal user request signaling. Numerical results show that our solutions provide superior performance from both user and infrastructure perspectives, with 20% improved operator profits, 17% enhanced service provisioning rates, and 80% reduced delay when simultaneously serving augmented reality and large language model's quality demands. This innovation exploits a generalized rein-forcement learning approach to minimize the signaling overheads and computation complexity while agilely adapting to practical converged networks, thus benchmarking AI-driven multi-domain network slicing development.

The rapid evolution of Industry 4.0 has led to the emergence of smart factories, where multirobot system autonomously operates to enhance productivity, reduce operational costs, and improve system adaptability. However, maintaining reliable and efficient network operations in these dynamic and complex environments requires advanced automation mechanisms. This study presents a zero-touch network platform that integrates a hierarchical Open Radio Access Network (O-RAN) architecture, enabling the seamless incorporation of advanced machine learning algorithms and dynamic management of communication and computational resources, while ensuring uninterrupted connectivity with multirobot system. Leveraging this adaptability, the platform utilizes federated deep reinforcement learning (FedDRL) to enable distributed decision-making across multiple learning agents, facilitating the adaptive parameter reconfiguration of transmitters (i.e., multirobot system) to optimize long-term system throughput and transmission energy efficiency. Simulation results demonstrate that within the proposed O-RAN-enabled zero-touch network platform, FedDRL achieves a 12% increase in system throughput, a 32% improvement in normalized average transmission energy efficiency, and a 28% reduction in average transmission energy consumption compared to baseline methods such as independent DRL.

The need for prioritized multi-tenant quality of service (QoS) management in emerging mobile edge systems is particularly critical for high-throughput next generation networks. Current traffic engineering tools rely on network administrator driven, complex functions embedded in closed, proprietary infrastructures, which significantly restrict design flexibility, scalability, and adaptability. This study addresses these challenges by proposing a software-defined networking (SDN) based dynamic QoS provisioning scheme, powered by a digital twin (DT) of networks. By separating the control and data planes, the scheme enables automated traffic management through SDN programmability and data-driven decision-making. It incorporates few-shot learning to dynamically identify and prioritize multi-tenant network traffic utilizing flow statistics from SDN. The proposed QoS provisioning mechanism allocates sufficient bandwidth to high-priority flows while optimizing the remaining bandwidth for lower-priority traffic. Performance evaluations show that the model achieves up to 98% accuracy in identifying the priority of previously unseen traffic flows. Hardware-in-the-loop (HiL) simulations further validate the scheme's effectiveness in meeting multi-tenant QoS requirements, offering a robust and scalable solution for traffic prioritization in SDN based edge networks.

The integration of 5G terrestrial networks with Low Earth Orbit (LEO) satellites has the potential to provide seamless global connectivity and enhanced service quality, particularly in regions with limited terrestrial infrastructure such as rural areas. Furthermore, the incorporation of Multi-connectivity (MC) enables user equipment (UEs) to maintain simultaneous connections with both terrestrial 5G base stations and LEO satellites, improving system reliability. However, the high mobility of LEO satellites and the dynamic behavior of UEs present significant challenges, particularly in handover decision-making which can adversely impact system throughput, and quality of service (QoS). To address these challenges, we propose a novel deep reinforcement learning-based approach that integrates online Random Ensemble Mixture and Dual Experience Replay into a Dueling Double Deep Q-Network architecture. This proposed scheme intelligently optimizes handover decisions in MC-enabled terrestrial-satellite networks, improving decision accuracy in highly dynamic scenarios. Simulation results demonstrate substantial gains in system throughput, reduced system delay, average handover reduction, and increased transmission success probability, setting new performance benchmarks for integrated terrestrial-satellite networks while adhering to diverse QoS requirements.

The distributed and disaggregated architecture of next-generation (NextG) networks, including 6G has sparked growing interest in federated learning (FL) as a strategy for enabling privacy-preserving collaborative network traffic analysis at the edge. However, FL encounters significant challenges due to data heterogeneity driven by diverse data distributions across edge nodes, and the scarcity of labeled data further worsened by the time-intensive process of data labeling. Although a few studies have addressed these challenges in network traffic analysis tasks using Multi-Task Learning (MTL), existing approaches pre-dominantly focus on single-task FL, centralized model solutions and overlook the integration of MTL in NextG networks. To bridge this gap, we propose O-FedMTL, a novel framework that combines FL with MTL to enable cooperative traffic analysis within an Open Radio Access Network (O-RAN) environment in NextG networks. MTL enhances FL by mitigating the issues of data heterogeneity and labeled data scarcity through shared knowledge derived from multiple interconnected traffic analysis tasks, i.e., traffic classification, flow duration analysis, and bandwidth estimation. Additionally, MTL offers significant benefits by reducing energy consumption and computation costs at the edge through the simultaneous processing of these tasks within a single model. Extensive experimental results demonstrate that O-FedMTL achieves the target global accuracy for traffic classification, flow duration analysis, and bandwidth estimation with 20, 12, and 23 fewer global communication rounds, respectively, compared to the baseline federated averaging. Additionally, O-FedMTL reduces computation costs by 43% compared to the baseline-combined.

With the evolution of modern telecommunications, the fronthaul network has become an indispensable component of the infrastructure, particularly in the context of Cloud Ra-dio Access Networks. Nevertheless, fronthaul networks still face significant challenges such as power outages, particularly when a disaster, like an earthquake or severe weather event, occurs. Damage to power supply facilities may cause operational failures while communication is one of the most crucial needs in the disaster area to make rescue operations more effective. Specialized fibers that can deliver electrical power can help mitigate this problem. However, power losses may be extremely large over distances and there is no flexibility after the installation of the fibers. The network topology design that reduces capital and operational costs while satisfying power constraints is an important problem. In this paper, we focus on network topology design in an urban area by taking into account both fiber and power costs. We then propose integer linear programming and fast algorithms based on methods for single-facility location problems. The results demonstrate that multiple approaches help to achieve the optimal design, with our proposed method standing out due to its efficiency in finding feasible solutions, scalability, and ≈656x reduction in execution time.

In this paper, we examine a multi-sensor system where each sensor may monitor more than one time-varying information process and send status updates to a remote monitor over a common channel. We consider that each sensor's status update may contain information about more than one information process in the system subject to the system's constraints. To investigate the impact of this correlation on the overall system's performance, we conduct an analysis of both the average Age of Information (AoI) and source state estimation error at the monitor. Building upon this analysis, we subsequently explore the impact of the packet arrivals, correlation probabilities, and rate of processes' state change on the system's performance. Next, we consider the case where sensors have limited sensing abilities and distribute a portion of their sensing abilities across the different processes. We optimize this distribution to minimize the total AoI of the system. Interestingly, we show that monitoring multiple processes from a single source may not always be beneficial. Our results also reveal that the optimal sensing distribution for diverse arrival rates may exhibit a rapid regime switch, rather than smooth transitions, after crossing critical system values. This highlights the importance of identifying these critical thresholds to ensure effective system performance.

In the realms of the internet of vehicles (IoV) and intelligent transportation systems (ITS), software defined vehicular networks (SDVN) and edge computing (EC) have emerged as promising technologies for enhancing road traffic efficiency. However, the increasing number of connected autonomous vehicles (CAVs) and EC-based applications presents multi-domain challenges such as inefficient traffic flow due to poor CAV coordination and flow-table overflow in SDVN from increased connectivity and limited ternary content addressable memory (TCAM) capacity. To address these, we focus on a data-driven approach using virtualization technologies like digital twin (DT) to leverage real-time data and simulations. We introduce a DT design and propose two data-driven solutions: a centralized decision support framework to improve traffic efficiency by reducing waiting times at roundabouts and an approach to minimize flow-table overflow and flow re-installation by optimizing flow-entry lifespan in SDVN. Simulation results show the decision support framework reduces average waiting times by 22% compared to human-driven vehicles, even with a CAV penetration rate of 40%. Additionally, the proposed optimization of flow-table space usage demonstrates a 50% reduction in flow-table space requirements, even with 100% penetration of connected vehicles.

Bayesian optimization (BO) is a popular global optimization scheme for sample-efficient optimization in domains with expensive function evaluations. The existing BO techniques are capable of finding a single global optimum solution. However, finding a set of global and local optimum solutions is crucial in a wide range of real-world problems, as implementing some of the optimal solutions might not be feasible due to various practical restrictions (e.g., resource limitation, physical constraints, etc.). In such domains, if multiple solutions are known, the implementation can be quickly switched to another solution, and the best possible system performance can still be obtained. This paper develops a multimodal BO framework to effectively find a set of local/global solutions for expensive-to-evaluate multimodal objective functions. We consider the standard BO setting with Gaussian process regression representing the objective function. We analytically derive the joint distribution of the objective function and its first-order derivatives. This joint distribution is used in the body of the BO acquisition functions to search for local optima during the optimization process. We introduce variants of the well-known BO acquisition functions to the multimodal setting and demonstrate the performance of the proposed framework in locating a set of local optimum solutions using multiple optimization problems.

C-band enabled Elastic optical networks (EONs) have been one of the most deployed optical network solutions in the world. However, as traffic demands continue to increase, capacity exhaustion is inevitable. There are two major technologies, namely, multiband elastic optical networks (MB-EONs) and space division multiplexed elastic optical networks (SDM-EONs) that can enhance capacity. Each technology offers a tradeoff between better capacity and deployment overhead which directly affects the network performance. Considering the different characteristics of these two technologies, we present our proposed strategy, Progressive Optics Deployment and Integration for Growing Yields (PRODIGY), to gradually migrate the current C-band EONs. PRODIGY uses various proactive measures, inspired by Swiss Cheese Model, to make the network robust for handling network traffic peaks and ensure that the service level agreement is met. We present a detailed comparison of our proposed strategy with customized baseline strategies, and demonstrate the superiority of our proposed approach.

Age of Information (AoI) is a well-investigated timeliness metric for data collected from sensors. Various packet scheduling policies have been proposed in order to optimize AoI. It is natural to raise the question of how much these scheduling policies could be improved using prediction and how many predictive packets are necessary to achieve the optimum. Given a packet sequence, there must be at least one packet combination to achieve the optimal AoI, which can be obtained by exhausting all possible decisions of preserving/rejecting packets. However, it is impractical to obtain prior knowledge of each update’s arrival and service time. In addition, this exhaustive scheduling policy consumes unaffordable resources of time, energy, and computing power.In this paper, we show that a sufficiently long packet sequence may be segmented into local epochs with invariant global optimal policy, and local optimization of each epoch incrementally aggregates the global optimal AoI of the entire sequence. This new perspective also explains the counter-intuitive phenomenon [1]–[3] of idle waiting time outperforming transmitting updates immediately. After comparing with the AoIs obtained from other scheduling policies, we find that the optimization performance of predictive scheduling prevails over others.

In this paper, we introduce a new metric, named Penalty upon Decision (PuD), for measuring the impact of communication delays and state changes at the source on a remote decision maker. Specifically, the metric quantifies the performance degradation at the decision maker's side due to delayed, erroneous, and (possibly) missed decisions. We clarify the rationale for the metric and derive closed-form expressions for its average in M/GI/1 and M/GI/1/1 with blocking settings. Numerical results are then presented to support our expressions and to compare the infinite and zero buffer regimes. Interestingly, comparing these two settings sheds light on a buffer length design challenge that is essential to minimize the average PuD.

Unmanned aerial vehicle (UAV)-empowered edge computing has been widely investigated in obstacle-free scenarios, where a moving UAV is in charge of handling offloaded singleton tasks from mobile devices on the ground. However, little attention has been paid to the scenario, in which the UAV serves a complex area withmultiple obstaclesanddependent tasks. A dependent task can be formulated as a directed acyclic graph (DAG) that contains a number of subtasks; and each subtask can be executed by a corresponding service function (SF) deployed on the UAV. In this backdrop, the joint UAV trajectory planning, DAG task scheduling, and SF deployment is formulated as an optimization problem in this article. Afterwards, a deep reinforcement learning (DRL)-based algorithm is presented to tackle the NP-hard problem. The state space, action space, and the reward function of the agent, i.e., the UAV, are defined, respectively, under the DRL framework. To evaluate the effectiveness of the proposal, a series of experiments is conducted with different parameter settings. Results show that the DRL-based algorithm performs much better than three heuristic algorithms in success rate of trajectory planning, the number of executed tasks, and the average task response latency.

Routing and Spectrum Assignment (RSA) is a well-known problem in Elastic Optical Networks (EONs), and the literature abounds with numerous heuristic algorithms that attempt to efficiently utilize spectrum and mitigate spectrum fragmentation. However, blocking performance modeling of EONs is a relatively under-explored topic, mainly because of the complexity of handling variable request sizes. In this paper, we analyze three spectrum assignment policies and compare their blocking performance. In the full slicing policy, every request is sliced into multiple one-slot calls. In the partitioning policy, each request size is allocated a dedicated spectrum partition, and each request can be only assigned spectrum from its corresponding partition. The partition with slicing policy is a hybrid of the two – there are dedicated partitions for small request sizes, but large requests are sliced into multiple sub-requests of the smaller sizes and assigned spectrum from those partitions. We obtain expressions for the bandwidth blocking probability for these three policies, and validate the analysis with extensive simulation results. Our results show that partitioning with slicing has the best performance.

Age of Information (AoI) is a widely used metric of information freshness in a status update communication system. In this paper, we study the AoI performance in a multi-source system in which multiple sources generate updates, but only one can be transmitted to a monitor at a time. Scheduling, or selecting the source update to transmit, is an important problem in such a system. We consider a time-slotted system with Bernoulli sources with no buffer. An update selected for transmission in a slot is delivered at the end of the slot. The AoI is analyzed for four different scheduling policies - random scheduling, round robin scheduling, age-greedy scheduling, and the Whittle index-based policy proposed in [1], [2]. Our analytical results are validated with simulation results, and a numerical performance comparison of the scheduling policies is presented.

Elastic optical networks (EONs) have emerged as attractive candidates to satisfy the dramatic growth of demand in 5G and cloud applications. EONs promise to provide high spectrum utilization due to flexibility in resource assignment. In translucent EONs, the spectrum efficiency can be further improved by deploying regenerators. Because of their extremely high flexibility, developing efficient mechanisms and strategies to ensure the survivability of translucent EONs is a challenging problem. In this paper, we consider disaster mitigation in translucent EONs. We propose a new approach to disaster management by introducing the concept of mitigation zone, which identifies a region surrounding the disaster zone wherein lightpaths may be reconfigured with degraded service (with a penalty) in order to improve overall performance. We formulate an integer linear program (ILP) to minimize the penalty due to service degradation after a disaster, and present a heuristic algorithm named Disaster Management Algorithm with Mitigation Awareness and 3R regenerators (DRAMA+). Simulation results demonstrate that the proposed algorithms have a better performance in terms of total penalty and blocking ratio than conventional disaster recovery algorithms.

Elastic Optical Networks (EONs) have become a promising solution to satisfy the dramatic growth of bandwidth demand due to 5G and cloud applications. Due to the flexibility of resource allocation, EONs provide high spectrum utilization efficiency, and because of this, developing efficient policies to ensure the survivability of EONs is a challenging problem. A well-designed disaster management plan is needed to prevent data loss during network failures and large-scale disasters. The bottleneck problem caused by disabled parts of the network causes difficulties for disaster recovery. Depending on the disaster, even traffic that may be far away from the disaster may be impacted by it. In this paper, we propose a new approach to disaster management using machine learning to facilitate efficient recovery. In addition to traffic immediately affected by the disaster, all traffic which is “close to” the disaster is re-routed and re-assigned with possibly degraded service, while requests “far from” the disaster are left unaffected. A deep reinforcement learning disaster recovery algorithm with mitigation awareness (DeepDRAMA) is proposed for recovery. A novel deep reinforcement learning agent is designed and trained for the agent to select the appropriate level of service degradation for re-assigned traffic. Simulation results show the performance improvement with DeepDRAMA.

Recent advances in Spectrally-Spatially Flexible Optical Networks (SS-FONs) have enhanced service provisioning by leveraging elastic optical transmission with a fine-grained and flexible frequency grid, multiple spatial modes (fiber cores), and a variety of advanced modulation formats (MFs) to increase the capacity of optical fiber. An important lightpath resource assignment problem in SS-FONs is the routing, modulation, core, and spectrum assignment (RMCSA) problem. Crosstalk (XT) between connections on different cores degrades the quality of transmission, and the RMCSA algorithm must ensure that XT constraints are met while maximizing performance. In this paper, we propose an RMCSA algorithm called Tridental Resource Assignment algorithm (TRA), as it balances network capacity, spectrum utilization and spectrum fragmentation that affect network performance. Our resource assignment approach includes both an offline/static network planning component and an online/dynamic provisioning component. In the former, MFs and spectrum utilization are used to compute path priorities for a lightpath. The dynamic provisioning component then allocates the resources on a selected path using TRA. Extensive simulation experiments performed in realistic network scenarios indicate that TRA significantly reduces the bandwidth blocking probability (BBP) in a variety of scenarios by maintaining a good balance between spectrum utilization and XT.

A combination of mobile edge computing (MEC) and cloud computing paradigms has the potential to greatly alleviate the challenges facing Internet of Things (IoT). We consider a tiered IoT infrastructure in which data generated by an IoT sensor/device is delivered to a data center for processing through an intermediate MEC server. The MEC server can either directly transmit the data to the data center or pre-process the data and then transmit it to the data center over a shared channel. The goal is to maintain the freshness of the data delivered to the data center. In this paper, we assume a probabilistic model for pre-processing by the MEC server. Sensor data is assumed to be generated as a Poisson process and the transmission times over the two paths are assumed to have general distributions.We use Age of Information (AoI) as a measure of data freshness at the data center. We perform stationary distribution analysis in this system and obtain closed form expressions for average AoI and average peak AoI. We focus on selecting the offloading probabilities in conjunction with the mean service times for each server for optimal operation determined by average AoI and peak AoI. Our numerical results show the effect of path diversity in the selection of best offloading probability and service times.

Distributed computing systems often need to consider the scheduling problem involving a collection of highly dependent data-processing tasks that must work in concert to achieve mission-critical objectives. This paper considers the unrelated machine scheduling problem for minimizing weighted sum completion time under arbitrary precedence constraints and on heterogeneous machines with different processing speeds. The problem is known to be strongly NP-hard even in the single machine setting. By making use of Queyranne’s constraint set and constructing a novel Linear Programming relaxation for the scheduling problem under arbitrary precedence constraints, our results in this paper advance the state of the art. We develop a 2(1+(m-1)/D)-approximation algorithm (and 2(1+(m-1)/D)+1-approximation) for the scheduling problem with zero release time (and arbitrary release time), where m is the number of servers and D is the task-skewness product. The algorithm can be efficiently computed in polynomial time using the Ellipsoid method and achieves nearly optimal performance in practice as D>O(m) when the number of tasks per job to schedule is sufficiently larger than the number of machines available. Our implementation and evaluation using a heterogeneous testbed and real-world benchmarks confirms significant improvement in weighted sum completion time for dependent computing tasks.

A datacenter that consists of hundreds or thousands of servers can provide virtualized environments to a large number of cloud applications and jobs that value the requirement of reliability very differently. Checkpointing a virtual machine (VM) is a proven technique to improve reliability. However, existing checkpoint scheduling techniques for enhancing reliability of distributed systems fails to achieve satisfactory results, either because they tend to offer the same, fixed reliability to all jobs, or because their solutions are tied up to specific applications and rely on centralized checkpoint control mechanisms. In this work, we first show that reliability can be significantly improved through contention-free scheduling of checkpoints. Then, inspired by the Carrier Sense Multiple Access (CSMA) protocol in wireless congestion control, we propose a novel framework for distributed and contention-free scheduling of VM checkpointing to provide reliability as a transparent, elastic service. We quantify reliability in closed form by studying system stationary behaviours, and maximize job reliability through utility optimization. Our design is validated via a proof-of-concept prototype that leverages readily available implementations in Xen hypervisors. The proposed checkpoint scheduling is shown to significantly reduce checkpointing interference and improve reliability by as much as one order of magnitude over contention-oblivious checkpoint schemes.

Edge computing applications typically require generated data to be preprocessed at the source and then transmitted to an edge server. In such cases, transmission time and preprocessing time are coupled, yielding a tradeoff between them to achieve the targeted objective. This paper presents analysis of such a system with the objective of optimizing freshness of received data at the edge server. We model this system as two queues in tandem whose service times are independent but the transmission service time is monotonically dependent on the computation service time in mean value. This dependence captures the natural decrease in transmission time due to lower offloaded computation. We analyze various queue management schemes in this tandem queue where the compute queue has a single server, Poisson packet arrivals, general independent service and no extra buffer to save incoming packets. The transmit queue has a single server receiving packets from the compute queue with memoryless service time. We consider the transmit queue in two forms: (i) No data buffer and (ii) One unit data buffer and last come first serve with discarding. We analyze various non-preemptive as well as preemptive cases. We perform stationary distribution analysis and obtain closed form expressions for average age of information (AoI) and average peak AoI. Our numerical results illustrate analytical findings on how computation and transmission times could be traded off to optimize AoI and reveal a consequent tradeoff between average AoI and average peak AoI.

The Internet gateway selection problem is becoming very important as the number of Internet-connected devices increases and stresses the limited number of Internet gateway nodes. The gateway nodes often experience frequent performance fluctuations, and the best gateway selection candidate changes frequently with growing network dynamics. We propose GateSelect, a customized selection algorithm for each client node that not only provides the best-effort selection candidate but also ensures the global, balanced distribution of the gateway nodes. We utilize over the counter, lightweight calculations to optimize the client-side selection algorithm by combining classification, short term performance prediction, and randomized selection. We compare our algorithm with several baseline algorithms, and the experiment results show that our proposal provides better performance and balanced distribution of gateway nodes.

Elastic optical networks are promising candidates for next-generation backbone networks due to their capability to efficiently and flexibly allocate optical bandwidth to demands with heterogeneous requirements. Scientific applications involving data processing, analysis, and transmission among multiple data centers can be modeled as virtual network requests. For delay-tolerant applications, the virtual network request has a specific deadline and is successfully mapped as long as sufficient resources are allocated to it before the deadline. In this paper, we investigate the dynamic virtual network mapping problem for advance reservation virtual network requests in elastic optical networks. We propose a stop-and-resume allocation scheme to map virtual network requests in order to satisfy deadline constraints and resource requirements, where the service duration of a request is divided into periods with different resources being allocated in different periods. Simulation results are presented to show the effectiveness of the proposed approach compared with a baseline algorithm in terms of request blocking ratio.

Coping with diverse channel access attacks (CAAs) has been a major obstacle to realize the full potential of wireless networks as a basic building block of smart applications. Identifying and classifying different types of CAAs in a timely manner is a great challenge because of the inherently shared nature and randomness of the wireless medium. To overcome the difficulties encountered in existing methods, such as long latency, high data collection overhead, and limited applicable range, a deep learning-based CAA detection framework is proposed in this paper. First, we show the challenges of CAA classification by analyzing the impacts of CAAs on wireless network performance using an event-driven network simulator. Second, a state-transition model is built for the channel access process at a node, whose output sequences characterize the changing patterns of the node's transmission status in different CAA scenarios. Third, a deep learning-based CAA classification framework is presented, which takes state transition sequences of a node as input and outputs predicted CAA types. The performance of three deep neural networks, i.e., fully-connected, convolutional, and Long Short-Term Memory (LSTM) network, for classifying CAAs are evaluated under our CAA classification framework in five CAA scenarios and the normal scenario without CAA. Experimental results show that LSTM outperforms the other two neural network architectures, and its CAA classification accuracy is higher than 95%. We successfully transferred the learned LSTM model to classify CAAs on other nodes in the same network and the nodes in other networks, which verifies the generality of our proposed framework.

Motivated by the inherent value of packets arising in many cyber-physical applications (e.g., due to precision of the information content or an alarm message), we consider status update systems with update packets carrying values as well as their generation time stamps. Once generated, a status update packet has a random initial value and a deterministic deadline after which it is not useful (ultimate staleness). In our model, value of a packet decreases in time (even after reception) starting from its generation to ultimate staleness when it vanishes. The value of information (VoI) at the receiver is additive in that the VoI is the sum of the current values of all packets held by the receiver. We investigate various queuing disciplines under potential dependence between value and service time and provide closed form expressions for average VoI at the receiver. Numerical results illustrate the average VoI for different scenarios and the contrast between average age of information (AoI) and average VoI.

In this paper, we explore the potential of server waiting before packet transmission in improving the Age of Information (AoI) in status update systems. We consider a non-preemptive queue with Poisson arrivals and independent general service distribution and we incorporate waiting before serving in two packet management schemes: M/GI/1/1 and M/GI/1/2*. In M/GI/1/1 there is no data buffer and a packet is taken to server only when it is idle whereas in M/GI/1/2* there is a single unit buffer that captures the latest arriving status update packet. In both schemes, we introduce “server waiting”: In M/GI/1/1 scheme, the server waits for a deterministic time immediately after a packet enters the server. In M/GI/1/2* scheme, depending on queue state, the server waits for a deterministic time before starting service. In both cases, if a newer arrival is captured, existing packet is discarded with no reset of the waiting period. Different from most existing works, we analyze AoI evolution by indexing the incoming packets, which is enabled by an alternative method of partitioning the area under the evolution of instantaneous AoI to calculate its time average. We obtain expressions for average AoI and average peak AoI for both queueing disciplines with waiting. Our numerical results demonstrate that waiting before service can bring significant improvement in average age, particularly, for heavy-tailed service distributions. This improvement comes at the expense of an increase in average peak AoI. We highlight the trade-off between average and average peak AoI generated by waiting before serving.

With the growing deadline-sensitivity of cloud applications, adherence to specific deadlines is becoming increasingly crucial, particularly in shared clusters. A few slow tasks called stragglers can potentially adversely affect job execution times. Equally, inadequate slotting of data analytics applications could result in inappropriate resource deployment, ultimately damaging system performance. Against this backdrop, one effective way of tackling stragglers is by making extra attempts (or clones)1 for every single straggler after the submission of a job. This paper proposes Shed+, which is an optimization framework utilizing dynamic speculation that aims to maximize the jobs' PoCD (Probability of Completion before Deadline) by making full use of available resources. Notably, our work encompasses a new online scheduler that dynamically recomputes and reallocates resources during the course of a job's execution. According to our findings, Shed+ successfully leverages cloud resources and maximizes the percentage of jobs meeting their deadlines. In our experiments, we have seen this percentage for heavy load going up to 98% for Shed+ as opposed to nearly 68%, 40%, 35% and 37% for Shed, Dolly, Hopper and Hadoop with speculation enabled, respectively.

This paper considers the protection of translucent elastic optical networks (EONs) through p-cycles. Such networks improve spectrum efficiency by employing regenerators and using advanced modulation formats for transmission. P-cycles provide fast restoration and high protection efficiency, and have been studied for conventional fixed-grid WDM networks as well as EONs. In this paper, we consider the design and selection of p- cycles for translucent EONs with 3R regenerators in a network. We propose two novel link-protection p-cycle evaluation methods in translucent EONs: individual p-cycle selection and p-cycle set selection. Based on these two metrics, Traffic Independent P-cycle Selection with 3R regenerator (TIPS-3R) and Traffic-Oriented P-cycle Selection with 3R regenerator (TOPS-3R), are designed to find the best set of p-cycles under a given 3R regenerator placement. We evaluate our algorithms using both static traffic and dynamic traffic. Simulation results indicate that the proposed algorithms have a lower spectrum usage and lower blocking ratio compared with baseline algorithms.

Massive data centers are at the heart of the Internet. The rapid growth of Internet traffic and the abundance of rich data-driven applications have raised the need for enormous network bandwidth. Towards meeting this growing traffic demand, optical interconnects have gained significant attention, as they can provide high throughput, low latency, and scalability. In particular, optical Wavelength Division Multiplexing (WDM) provides the possibility to build data centers comprising of millions of servers, while providing hundreds of terabits per second bandwidth. In this paper, we propose a WDM-based Reconfigurable Hierarchical Optical Data Center Architecture (RHODA) that can satisfy future Internet traffic demands. To improve scalability, our DCN architecture is hierarchical, as it groups server racks into clusters. Cluster membership is reconfigurable through the use of optical switches. Each cluster enables heavy-traffic communication among the racks within. To support varying traffic patterns, the inter-cluster network topology and link capacities are also reconfigurable, which is achieved through the use of optical space switches and Wavelength Selective Switches (WSSs). Our simulation results demonstrate that in terms of average hop distance, RHODA outperforms OSA, FatTree and WaveCube by up to 81%, 66% and 60%, respectively.

In this paper, we introduce a new data freshness metric, relative Age of Information (rAoI), and examine it in a single server system with various packet management schemes. The (classical) AoI metric was introduced to measure the end to end freshness of status updates at the receiver with respect to their generation at the source. In order to understand the efficiency of the update delivery system in isolation, we introduce rAoI metric to measure how fresh the data is at the receiver with respect to the data at the transmitter. This metric introduces an explicit dependence on the arrival process in the evaluation of age. We investigate several queuing disciplines and provide closed form expressions for rAoI and numerical comparisons.

This paper is motivated by emerging edge computing applications in which generated data are preprocessed at the source and then transmitted to an edge server. In such a scenario, there is typically a tradeoff between the amount of pre-processing and the amount of data to be transmitted. We model such a system by considering two non-preemptive queues in tandem whose service times are independent over time but the transmission service time is dependent on the computation service time in mean value. The first queue is in M/GI/1/1 form with a single server, memoryless exponential arrivals, general independent service and no extra buffer to save incoming status update packets. The second queue is in GI/M/1/2*form with a single server receiving packets from the first queue, memoryless service and a single data buffer to save incoming packets. Additionally, mean service times of the first and second queues are dependent through a deterministic monotonic function. We perform stationary distribution analysis in this system and obtain closed form expressions for average age of information (AoI) and average peak AoI. Our numerical results illustrate the analytical findings and highlight the tradeoff between average AoI and average peak AoI generated by the tandem nature of the queueing system with dependent service times.

Unmanned aerial vehicles (UAVs) are considered a promising solution for carrying communications and computational facilities to increase the flexibility of cloud-to-thing continuum, where short-range and long-range wireless links are adopted to connect mobile devices to the fog node and the fog node to the remote data center, respectively. Most existing UAV-involved resource allocation algorithms focus mainly on the radio resource allocation problem, and much less attention has been paid to the allocation of computational resources. Moreover, the dynamic arrival of tasks and the queueing delay at each computation entity is usually neglected. In this paper, a joint optimization problem is formulated that takes the weighted sum of energy consumption and delay experienced by tasks as the objective function. Processing frequencies and transmission powers of mobile devices and the fog node are the decision variables in the problem. To solve this problem, three decision-making algorithms are presented. The first one is used to decide the UAV's position. The processing frequency, transmission power, and task assignment results at mobile devices are determined by the second algorithm. The last one is adopted by the fog node to optimize its processing frequency and transmission power. A series of simulation experiments are conducted to evaluate the effectiveness of the proposed algorithms. Compared with the random task assignment scheme with fixed parameters, the combination of our three algorithms always perform much better for a wide range of parameter settings.

As cloud applications are becoming increasingly deadline-sensitive, meeting desired deadlines is more critical, especially in shared clusters. It has been shown that a few slow tasks, called stragglers, could significantly adversely impact job execution times. Moreover, poor scheduling of data analytics applications can lead to inefficient resource usage, and eventually hurt system performance. One way to mitigate stragglers is by launching extra attempts (clones) for each task upon job submission. In this paper, we propose Shed, an optimization framework that leverages dynamic cloning to jointly maximize jobs' Probability of Completion before Deadline (PoCD) by fully utilizing the available resources. Our work includes a novel online scheduler that dynamically recomputes and reallocates resources during a job's execution for PoCD maximization. The results show that Shed is able to leverage cloud resources and maximize the percentage of jobs that meet their deadlines - up to 100% in our experiments compared to typically around 60% and 40% for another cloning approach called Dolly, and Hadoop with speculation enabled, respectively.

Data centers have become a popular computing platform for various applications, and account for nearly 2% of total US energy consumption. Therefore, it has become important to optimize data center power, and reduce their energy footprint. With newer power- efficient design in data center infrastructure and cooling equipment, active components such as servers and the network consume most of the power with emerging sets of workloads. Most existing work optimizes power in servers and networks independently, and do not address them together in a holistic fashion that can achieve greater power savings. In this paper, we present PopCorns, a cooperative server-network framework for power optimization. We propose power models for switches and servers with low-power modes. We also design job scheduling algorithms that place tasks onto servers in a power-aware manner, such that servers and network switches can take effective advantage of low-power states. Our experimental results show that we are able to achieve more than 20% higher power savings compared to a baseline strategy that performs balanced job allocation across the servers.

Meeting desired application deadlines in cloud processing systems such as MapReduce is crucial as the nature of cloud applications is becoming increasingly mission-critical and deadline-sensitive. It has been shown that the execution times of MapReduce jobs are often adversely impacted by a few slow tasks, known as stragglers, which result in high latency and deadline violations. While a number of strategies have been developed in existing work to mitigate stragglers by launching speculative or clone task attempts, none of them provide a quantitative framework that optimizes the speculative execution for offering guaranteed Service Level Agreements (SLAs) to meet application deadlines. In this paper, we bring several speculative scheduling strategies together under a unifying optimization framework, called Chronos, which defines a new metric, Probability of Completion before Deadlines (PoCD), to measure the probability that MapReduce jobs meet their desired deadlines. We systematically analyze PoCD for popular strategies including Clone, Speculative-Restart, and Speculative-Resume, and quantify their PoCD in closed-form. The results illuminate an important tradeoff between PoCD and the cost of speculative execution, measured by the total (virtual) machine time required under different strategies. We propose an optimization problem to jointly optimize PoCD and execution cost in different strategies, and develop an algorithmic solution that is guaranteed to be optimal. Chronos is prototyped on Hadoop MapReduce and evaluated against three baseline strategies using both experiments and trace-driven simulations, and achieves 50% net utility increase with up to 80% PoCD and 88% cost improvements.

With the growing energy demands from server farms, it becomes necessary to understand the tradeoffs between energy consumption and application performance. Typically, server farms are provisioned for peak load even when they are mostly operating at low utilization levels. This results in wasteful energy consumption. At the same time, application workloads have Quality of Service (QoS) constraints that need to be satisfied. Optimizing server farm energy consumption with QoS constraints is a challenging task since the workload can have variabilities in job sizes, job arrival patterns and system utilization levels. In this paper, we present WASP, where we explore techniques that make smart use of the processor and system low-power states, and orchestrate their use with workload adaptivity for more effective energy management. We perform an extensive study of Energy-Latency tradeoffs with simulations, and evaluate WASP on a testbed with a cluster of servers. Our experiments on real systems show that WASP achieves up to 57% energy reduction over a naive policy that uses a shallow processor sleep state when there are no jobs to execute, and 39% over a delay timer based approach while maintaining the 90th percentile job service latency to be under 2x job execution time.

With the proliferation of the Internet of Things (IoT), the current "cloud-only" architectures cannot efficiently handle IoT's data processing and communications needs, while providing satisfactory service latency to support emerging mobile applications on the horizon that require almost real-time responses. fog computing is introduced as a new computing paradigm that distributes computation, communication, control, and storage closer to the end users along the "cloud- to-things" continuum. In this paper, we present a deadline-aware task scheduling mechanism for fog computing in a tiered IoT infrastructure, where service providers exploit the collaboration between their own fog nodes and the rented cloud resources to efficiently execute users' offloaded tasks, at large geographical scale. We first formulate the task-scheduling problem in such a cloud-fog environment as a multi-dimensional 0-1 knapsack problem that is NP-hard, and then propose an efficient algorithmic solution based on ant colony optimization heuristic. The main objective is to maximize the profits of fog service provider while meeting the tasks' deadline constraint. Extensive experimental results show that our proposed optimization and solution significantly improves the system performance compared with existing heuristics.

The rapid growth of Internet traffic has caused researchers continue to seek new ways to increase fiber bandwidth and spectrum utilization efficiency through elastic optical networking (EON). The advantage of EON comes from the fine-grained grid, which allows traffic demands to be better matched through a flexible allocation of fiber bandwidth. To further increase capacity, multiple fibers per link will be desired. Conventional optical crossconnects (OXCs) that use wavelength selective switches (WSSs) to switch the slots of a lightpath from input fibers to output fibers do not scale well. A more scalable and cost-effective node architecture called a flexible wavebanding crossconnect (FLEX) has been proposed recently. The FLEX architecture considerably reduces the cost of the crossconnect while introducing a small performance penalty in the form of reduced switching flexibility. In order to alleviate the limited switching capability, a cost-function-pluggable auxiliary layered-graph framework has also been proposed recently to solve the routing, fiber, waveband, and spectrum assignment (RFBSA) problem in multi-fiber EON with FLEX nodes. In this paper, we address the following problem. Given a budget in terms of the number of available WSSs for the network, determine the number and placements of FLEX nodes, and solve the RFBSA problem jointly in order to optimize network performance. We present an integer linear programming formulation, and propose a heuristic algorithm to solve this joint problem. The results show that our heuristic algorithm achieves good network performance, as measured by the average maximum spectrum usage (MSU), while saving significant hardware costs.

Data centers that run latency-critical workloads are typically provisioned for peak load even when they are operating at low levels of system utilization. Optimizing energy in data centers with Quality of Service (QoS) constraints is challenging since variabilities exist in job sizes, system utilization, and server configurations. Therefore, it is impractical to have a single configuration for energy management that works well across various scenarios. In this paper, we propose TS-Bat, a new data center energy optimization framework that judiciously integrates spatial and temporal job batching while meeting QoS constraints. TS-Bat works on commodity server platforms and comprises two major components: a temporal batching engine that batches the incoming jobs and creates opportunities for the processor to enter low power modes, and a spatial batching engine that schedules the batched jobs on to a server that is estimated to be idle. We implement a prototype of TS-Bat on a testbed with a cluster of servers, and evaluate TS-Bat on a variety of workloads. Our results show that pure temporal batching achieves 49% savings in CPU energy compared to a baseline configuration without batching. Through combining temporal and spatial batching, TS-Bat increases the energy savings by up to 68%.

The dramatic growth of Internet traffic brings challenges for optical network designers. There have been a number of advances recently in increasing fiber bandwidth and spectrum utilization efficiency through elastic optical networking (EON). In EON, a flexible and more fine-grained grid than conventional approaches is employed, and this allows allocated fiber bandwidth to better match traffic demands. Despite these advances, imminent fiber capacity exhaustion means that multiple fibers per link will be inevitable. In an effort to reduce the complexity of optical crossconnects, a flexible wavebanding crossconnect has been proposed recently. Elastic networking and flexible wavebanding introduce a new problem, namely, the routing, fiber, waveband, and spectrum assignment (RFBSA) problem. In this work, we propose new cost functions that are pluggable into an auxiliary layered-graph framework to solve the RFBSA problem with different objectives. We focus on minimizing the maximum spectrum usage for a set of traffic demands, and show that our approach outperforms traditional approaches.

Meeting desired application deadlines is crucial as the nature of cloud applications is becoming increasingly mission-critical and deadline-sensitive. Empirical studies on large-scale clusters reveal that a few slow tasks, known as stragglers, could significantly stretch job execution times. A number of strategies are proposed to mitigate stragglers by launching speculative or clone (task) attempts. These strategies often rely on a model-based approach to optimize key operating parameters and are prone to inaccuracy/incompleteness in the underlying models. In this paper, we present LASER, a deep learning approach for speculative execution and replication of deadline-critical jobs. Machine learning has been successfully used to solve a large variety of classification and prediction problems. In particular, the deep neural network (DNN), consisting of multiple hidden layers of units between input and output layers, can provide more accurate regression (prediction) than traditional machine learning algorithms. We compare LASER with SRQuant, a speculative- resume strategy that is based on quantitative analysis. Both these scheduling algorithms aim to improve Probability of Completion before Deadlines (PoCD), i.e., the probability that MapReduce jobs meet their desired deadlines, and reduce the cost of speculative execution, measured by the total (virtual) machine time. We evaluate and compare the two strategies through testbed experiments. The results show that our two strategies outperform Hadoop without speculation (Hadoop-NS) and Hadoop with speculation (Hadoop-S) by up to 89% in PoCD and 13% in cost.

As demands for cloud-based data processing continue to grow, cloud providers seek effective techniques that deliver value to the businesses without violating Service Level Agreements (SLAs). Cloud right-sizing has emerged as a very promising technique for making cloud services more cost-effective. In this paper, we present CRED, a novel framework for cloud right-sizing with execution deadlines and data locality constraints. CRED jointly optimizes data placement and task scheduling in data centers with the aim of minimizing the number of nodes needed while meeting users' SLA requirements. We formulate CRED as an integer optimization problem and present a heuristic algorithm with provable performance guarantees to solve the problem. Competitive ratios of the proposed algorithm are quantified in closed form for arbitrary task parameters and cloud configurations. We also extend our work to obtain a resilient solution, which allows successful recovery at run time from any single node failure and is guaranteed to meet both deadline and locality constraints. Simulation results using Google trace show that our proposed algorithm significantly outperforms existing heuristics such as first-fit by reducing the number of required active servers by up to 47 percent, and achieves near-optimal performance. We also show that our algorithm can significantly improve utilization of both computational resources and storage space by up to 28 and 15 percent, respectively.

Modern day data centers coordinate hundreds of thousands of heterogeneous tasks and aim at delivering highly reliable cloud computing services. Although offering equal reliability to all users benefits everyone at the same time, users may find such an approach either inadequate or too expensive to fit their individual requirements, which may vary dramatically. In this paper, we propose a novel method for providing elastic reliability optimization in cloud computing. Our scheme makes use of peer-to-peer checkpointing and allows user reliability levels to be jointly optimized based on an assessment of their individual requirements and total available resources in the data center. We show that the joint optimization can be efficiently solved by a distributed algorithm using dual decomposition. The solution improves resource utilization and presents an additional source of revenue to data center operators. Our validation results suggest a significant improvement of reliability over existing schemes.

As demands for cloud-based data processing continue to grow, cloud providers seek effective techniques that deliver value to the business without violating Service Level Agreements (SLAs). Cloud right-sizing has emerged as a very promising technique for making cloud services more cost-effective. In this paper, we present CRED, a novel framework for cloud right-sizing with execution deadlines and data locality constraints. CRED jointly optimizes data placement and task scheduling in data centers with the aim of minimizing the number of nodes needed while meeting users' SLA requirements. We formulate CRED as an integer optimization problem and present a heuristic algorithm with provable performance guarantees to solve the problem. Competitive ratios of the proposed algorithm are quantified in closed form for arbitrary task parameters and cloud configurations. Simulation results using Google trace show that our proposed algorithm significantly outperforms existing heuristics such as first-fit by reducing up to 47% of required active servers, and achieves nearly-optimal performance in terms of cloud-right sizing.

Optical data center network architectures are becoming attractive because of their low energy consumption, large bandwidth, and low cabling complexity. In [1], an AWGR-based passive optical data center architecture (PODCA) is presented. Compared with other optical data center architectures, e.g., DOS [2], Proteus [3], and Petabit [4], PODCA can save up to 90% on power consumption and 88% in cost. Also, average latency can be low as 9 s at close to 100% throughput. However, PODCA is not reconfigurable and cannot optimize the network topology to dynamic traffic. In this paper, we present a novel, scalable and flexible reconfigurable architecture called RODCA. RODCA is built on and augments PODCA with a flexible localized intra-cluster optical network. With the reconfigurable intra-cluster network, racks with mutually large traffic can be located within the same cluster, and share the large bandwidth of the intra-cluster network. We present an algorithm for DCN topology reconfiguration, and present simulation results to demonstrate the effectiveness of reconfiguration.

Optical interconnects for data centers can offer reduced power consumption, low latency, and high scalability, compared to electrical interconnects. However, active optical components such as tunable wavelength converters and Micro-Electro-Mechanical Systems (MEMS) switches suffer from high cost or slow reconfiguration times. In this paper, we propose three different Passive Optical Data Center Architectures (PODCAs), depending on the size of the network. Our key device is the Arrayed Waveguide Grating Router (AWGR), a passive device that can achieve contention resolution in the wavelength domain [1]. In our architectures, optical signals are transmitted from fast tunable transmitters and pass through couplers, AWGR, demultiplexers, and are received by wide-band receivers. Our architecture can easily accommodate over 2 million servers. Simulation results show that packet latency is below 9μs, and 100% throughput is achievable. We compare the power consumption and capital expenditure (CapEx) cost of PODCA with other recent optical data center network architectures such as DOS, Proteus, and Petabit. Results show that our architectures can save up to 90% on power consumption and 88% on CapEx.

Network-wide local unambiguous failure localization (NL-UFL) has been demonstrated as an interesting scenario of monitoring trails (m-trails). It attempts to enable every node to autonomously localize any failure event in the network in a distributed and all-optical manner by inspecting a set of m-trails traversing through the node. This paper investigates the m-trail allocation problem under the NL-UFL scenario by taking each link and node failure event into consideration. Bound analysis is performed using combinatorial group testing (CGT) theory and this is followed by the introduction of a novel heuristic on general topologies. Extensive simulation is conducted to examine the proposed heuristic in terms of the required cover length and the number of m-trails to achieve NL-UFL.

Server farms are becoming increasingly energy-hungry with the growing popularity of web-based applications and services. Servers consume nearly 60% of peak power even when operating at relatively low utilization levels of around 30%. Unfortunately, most server farms are generally provisioned to accommodate the peak load, and wasteful energy is often spent on unnecessarily keeping the servers active. Recent work on utilizing processor sleep states has mitigated the energy problem, but more opportunities to optimize energy remain to be explored. In this paper, we explore techniques that make smart use of processor deep sleep states through augmenting them with dual delay timers for more effective energy management in the multi-server environment. We find that our exploratory studies on smarter use of processor sleep states with dual delay timers show good promise in achieving higher energy savings on different kinds of synthetic and real workloads. Our experimental results show that our techniques achieve up to 71% savings in energy over naive energy management without the use of low-power sleep states, and up to 31% energy savings over a relatively smarter energy management mechanism with just a single delay timer to enter the sleep state. We also show that the normalized latency of jobs on a server farm with our dual delay timer strategy is almost similar to the one that is always ready to accept incoming jobs.

Survivability is a critical issue in high-capacity wavelength division multiplexing based optical networks, since even a disruption for a very short time could result in the loss of a large amount of data. Each accepted connection request brings revenue to the network. On the other hand, a service level agreement (SLA) requirement (e.g., 99.9% uptime in terms of availability) is associated with a connection, and a penalty is paid if the requirement is violated. In addition, physical impairments degrade the quality of signals as they traverse along lightpaths. In this paper, the survivable routing and wavelength assignment problem is studied for dynamic traffic in translucent optical networks considering randomly occurring failures of optical fibers and nodes, so that the quality of transmission (QoT) requirement of each connection is satisfied, and the network-level performance metric of total profit is maximized. Two heuristics are developed - the first one aims to allocate resources to each connection to minimize the expected penalty; while the other heuristic prefers to allocate the minimum possible resources to each connection. Simulation results are presented to demonstrate their effectiveness.

Creating wireless backhaul using a network of high altitude balloons is a promising solution for future high speed Internet access. Especially, it appears to be an easy and affordable solution in remote and rural areas thanks to its infrastructureless characteristics. Ease of deployment and affordability of such networks, however, comes at the price of complicated issues introduced by dynamic and complex topology of such networks. In this paper, we address the wireless backhaul reliability problem which arises due to limited control on horizontal mobility of the balloon transceivers. Despite the possibility of moving the balloons vertically, their overall trajectories are greatly dependent on the stratospheric wind which directly affects the availability of the links. In this paper, we show that using a few unmanned aerial vehicles (UAV) as relays can drastically improve the reliability of the wireless backhaul networks. We investigate the optimal paths that UAVs must follow to maximize the network reliability.

A novel coarse granular routing scheme for elastic optical networks is proposed in this paper together with a node architecture and network design algorithm. The proposed scheme bundles optical paths to be routed together and each bundle of paths, named flexible waveband, is routed as an entity. Path bundling is done by a small port count wavelength-selective switch (WSS) while flexible waveband routing is done by the other optical switches, i.e., two stage routing. The proposed network design algorithm resolves the routing and frequency slot assignment problem while considering specific constraints imposed by the routing scheme. Numerical experiments on several topologies confirm that the routing performance degradation caused by the coarse granular routing is small while the number of WSSs is substantially reduced.

Large scale optical cross-connects (OXCs) are required due to the increasing traffic demands. Currently, wavelength-selective switches (WSS) are utilized to create the OXCs. However, the port count of commercially available WSSs is limited. To achieve high port counts in OXCs, the existing WSS-based approach is to cascade WSSs, which results in a square order increment in the number of required WSSs. To save the hardware costs in terms of number of WSSs, two novel OXC architectures utilizing the waveband switching technique have been proposed. In this paper, we conduct a detailed comparison among the conventional cascading architecture and the two new architectures. We propose algorithms to accommodate dynamic traffic demands for all architectures and compare the blocking rates. The results show that the blocking rates of the novel architectures are very small and close to that of the cascading architecture, while the novel architectures have much less node complexity in terms of hardware requirement.

Today's applications such as cloud computing and e-science involve the processing of complex jobs consisting of several inter-dependent tasks executing on heterogeneous clusters of computing resources, which are interconnected by high-speed optical networks. The emerging technology of flexible grid through the use of Optical Orthogonal Frequency-Division Multiplexing (OOFDM) allows fiber bandwidth to be more suitably matched up with application requirements, thereby making the network more elastic. This is done by partitioning the bandwidth into hundreds or even thousands of OFDM subcarriers that may be allocated to services. An important problem in such applications is the joint scheduling (or co-scheduling) of computational and network resources. In this paper, we formulate a problem of co-scheduling computational and networking resources to multiple jobs in elastic optical networks. We consider both static and dynamic versions of the problem; in the static case, our objective is to minimize the makespan of all the jobs, while minimizing the job blocking is the aim when jobs arrive dynamically. We formulate an integer-linear program for the static version of the problem. Two efficient heuristics are then proposed and compared. Simulation results are presented to demonstrate the effectiveness of the proposed approaches.

Advances in data intensive computing and high performance computing facilitate rapid scaling of data center networks, resulting in a growing body of research exploring new network architectures that enhance scalability, cost effectiveness and performance. Understanding the tradeoffs between these different network architectures could not only help data center operators improve deployments, but also assist system designers to optimize applications running on top of them. In this paper, we present a comparative analysis of several well known data center network architectures using important metrics, and present our results on different network topologies. We show the tradeoffs between these topologies and present implications on practical data center implementations.

Network-wide local unambiguous failure localization (NL-UFL) [1] has been demonstrated as an interesting scenario of monitoring trails (m-trails). It attempts to enable every node to autonomously localize any failure event in the network in a distributed and all-optical manner by inspecting a set of m-trails traversing through the node. This paper investigates the m-trail allocation problem under the NL-UFL scenario by taking each link and node failure event into consideration. Bound analysis is performed using combinatorial group testing (CGT) theory and this is followed by the introduction of a novel heuristic on general topologies. Extensive simulation is conducted to examine the proposed heuristic in terms of the required cover length and the number of m-trails to achieve NL-UFL.

The articles in this special issue focus on energy efficiency techniques deployed in optical fiber networking.

Savings in switching costs of an optical cross-connect can be achieved by grouping together a set of consecutive wavelengths and switching them as a single waveband. This technique is known as waveband switching. While previous work has focused on either uniform band sizes or nonuniform band sizes considering a single node, in this paper we focus on the number of wavebands and their sizes for ring topologies. First, we show that such solutions are inadequate when considering the entire network. We then present a novel framework for optimizing the number of wavebands in a ring network for deterministic traffic. The objective of the Band Minimization Problem is to minimize the number of nonuniform wavebands in the network while using the minimum possible number of wavelengths. We show that the problem is NP-hard and present heuristics for it. We then consider a specific type of traffic, namely all-to-all traffic, and present a construction method for achieving the minimum number of wavebands in the ring. Our results show that the number of ports can be reduced by a large amount using waveband switching compared to wavelength switching, for both all-to-all traffic and random traffic. We also numerically evaluate the performance of our waveband design algorithms under dynamic stochastic traffic.

One of the major problems facing optical networking is to intelligently assign physical resources such as lightpaths and regenerators to connection requests, namely the grooming, routing and wavelength assignment (GRWA) problem. Due to the high computation complexity, many heuristic methods have been proposed to solve the problem. We first propose an extended Dijkstra shortest path algorithm to solve GRWA for dynamic network while considering regeneration, quality of transmission (QoT), mixed-line-rate (MLR) and traffic grooming. This generalized adaptive shortest path (GASP) algorithm requires that each node maintain global view of network state information. In order to perform GRWA in a distributed fashion, which allows for greater network scalability, and gives individual domains more control over their data, we then apply an ant colony optimization (ACO) technique, a metaheuristic optimization algorithm used to solve dynamic problems, to real-time optical networks. We compare the two proposed algorithms and show that the ACO algorithm outperforms the GASP algorithm in terms of connection request blocking probability and network throughput while maintaining a reasonable computation complexity.

Free Space Optics (FSO) relays can be used to recover a network which is partially disconnected due to natural disasters or terrorist attacks. Rapid and efficient recovery can be achieved thanks to FSO technology being wireless and providing high bandwidth. However, placement of such relays is a challenging problem as FSO links greatly depend on weather conditions. In this paper, we find the minimum number of transceivers and their optimal placement which guarantees the recovery of a certain fraction of network capacity in the worst weather conditions and maximizes the throughput in the best weather conditions through transceiver reconfiguration. The problem is formulated as an integer linear program (ILP) which takes the link availability prediction as an input and guarantees fairness to all existing traffic flows. To avoid the complexity of the ILP, an efficient probabilistic heuristic that computes the placement of FSO transceivers is proposed. We show through extensive simulations that the heuristic performs within 12% of the optimal performance.

Virtualization improves the efficiency of networks by allowing multiple virtual networks to share a single physical network's resources. Next-generation optical transport networks are expected to support virtualization by accommodating multiple virtual networks with different topologies and bit rate requirements. Meanwhile, Optical Orthogonal Frequency-Division Multiplexing (OOFDM) is emerging as a viable technique for efficiently using the optical fiber's bandwidth in an elastic manner. OOFDM partitions the fiber's bandwidth into hundreds or even thousands of OFDM subcarriers that may be allocated to services. In this paper, we consider an OOFDM-based optical network and formulate a virtual network mapping problem for both static and dynamic traffic. This problem has several natural applications, such as e-Science, Grid, and cloud computing. The objective for static traffic is to maximize the subcarrier utilization, while minimizing the blocking ratio is the aim for dynamic traffic. Two heuristics are proposed and compared. Simulation results are presented to demonstrate the effectiveness of the proposed approaches.

Despite the increasing popularity of Free-Space Optics (FSO) in wireless mesh networks, reliability is still a major concern. Due to the strong dependence of FSO link availability on weather conditions (e.g., fog, cloud, air turbulence) using RF links is inevitable, both for backup and signaling. In addition, reconfiguration of FSO transceivers could preserve network reliability when several connections suffer outages. In this paper, we find the optimal reconfiguration of transceivers and rerouting of traffic flows using an integer-linear program (ILP). The link availability prediction is used as an input and the overall throughput of the network is maximized while guaranteeing fairness to all traffic demands. To avoid the complexity of the ILP, an efficient probabilistic heuristic that computes a set of schedules for dynamic reconfiguration of FSO transceivers is proposed. We show that this heuristic achieves a high fraction of the optimal throughput, as verified by extensive simulations.

Translucent fiber-optic networks are carefully planned to achieve high capacity utilization efficiency as required by society's ever-increasing traffic demand. Existing research treats the problem of resource placement largely as a static design problem, which is solved with linear programming (LP) to find the optimal solution. The dynamic operational problem (grooming, regeneration, routing, and wavelength assignment) is approached using heuristic methods with the goal of improving the overall network performance given an existing network infrastructure. Our work combines these two approaches and solves a real-time dynamic traffic scenario with integer linear programming (ILP), seeking to maximize the overall network throughput. The traffic is served in a time-slotted fashion so that the network throughput is optimized at each time slot given the existing network state. The solution is compared with results from existing heuristic methods. We incorporate physical impairment limitations into our network model, and consider several grooming options.

Modern day data centers coordinate hundreds of thousands of heterogeneous tasks and aim at delivering highly reliable cloud computing services. Although offering equal reliability to all users benefits everyone at the same time, users may find such an approach either too inadequate or too expensive to fit their individual requirements, which may vary dramatically. In this paper, we propose a novel method for providing reliability as an elastic and on-demand service. Our scheme makes use of peer-to-peer checkpointing and allows user reliability levels to be jointly optimized based on an assessment of their individual requirements and total available resources in the data center. We show that the joint optimization can be efficiently solved by a distributed algorithm using dual decomposition. The solution improves resource utilization and presents an additional source of revenue to data center operators. Our validation results suggest a significant improvement of reliability over existing schemes.

Wavelength Division Multiplexing (WDM)-based optical networks are ideal candidates for core backbone networks because of their ability to carry large amounts of traffic. Optical amplification has increased the reach of long-haul optical links. Nevertheless, it is impossible today to construct a truly optical network without converting optical signals to electrical signals and regenerating them, because of the deleterious effects of physical impairments such as amplifier noise, dispersion, and non-linear effects such as four-wave mixing and cross-phase modulation. Being expensive devices, these regenerators are expected to be sparsely located and used in such a network called as a translucent optical network. In this paper, we consider the routing and wavelength assignment (RWA) problem so that the Quality of Transmission (QoT) for connections is satisfied, and the network-level performance metric of blocking probability is minimized. Cross-layer heuristics that are based on dynamic programming to effectively allocate the sparse regenerators are developed, and extensive simulation results are presented to demonstrate their effectiveness.

Advances in optical transmission technology are driving the evolution from legacy opaque toward transparent backbone networks. Aiming at reducing the costs and energy consumption, transparent networks allow the data to be kept optical end to end and thus removing the majority of the expensive optical-electrical-optical (OEO) devices. Nonetheless, physical impairments accumulate on the signal along optical transparent paths, therefore limiting the system reach and the overall network performance. Furthermore, failures propagate in a transparent network environment and they cannot be easily localized and isolated. So then, transparent networks need to deal with a number of challenges impacting both planning and operation phases. Such challenges require the use of cross-layer approaches, which involve dynamic interactions between the physical layer and the network layer to enable the compensation for mismatching of requirements and resources. In this paper, we aim to provide an overview of such challenges, reporting comparative analysis with a selection of existing solutions and to cast a glance at the open issues for future research.

Wireless mesh networks play a key role in the overall performance of hybrid wireless optical access networks. In many cases, the wireless multihop network causes a bottleneck and under- utilizes the network resources. In this paper, we propose a new approach to increase the wireless network capacity by enhancing the congested links, normally those which are closer to the gateways, using MIMO antennas. We are mainly interested in determining the links which provide the network with the highest improvement if they get enhanced. We suggest a TDMA-based framework for link activation to avoid interference, and the problem is formulated as an integer linear program. Based on the optimal solution, we assign a few extra antenna modules to some critical transceivers to significantly increase the entire capacity of the network. We also propose a simple routing algorithm based on the optimal schedule which handles the variations of traffic demand efficiently.

Wireless mesh network addresses the last mile issue; however, RF-based technologies are still a bottleneck for todays' media-rich applications. In this paper, we propose enhancing the capacity of an existing wireless mesh backhaul using free-space optical links. Such a hybrid RF/FSO network is capable of multiplying the available bandwidth for each subscriber if the FSO transceivers are placed at strategically located nodes. Adopting a TDMA- based framework, we formulate a problem for the optimal placement of FSO transceivers and RF link activation schedules which maximize the network capacity. An integer-linear program (ILP) is developed to obtain an optimal solution. Due to the complexity of this ILP, we also propose a relaxation which yields an upper bound on the network capacity in such an architecture, and develop a probabilistic heuristic that approaches this capacity. We show by simulations that our heuristic achieves a high fraction of the network capacity upper bound, and it can be implemented efficiently.

Transparent all-optical networks introduce a challenging problem of achieving efficient and accurate full-coverage fault-localization. In this paper we present a novel and efficient monitoring approach that exploits the benefits of provisioned user lightpaths for in-band monitoring and achieves complete fault localization coverage at the minimum resource cost through the use of complementary supervisory lightpaths. We formulate the problem and present an Integer Linear Program (ILP) solution. An efficient heuristic for provisioning complementary supervisory paths and cycles is also introduced and the solution quality is compared with ILP for different sets of random user lightpaths and different parameters. Extensive numerical comparisons with several recently proposed monitoring algorithms demonstrate the advantages of the proposed approach.

Location-based services (LBS) have become an immensely valuable source of real-time information and guidance. Nonetheless, the potential abuse of users' sensitive personal data by an LBS server is evolving into a serious concern. Privacy concerns in LBS exist on two fronts: location privacy and query privacy. In this paper we investigate issues related to query privacy. In particular, we aim to prevent the LBS server from correlating the service attribute, e.g., bar/tavern, in the query to the user's real-world identity. Location obfuscation using spatial generalization aided by anonymization of LBS queries is a conventional means to this end. However, effectiveness of this technique would abate in continuous LBS scenarios, i.e., where users are moving and recurrently requesting for LBS. In this paper, we present a novel query-perturbation-based scheme that protects query privacy in continuous LBS even when user-identities are revealed. Unlike most exiting works, our scheme does not require the presence of a trusted third party.

Resiliency is a crucial performance measure for high-speed all-optical networks. Previous work has shown that any algorithm designed for such purpose should take into account the physical layer characteristics of these environments, in other words it must employ a cross-layer design. In particular, cross-layer p-Cycle design has been developed as a desirable approach. In this paper, we propose a classification of network traffic resiliency requirements and present a multi-class link protection scheme. In particular we allow pre-computed but not pre-crossconnected protection routes for some portion of the network traffic. We evaluate the performance of these algorithms in the face of physical layer impairments (PLIs) and show that this classification of traffic can result in higher resiliency or, equivalently, lower vulnerability ratio in the network.

An all-optical approach to achieve finer bandwidth granularity is to time division multiplex low capacity circuits on each wavelength channel and to switch time-wavelength slots within the network. In such a Time-Wavelength-Switched- Network (TWSN), the Time-Wavelength-Space-Routers (TWSRs) are configured to change their routing pattern on a time slot basis. Another kind of time slotted network proposed in the literature is the Time-Wavelength-Interleaved Network (TWIN), which eliminates time switching within the network by using a non-reconfigurable core and an intelligent edge utilizing a fast tunable laser to emulate fast switching. A drawback of the TWIN network is that it assigns a unique wavelength to each node in the network. Thus it requires a total of W = N wavelengths for an N-node network and hence is not scalable. In this paper we propose to design a wavelength-constrained (i.e., W <; N) TWIN network with no switching (TWIN-NS) by using a multicasting strategy. We also propose a variant of the TWIN network which possesses switching capabilities only at the edge nodes (TWIN-ES), and compare the performances of these TWIN networks to that of the TWSN. We present integer linear programs and heuristic algorithms to solve the connection scheduling problem in the TWSN and the TWIN, and investigate the benefits of having a fast reconfigurable switch as opposed to a non-reconfigurable core.

Waveband switching saves port costs in optical crossconnects by grouping together a set of consecutive wavelengths and switching them as a single waveband. Previous work has focused on either uniform band sizes or non-uniform band sizes considering a single node. In this paper, we show that such solutions are inadequate when considering the entire network, and present a novel framework for optimizing the number of wavebands in a ring network for deterministic traffic. We then consider a specific type of traffic, namely, all-to-all traffic and present bounds and heuristic solutions for the problem. Our results show that the number of ports can be reduced by a large amount using waveband switching compared to wavelength switching. We also numerically evaluate the performance of our waveband design algorithms under dynamic stochastic traffic.

Wireless infrastructure networks that provide ubiquitous connectivity to mobile nodes in metro areas are becoming increasingly popular. The nodes in such backbone networks are often equipped with multiple transceivers to allow for simultaneous transmissions in multiple non-overlapping channels. In this paper, we develop an analytical model for evaluating the throughput and delay performance of wireless infrastructure networks employing Slotted Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) over multiple channels. The analytical model, which takes into account the correlation due to multi-hop transmissions, approximates the performance observed through simulations accurately.

Achieving accurate and efficient fault localization in large transparent all-optical networks (TONs) is an important and challenging problem due to unique fault-propagation, time constraints, and scalability requirements. In this paper, we introduce a novel technique for optimizing the speed of fault-localization through the selection of an active set of monitors for centralized and hierarchically-distributed management. The proposed technique is capable of providing multiple levels of fault-localization-granularity, from individual discrete optical components to the entire monitoring domains. We formulate and prove the NP-completeness of the optimal monitor activation problem and present its Integer Linear Program (ILP) formulation. Furthermore, we propose a novel heuristic whose solution quality is verified by comparing it with an ILP. Extensive simulation results provide supporting analysis and comparisons of achievable alarm-vector reduction, localization coverage, and time complexity, for flat and hierarchically distributed monitoring approaches. The impact of network connectivity on fault localization complexity in randomly generated topologies is also studied. Results demonstrate the effectiveness of the proposed technique in efficient and scalable monitoring of transparent optical networks.

Their high recovery speed and efficiency have made p-Cycles a premier approach to link protection. Recent research has discussed their capabilities and provided different methods for p-Cycle selection in communication networks. In this work, we consider the p-Cycle approach to resilient all-optical network design and present an ILP formulation that preconfigures p-Cycles in such an environment. Our proposed formulation strives to find p-Cycles that have good recovery capability in the face of physical layer impairments (PLIs) and are capacity-efficient at the same time. Our numerical results show a trade-off between these two performance measures. We evaluate the performances of different outcomes of the formulation using our accurate physical layer model. We show that applying our previously proposed cross-layer RWA algorithm, HQ, greatly improves the performance.

Link protection schemes for WDM-based optical networks have been extensively researched. Most work to date has ignored the physical layer impairments (PLIs) that could be dominant in transparent optical networks with long links. In this paper, we first evaluate the performance of two link protection schemes - namely, p-cycles and generalized loopback - when PLIs are considered. Our evaluation shows that a choice of p-cycles that merely tries to optimize the wavelength usage can be more susceptible to failures in a realistic environment. In particular, we show that the Hamiltonian p-cycle is not the optimal choice for p-cycle selection in all-optical networks. This is mainly due to the interference caused by the long lightpaths that appear in the network after a failure occurs. We show that although a selection of smaller p-cycles can cause higher blocking probability, it is less vulnerable to failures. We also compare the performance of p-cycles to the generalized loopback link cover approach in these environments. Finally, we apply a cross-layer routing and wavelength assignment algorithm to these schemes that significantly enhances performance in physically impaired optical networks.

Time-Wavelength-Switched Networks (TWSNs) provide a finer bandwidth granularity than traditional wavelength routing networks. In TWSNs, time on every wavelength is slotted and the crossconnects are configured to switch slots within a frame, and connections are assigned one or more slots per frame. Fiber Delay Line (FDL) banks within the crossconnects can reduce the probability of connection blocking due to output slot contention. In this paper, we consider the optimal construction of an FDL bank with a limited number of recirculations through the FDLs. Specifically, given a frame consisting of M slots, an FDL bank of D FDLs per output link, we develop a method to find a set of delay values D={d1, d2,...dD} such that the set of achievable delays using no more than K recirculations through the bank is maximized. Maximizing the number of achievable delays provides more flexibility in slot provisioning for connections, and is expected to decrease the blocking probability. We investigate the impact of the FDL configuration on a CrossConnect's performance. Our results show that the optimal FDL bank construction presented here achieves better blocking than previously considered FDL bank configurations. We also present a solution to the FDL bank design problem for an optical packet switch, and show through simulations that packet drop probability is considerably reduced for the optimal configuration.

Wavelength routed optical networks can experience increased call blocking due to insufficient quality of transmission. The routing and wavelength assignment algorithm must verify the quality of the lightpath before accepting it. In this paper, analytical expressions for the total blocking probability are derived for first fit wavelength assignment for networks suffering from transmission impairments. Physical layer degradations considered include amplifier noise and crosstalk between WDM channels. The technique effectively predicts the performance for wavelength selection techniques that consider a single candidate channel or all channels for quality of transmission compliance. The analysis is also applicable to first-fit algorithms with different static channel orderings.

Protection against failure in high bandwidth optical networks can prevent significant potential losses of data but at the same time it requires a considerable fraction of the network resources to be reserved for it. Moreover, in optical networks suffering from physical layer impairments, inefficient use of network resources leads to significant increase in the blocking probability and higher vulnerability due to degradation of the QoT (quality of transmission) in the network. These trade offs call for the need to consider the physical layer impairments in designing higher layer protection schemes, i.e., a QoT-aware or cross-layer design. In this paper, we look at the performance of several link protection and link and path restoration algorithms in all-optical networks with realistic physical layer impairments, and propose a new cross-layer restoration method that exhibits both low blocking probability and low vulnerability ratio.

An all-optical approach to reduce the speed mismatch between electronic sources and high-speed wavelength channels is to time-division multiplex low-capacity circuits onto each wavelength, and switch time-wavelength-slots within the network. In such a Time-Wavelength-Switched Network (TWSN), fiber delay lines (FDLs) and wavelength conversion (WC) can increase the flexibility in connection slot scheduling and decrease the connection blocking probability. In this paper, we examine the impact of various FDL configurations and WC on a CrossConnect's performance by looking at its ability to schedule as many connections as possible from a given traffic matrix. We present several crossconnect architectures and develop graph formulations that can be used to optimally solve the scheduling problem. Using numerical simulation results, we then compare several FDL configurations. Results show that FDL configurations and WC could play an important role in determining performance.

Rapid fault detection and localization in transparent optical networks is crucial due to the high data rates at which they operate and optical transparency. Furthermore, large all- optical networks require distributed fault-management in order to achieve scalable, accurate, and rapid fault localization. In this work we present an efficient scheme for hierarchically distributed monitoring and fault-localization in transparent all- optical networks. We prove that the proposed scheme yields identical fault-localization capability as the centralized and flat fault-localization scheme, while allowing for distributed optimization of alarm-vector size and fault-localization. Simulation results provide supporting comparison of achievable fault localization capability and effective reduction of alarm- vector lengths and fault localization complexity for flat and hierarchically distributed monitoring approaches.

An important factor that affects the cost of WDM networks is the number of wavelengths to be added/dropped by reconfigurable optical add drop multiplexers (ROADMs). Limiting the range of wavelengths that can be accessed at a node reduces costs. In this paper, we study wavelength assignment (WA) in WDM bi-directional ring networks with limited reconfigurability and take up the problem of supporting two traffic types - permutation traffic and all-to-all traffic - while minimizing the worst-case wavelength range of the ROADMs and respecting the wavelength termination and continuity constraints. ROADMs can be limited in tuning range (L-ROADMs) or have full range (F- ROADMs). In this work, a node can either possess a limited-add L-ROADM or an F-ROADM. For permutation traffic, we show that at least one F-ROADM is necessary, and we also give an upper bound on the number of F-ROADMs needed as a function of the range of L-ROADMs at other nodes. For all-to-all traffic, we develop an integer linear programming (ILP) formulation, and also propose a new wavelength assignment algorithm that uses Wu= N2/4 - N/2 + 1/4 wavelengths. Our heuristic's results are similar to the ILP's. The results show that a saving of 85-93.3% in tuning range is achievable for this type of traffic.

In WDM all-optical networks where electrical regeneration is not available, physical impairments due to propagation in the fibers, amplifier noise, and leaks between channels and in the switches cannot be removed at the physical layer. These effects cause calls, especially between physically distant nodes, to be rejected because they cannot meet minimum Quality of Transmission (QoT) requirements, as measured by signal bit-error rates. It is possible to mitigate physical layer effects at the network layer using appropriate Routing and Wavelength Assignment (RWA) algorithms. We present new RWA algorithms which account for physical impairments in their design and increase QoT and fairness among users without sacrificing low blocking probabilities in metropolitan-sized networks. We also present RWA algorithms that can sharply decrease blocking probabilities in regional-sized networks using optional channel coding. All algorithms are evaluated through simulation in realistic scenarios and shown to successfully mitigate crosstalk effects and to perform better in terms of QoT and network access fairness than traditional algorithms.

As an optical signal propagates along a lightpath to its destination in wavelength-routed optical networks (WRONs), the quality of transmission (QoT) is degraded by transmission impairments such as crosstalk and amplified spontaneous emission noise. Consequently, the signal’s bit error rate at the destination’s receiver can become unacceptably high. Recently, many BER-aware connection provisioning algorithms have been proposed that incorporate physical layer impairments in terms of increased blocking rate from the BER constraint. However, they have a high computational complexity compared to traditional connection provisioning schemes because of the complexity of BER estimation. The delay involved in the computation greatly affects the performance in distributed networks because the latency worsens the contention and makes instantaneous network status impossible to obtain. In this paper, distributed connection provisioning schemes are compared using two BER estimation procedures and several wavelength assignment algorithms. Simulation results show the effects of delay in connection provisioning and measure the performance in terms of blocking probability.

Rapid fault identification and localization in optical networks are crucial due to high data rates. These problems are more challenging than in traditional electronic networks because of optical transparency. In a transparent optical network which does not regenerate optical signals, a fault may propagate to various parts of the network from the origin, and multiple alarms can be generated for a single failure. In order to reduce the number of redundant alarms, simplify fault localization, and reduce the fault localization time, the number and location of fault monitors that are turned on should be optimized for a given network. In this paper, we formulate a problem on the optimal activation of network monitoring devices and propose a solution approach. First, we provide a brief summary of available physical-layer monitoring devices, and then present a scheme for optimal monitor activation. We show the NP-completeness of the problem and present a mixed-integer-linear-program (MILP) formulation of it. We also present a heuristic, whose performance is evaluated through comparisons with the solutions to the MILP, as well as a naive monitor activation approach.

Reconfigurability of optical networks is a very important feature for supporting dynamic traffic demands and to enable new services. Optical network reconfigurability is enabled by reconfigurable optical add drop multiplexers (ROADMs) and tunable transceivers. Limiting the range of wavelengths that can be accessed at a node reduces costs, and possibly reduces flexibility in reconfigurability. The limited range (band-size) of wavelengths is a result of the limited add/drop range of the ROADMs and/or limited tuning range of transceivers. In either case, limited reconfigurability introduces a new constraint in the wavelength assignment of connections called as the wavelength termination constraint. In this paper, we consider a uni-directional ring network with limited reconfigurability and take up the problem of supporting two traffic types - permutation traffic and all-to-all traffic - while minimizing the worst-case wavelength range of the ROADMs and respecting the wavelength termination and continuity constraints. ROADMs can be limited in range (L-ROADMs) or have full range (F-ROADMs). For permutation traffic, we show that at least one F-ROADM is necessary, and we also give an upper bound on the number of F-ROADMs needed as a function of the range of the other nodes that are L-ROADMs. For all-to-all traffic, we give an integer linear programming formulation and present a heuristic algorithm to minimize the worst-case range. Our results show that all ROADMs can be L-ROADMs and the worst-case range is only about 65% of the full range for large network sizes.

In this paper, we consider the problem of routing and wavelength assignment (RWA) in all-optical networks, where signals propagate in the optical domain (with no electrical regeneration) from end-to-end. In such networks, physical impairments are an issue and can cause calls to be blocked if a minimum quality of transmission (QoT) measured in terms of bit-error rate (BER) cannot be guaranteed for all calls, at all times. We focus on the routing component of RWA to incorporate QoT: we propose a novel RWA algorithm that finds a route based on both the network utilization and the physical impairments experienced over the tentative route. Our algorithm is fully decentralized and is shown using simulations to perform better than previously proposed algorithms on a regional-sized network.

Crosstalk originating from optical switching devices and demultiplexers can be the dominant physical impairment in large all-optical networks. When a centralized network controller must estimate the impact of crosstalk and other physical impairments on the quality of a lightpath, this can add significant latency. In this paper, QoS-aware wavelength assignment algorithms are proposed that consider both bit error rate and latency guarantees. A technique called wavelength ordering is presented to alleviate physical impairments and also decrease the processing delay due to bit-error-rate (BER) estimation. Simulations presented show that the wavelength ordering algorithm outperforms other wavelength assignments when both BER and response latency guarantees are enforced.

As an optical signal propagates along a lightpath to its destination in distributed wavelength-routed optical networks, the signal's quality is degraded by transmission impairments, such as node crosstalk, which is induced by power leaking in the network nodes. Consequently, the bit error rate of the received signal at the destination can be unacceptably high. In this paper, a new adaptive wavelength assignment scheme based on wavelength spectrum separation technique is proposed to combat physical impairments in distributed optical networks. The proposed algorithm outperforms traditional techniques as measured by the blocking probability, adding little complexity.

There has been a significant amount of recent research on routing and wavelength assignment for DWDM networks suffering from physical layer transmission impairments. These algorithms attempt to increase the quality of transmission or service and are called QoT (quality of transmission) or QoS aware algorithms. However, protection has not been considered in this context so far, as far as we are aware. In this paper, we investigate the effect of physical layer impairments on dedicated path protection schemes. In dedicated path protection, every connection has resources reserved and dedicated on both a primary and a backup lightpath. While there is no difference in network performance whether the backup path is lit or kept dark in networks that are not transmission-impaired, lighting the backup path has an adverse effect on the network in transmission-impaired networks. By considering two different RWA algorithms - a QoT-aware one and a QoT-unaware one, we study the blocking performance and the vulnerability of the connections to failures for the two cases - dark backup and lit backup. Our results show that there are significant penalties in backup lit case, and that the QoT-aware algorithm considerably outperforms the QoT-unaware algorithm in terms of both blocking and vulnerability to failures.

The blocking performance of limited-reconflgurable all-optical networks is investigated in this paper. Reconfigurability is achieved by reconflgurable optical add-drop multiplexers (ROADMs) and tunable transponders. An analytical model is developed adopting a share-per-node model for the sharing of transponders within a reconfigurable node. The set of wavelengths that can be added/dropped at a reconfigurable node depends on the constraints of ROADMs as well as the (possibly limited or narrowly) tunable transponders. The size of the wavelength set to which a transponder is tunable (tuning range), and the number of transponders are the key parameters that affect the blocking. A lightpath request can be established depending on the transponders' ability to tune to the available wavelengths along the route. We call this constraint as wavelength termination constraint. Both analytical and simulation results show that narrowly tunable transponders achieve a similar performance as widely tunable transponders. The effect of the number of transponders or the number of ports of a ROADM on blocking is also investigated.

In all-optical networks with no wavelength converters, signals are switched optically inside the nodes and therefore propagate over hundreds or thousands of kilometers with no electrical regeneration. Over such distances, physical impairments accumulate and can lead to serious signal degradation, resulting in poor quality of transmission (QoT) as measured by signal bit-error rates. The role of routing and wavelength assignment (RWA) algorithms is to accommodate incoming calls in optical networks over a route and a wavelength. RWA algorithms block calls if a continuous wavelength from source to destination cannot be found (wavelength blocking), or when the QoT of the call is not acceptable (QoT blocking). In this paper, we present an analytical method to evaluate blocking probability in all-optical networks, accounting for several physical layer impairments: intersymbol interference (ISI), amplifier noise (both are static effects that only depend on the network topology only) and node crosstalk (a dynamic effect that depends on the network status). We successfully validate our model through simulations on large scale networks with realistic physical layer parameters.

In this paper, we investigate the blocking performance of all-optical ring and mesh networks with Reconfigurable Optical Add-Drop Multiplexers (ROADMs) and tunable transponders. The ROADMs and transponders together determine the number and set of wavelengths that can be added/dropped leading to a wavelength termination constraint in addition to the well-known wavelength continuity constraint. We develop an analytical model for the blocking probability by accounting for both constraints and validate the model using simulation results. Specifically, a model for computing the blocking probability in a reconfigurable network of arbitrary topology wherein the number and tunability of the transponders are given parameters is presented. It is found that narrowly tunable transponders with a tuning range of about 4 in a 32-wavelength network provides nearly as good a performance as widely tunable transponders (tunable over the entire range of wavelengths), for a wide range of loads and number of transponders. Moreover, waveband assignment to narrowly tunable transponders is found to be a factor in some cases, and we present some results for two different assignment schemes.

In all-optical networks with no wavelength conversion, signals must travel on the same wavelength over possibly very long distances. During transmission, the QoS of signals as measured by their Bit Error Rates is degraded not only by the propagation through fibers, but also by small optical leaks from other signals called crosstalk that occur in the nodes and cannot be removed at the physical layer. We present a set of Routing and Wavelength Assignment algorithms that mitigate the crosstalk effects on all-optical network operation. These algorithms incorporate QoS information at both the routing and the wavelength assignment steps and account for dynamic crosstalk to yield better performance in terms of average BER and fairness among network users without sacrificing blocking probabilities, as shown through simulation.

We consider an IP/MPLS over WDM network, in which label switched routers (LSRs) in the IP/MPLS layer are interconnected through optical cross-connects (OXCs) in the optical core network (WDM layer) providing an end-to-end wavelength routing capability. In this paper, we study a dynamic label switched path (LSP) routing problem for the three different network models of the IP/MPLS over WDM network, namely, Overlay, Augmented, and Peer models. For the overlay model, we propose two algorithms: ECF_OVLY and MLH_OVLY. In ECF_OVLY, a network always tries to use existing capacity first, whereas in MLH_OVLY, a network finds a path with the minimum number of logical hops for an LSP request. We also propose, for the augmented model, two simple and efficient dynamic LSP provisioning algorithms, called MCPI_AUG and DCPI_AUG, utilizing different type/amount of summarized capacity information from the WDM layer, namely Minimum CaPacity Information (MCPI) and Detailed CaPacity Information (DCPI). We compare the proposed algorithms with the existing algorithms available for the overlay [1] and peer [2], [3] models. The algorithms are compared and evaluated using two key performance measures: LSP blocking probability and network (lightpath) utilization. Simulation results show that at low loads with a limited number of ports available in the network, DCPI_AUG achieves an order of magnitude better blocking performance than the algorithm in [2] and outperforms the one in [3] by more than three times. It also achieves higher network utilization than the one in [2] by more than 10 % and the one in [3] by 2-7 % depending on the traffic load. Considering the small amount of information that is exchanged between the layers in the augmented model, these results suggest that the augmented model can be a practically good compromise between the overlay and peer models.

In transparent DWDM (Dense Wavelength Division Multiplexing) networks, high blocking rate of lightpaths due to unsatisfactory BERs (Bit Error Rates) can become a major traffic bottleneck. Most of the existing routing schemes have focused on reducing the call rejection rate that is caused by wavelength exhaustion on the assigned routes, without considering the inferior network BER performance that these schemes often produce. In this paper, we propose a dynamic AQoS (Adaptive Quality of Service) routing algorithm, which assigns routes based on real-time Q factor measurements collected from devices. Not only is it BER-friendly, the incorporated CLC (Constrained Least Congested) approach in route selection makes our algorithm wavelength-efficient as well. AQoS routing can work seamlessly with variable data rates and heterogeneous device characteristics. More importantly, it handles non-uniform networks and state changes in a much more flexible way than traditional schemes. Simulation results show that our routing algorithm gives the best overall performance in all tested scenarios.

The inherent complexity of analysis of a multi-hop ad hoc network together with the fact that the behavior of a node is dependent not only on its neighbors' behavior, but also on the behavior of other unseen nodes makes multi-hop network analysis extremely difficult. However, our approach in this paper to analyze multi-hop networks offers an accurate approximation with moderate complexity. Our approach is based on characterizing the behavior of a node by its state and the state of the channel it sees. This approach is used to carry out an analysis of single-hop and multi-hop ad hoc networks in which different nodes may have different traffic loads. In order to validate the model, it is applied to IEEE 802.11-based networks, and it is shown through extensive simulations that the model is very accurate.

In-band crosstalk has been widely considered as a major transmission impairment that significantly impacts the BER (bit error rate) performance of lightpaths in circuit-switched all-optical wavelength-routed networks. Such crosstalk usually occurs when multiple lightpaths occupying identical or adjacent wavelengths pass through an optical crossconnect node. Traditional WA (wavelength assignment) schemes pay little regard to the physical layer QoS (quality of service), and hence cannot provide optimized network performance in practice. In this paper, we propose four crosstalk-aware WA algorithms as variations of the well-known first-fit, random-pick, most-used and least-used schemes, with the crosstalk factor taken into consideration. Simulation results show that, independent of the network amplifier placement and the traffic, our WA algorithms can successfully suppress the in-band crosstalk throughout the network, and significantly decrease BER blocking rate.

Advances in optical WDM technology have paved the way for high-capacity wavelength channels capable of carrying information at Gb/s rates. However, with current traffic streams requiring only a fraction of a wavelength's bandwidth, it becomes necessary to groom these independent low rate traffic streams on to higher capacity wavelength channels. An all-optical approach to grooming is to allow many connections to time-share a wavelength. Accordingly, in a TDM wavelength routing network, the establishment of a connection requires the assignment of time slots in addition to routing and wavelength assignment. One of the primary challenges in such networks is the need for quick reconfiguration at the routing nodes. In this paper, we investigate the effects of switch reconfigurability, wavelength conversion and time slot interchangers (TSIs) on the blocking performance of connections with multiple rates. Heuristics for time slot assignment that consider constraints imposed by six different node architectures are proposed, and the blocking performance of the TDM wavelength routing network is evaluated through simulations. Results indicate that limited reconfigurability at the nodes is sufficient to attain the performance obtained with full reconfigurability, especially when connections occupy only a small fraction of the wavelength capacity. Furthermore, the blocking performance is not seen to benefit significantly with the introduction of wavelength converters and TSIs, thus signifying that the improvement in blocking is largely dependent on the switch reconfigurability at the nodes.

The MAC layer performance analysis of multi-hop ad hoc networks, is a challenging exercise due to the fact that the behavior of a node is dependent not only on its neighbors' behavior, but also on the behavior of other unseen nodes. In this paper, an approach to model the throughput performance of such networks that provides an accurate approximation is presented. The accuracy of the model is shown by applying it to 802.11-based ad hoc networks and comparing with simulation results.

In transparent WDM (wavelength division multiplexing) optical networks, some component and network vulnerabilities can be used by malicious attackers to implement QoS (quality of service) degrading/disruptive attacks. Among the attacks, some are easy to detect, locate and isolate, while some are not. For those covert attacks, which impair the BER of the legitimate channels without triggering alarms at the immediate attacking spot, the service impact may be much more severe than the overt attacks, since the NMS (network management system) might mis-diagnose the failure location. This paper studies one specific form of the covert attacks, which we name the low-power QoS attack. The attacking method, QoS impact, as well as the propagating nature of this attack are investigated. Our study shows that the low-power attack is easy to implement, yet its damage to the BER performance can be significant. Multiple colluding low-power attacks are capable of greater QoS impact using less attacking intensity. Measures to protect the network from low-power attack are also proposed.

Wavelength-routed WDM networks are expected to offer on-demand lightpath services in the near future. Lightpath bit error rate (BER) is one of the most important quality of service (QoS) performance measures. In a transparent WDM network, an admitted lightpath can potentially impact the BER performance of many other lightpaths that are already established, and hence admission control becomes necessary. In order to reduce the number of BER-related call blockings, a dynamic power management algorithm in contrast to the traditional fixed source power assignment is proposed in this paper. Our study shows that having adjustable source powers can achieve the same BER performance as in the fixed-power case while using much lower average source powers. As a result, detrimental fiber nonlinearities are significantly reduced, thus leading to much improved network BER performance.

We consider an IP/MPLS over WDM network, in which label switched routers (LSRs) in the IP/MPLS layer are interconnected through optical cross-connects (OXCs) in the optical core network (WDM layer) providing an end-to-end wavelength routing capability. In this paper, we study a dynamic label switched path (LSP) provisioning problem for the three different network models of the IP/MPLS over WDM network, namely, overlay, augmented, and peer models. For the overlay model, we propose an algorithm, called MLH/spl I.bar/OVLY, in which a network finds a path with the minimum number of logical hops for an LSP request. We also propose, for the augmented model, a simple and efficient dynamic LSP provisioning algorithm, called CAPA/spl I.bar/AUG, utilizing summarized capacity information from the WDM layer. We compare the proposed algorithms with the existing algorithms available for the overlay and peer models, and also provide an in-depth analysis of the algorithms. The algorithms are compared and evaluated using two key performance measures: LSP blocking probability and network (lightpath) utilization. Simulation results show that at low loads with a limited number of ports available in the network, CAPA/spl I.bar/AUG achieves an order of magnitude better blocking performance and outperforms all other algorithm. Considering the small amount of information that is exchanged between the layers in the augmented model, these results suggest that the augmented model can be a practically good compromise between the overlay and peer models.

This paper considers the problem of providing quality-of-protection (QoP) classes and improving the failure-recovery time performance in dense wavelength-division-multiplexing (DWDM) networks that use mesh-based path restoration schemes with capacity sharing. We focus on the signaling process required for reconfiguring the nodes along the preplanned restoration paths, and propose a novel approach for reducing the restoration time and meeting the QoP requirements by coordinating the setup procedures for the backup paths through scheduling. We present priority-based online scheduling algorithms that are amenable to distributed implementation for the problems of: 1) minimizing the worst case restoration time and 2) maximizing the number of connections that meet their QoP-class-specific restoration time deadlines. We also present mixed-integer-linear-program (MILP) formulations for both problems for comparison purposes. The online scheduling methods that we propose use simple connection and/or class-specific information and can be easily implemented with minor modifications to the currently proposed signaling protocols. We apply these methods to signaling protocols that require cross-connect configurations at different nodes to be done in sequence as in the current generalized multiprotocol label switching specification, as well as signaling protocols that allow cross-connect configurations to be done in parallel. It is shown that in both cases, significant performance improvements are achievable through scheduling in terms of both the QoP grades that can be supported and the restoration times, with both the MILP solutions and the heuristics. The improvement in restoration time and restorability through our heuristics can be quite high (e.g., increase from a network restorability performance of 40% to a network restorability of 88%, and a 12% reduction in worst case restoration time).

Network survivability is a crucial requirement in high-speed optical networks. Typical approaches of providing survivability have considered the failure of a single component such as a link or a node. We motivate the need for considering double-link failures and present three loopback methods for handling such failures. In the first two methods, two edge-disjoint backup paths are computed for each link for rerouting traffic when a pair of links fails. These methods require the identification of the failed links before recovery can be completed. The third method requires the precomputation of a single backup path and does not require link identification before recovery. An algorithm that precomputes backup paths for links in order to tolerate double-link failures is then presented. Numerical results comparing the performance of our algorithm with other approaches suggest that it is possible to achieve almost 100% recovery from double-link failures with a moderate increase in backup capacity. A remarkable feature of our approach is that it is possible to trade off capacity for restorability by choosing a subset of double-link failures and designing backup paths using our algorithm for only those failure scenarios.

We consider an IP/MPLS-over-WDM overlay network, which an IP/MPLS service layer exists independently over a WDM transport network. In the overlay model, two adjacent layers communicate with each other only through a pre-defined user-network interface (UNI) and the WDM topology information is not available to the service layer. In this paper, we concentrate on network design with a fault tolerance mechanism provided only in the IP/MPLS service layer in which the label-switched path (LSP) protection takes into account both node failures in the IP/MPLS service layer and failures in the WDM transport layer. In this design problem, the service layer must decide how lightpaths or logical links should be provisioned and how primary and backup LSPs should be routed in a cost-effective manner, such that failures do not affect both LSPs simultaneously. We formulate this survivable IP/MPLS overlay network design problem as a mixed integer linear program (MILP). Since the MILP is computationally intractable for large networks, we solve it only for a small network and present the optimal solution found by CPLEX 7.5, a commercial MILP solver. In order to find efficient solutions in larger networks we propose a heuristic algorithm that performs near-optimally in the cases studied. we then compare total costs for LSP protection from WDM link failures and LSR failures may not be much more expensive than protection from LSR failures only.

In this paper, we propose a new IEEE 802.11-based MAC protocol for ad hoc networks. The protocol incorporates multiple transmit power options in the MAC layer by letting stations adaptively select the transmit power, and thus vary the coverage area. It is shown that the protocol is capable of improving network latency, which is a fundamental QoS measure. The latency reduction factor is more than two under low and moderate load conditions, and close to one at high offered loads, when compared to the standard IEEE 802.11 DCF MAC protocol. This reduction in latency is achieved with almost no decrease in network throughput.

Multicasting is becoming increasingly important in today's networks. In optical networks, optical splitters facilitate the multicasting of optical signals. By eliminating the transmission of redundant traffic over certain links, multicasting can improve network performance. However, in a wavelength-division multiplexed (WDM) optical network, the lack of wavelength conversion necessitates the establishment of a single multicast circuit (light-tree) on a single wavelength. On the other hand, establishing several unicast connections (lightpaths) to satisfy a multicast request, while requiring more capacity, is less constrained in terms of wavelength assignment. The objective of the paper is to evaluate the tradeoff between capacity and wavelength continuity in the context of optical multicasting. To this end, we develop accurate analytical models with moderate complexity for computing the blocking probability of multicast requests realized using light-trees, lightpaths, and combinations of light-trees and lightpaths. Numerical results indicate that a suitable combination of light-trees and lightpaths performs best when no wavelength conversion is present.

Network survivability is a crucial requirement in high-speed optical networks. Typical approaches of providing survivability have considered the failure of a single component such as a link or a node. We consider a failure model in which any two links in the network may fail in an arbitrary order. Three loopback methods of recovering from double-link failures are presented. The first two methods require the identification of the failed links, while the third one does not. However, precomputing the backup paths for the third method is more difficult than for the first two. A heuristic algorithm that pre-computes backup paths for links is presented. Numerical results comparing the performance of our algorithm with other approaches suggests that it is possible to achieve 100% recovery from double-link failures with a modest increase in backup capacity.

The design of logical topologies in wavelength-routing multihop optical networks is a well-studied problem. We consider logical topology (LT) design over the popular ring and linear topologies. Our objective is the minimization of the electronic processing delay for the worst case traffic flow. For uniform traffic between nodes, this delay minimization corresponds to minimizing the number of hops on a shortest path between the farthest two nodes in the logical topology (the diameter of the logical topology). The simple structure of the physical topologies enables us to present a rigorous analysis of the problem. We present lower bounds for the achievable diameter wherever possible and propose practical logical topology design algorithms and corresponding upper bounds. We also present an application of the LT designs in the linear topology to the survivability of ring networks.

We consider a multichannel optical network that provides circuit-switched service. Different rates may be provided to sessions by allocating a variable number of channels to the sessions. In this paper, a new analytical model to compute the blocking probabilities of dual-rate sessions (sessions having one of two rates) is presented. The model considers channel-switching constraints along a session's path but ignores the correlation between link channel occupancies. The model is applicable to a variety of scenarios such as wavelength-grooming networks and all-optical TDM wavelength-routing networks. Simulations to show the usefulness of the model are also presented.

The efficient management of location information is an important issue in the design of future personal communication services (PCS) networks. Traditional architectures for location information databases have used a hierarchical tree structure with a single point of vulnerability, the root. In this paper, we consider a distributed database architecture for location management, and address the problem of load balance among the databases. We take into account the possibility that the databases may not have identical capabilities. Some conditions necessary to obtain load balance are derived and a location update algorithm that achieves load balance when the conditions are met is presented.

The blocking performance of wavelength routing WDM optical networks can be improved by employing wavelength conversion. In this paper, we address the problem of optimally placing a limited number of wavelength converters in mesh topologies. Two objective functions, namely, minimizing the average blocking probability and minimizing the maximum blocking probability over all routes, are considered. We first extend an earlier analytical model to compute the blocking probability on an arbitrary route in a mesh topology, given the traffic and locations of converters. We then propose heuristic algorithms to place wavelength converters, and evaluate the performance of the proposed heuristics using the analytical model.

We consider multiwavelength wavelength-routing networks operating in circuit-switched mode. Wavelength utilization is poor in such networks if sessions require only a fraction of a wavelength's capacity. An all-optical approach to improve wavelength utilization is to use time division multiplexing (TDM) on each wavelength, and switch time slots and wavelengths. In this paper, we address the off-line multirate session scheduling problem, i.e., the problem of assigning time slots and wavelengths to a given static set of multirate sessions, in ring topologies. Given a set of sessions and their relative rates, our objective is to maximize network throughput. This objective translates to the problem of minimizing the maximum length of a TDM frame over all wavelengths. We first show that the off-line single-rate session scheduling problem is equivalent to the off-line wavelength assignment problem, and hence obtain bounds on frame length. We then present scheduling algorithms with provable worst-case bounds on frame length for multirate session scheduling.

Wavelength converters increase the traffic-carrying capacity of circuit-switched optical networks by relaxing the wavelength continuity constraints. We consider the problem of optimally placing a given number of wavelength converters on a path to minimize the call-blocking probability. Using a simple performance model, we first prove that uniform spacing of converters is optimal for the end-to-end performance when link loads are uniform and independent. We then show that significant gains are achievable with optimal placement compared to random placement. For nonuniform link loads, we provide a dynamic programming algorithm for the optimal placement and compare the performance with random and uniform placement. Optimal solutions for bus and ring topologies are also presented. Finally, we discuss the effect of the traffic model on the placement decision.

We consider the problem of optimally placing a given number of wavelength converters on a path to minimize the call blocking probability. Using a simple performance model, we first prove that uniform spacing of converters is optimal for the end-to-end performance when the link loads are uniform and statistically independent. We then show that significant gains are achievable with optimal placement compared to random placement. For non-uniform link loads, we provide a dynamic programming algorithm for the optimal placement and compare the performance with random and uniform placement. Optimal solutions for bus and ring topologies are also presented.

We propose a new algorithm for the dynamic centralized wavelength assignment problem in fixed-routing WDM networks without wavelength conversion. The blocking performance of our algorithm is better in many cases (and no worse in tire cases we studied) than other previously proposed algorithms. The performance improvement of our algorithm over other algorithms is high for multi-fiber ring networks with a moderate number of fibers per link. In a multi-fiber mesh-torus network, the difference in performance is not as significant, but the blocking probabilities for all algorithms approach those achievable by wavelength conversion as the number of fibers per link increases. We also extend an earlier analytical model for predicting the blocking probability with and without wavelength conversion to dense multi-fiber networks. Finally, our simulation results on multi-fiber rings and mesh-tori reveal surprising results about the benefits of wavelength conversion as the number of fibers per link increases.

This paper makes the first known attempt to study wavelength-routing networks and the effects of wavelength conversion under dynamic non-Poisson traffic. An approximation that characterizes any non-Poisson traffic by its first two moments is utilized. The arrival occupancy distribution of busy wavelengths for this approximate process is derived and is used to analyze the effects of wavelength conversion. The model predicts that traffic peakedness plays an important role in determining the blocking performance, and also that wavelength conversion gain is insensitive to traffic peakedness over a large range.

Wavelength-routing networks offer the advantages of wavelength re-use and scalability over broadcast-and-select networks and are therefore suitable for wide area networks (WANs). We study the effects of topological connectivity and wavelength conversion in circuit-switched all-optical wavelength-routing networks. An approximate blocking analysis of such network is performed. We first propose an improved framework for the analysis of networks with arbitrary topology. We introduce a simple model for networks with a variable number of converters and analyze the effect of wavelength converter density on blocking probability. We then apply this framework to two sparse network topologies, the ring and the mesh-torus, and obtain the blocking performance. The results show that, in most cases, only a fraction of the network nodes need to be equipped with wavelength conversion capability for good performance. Finally, the tradeoff between physical connectivity, wavelength conversion, and the number of available wavelengths is studied through networks with random topologies.

Unlike broadcast-and-select networks, wavelength-routing networks offer the advantages of wavelength reuse and scalability and are thus suitable for wide-area networks (WANs) We study the effects of topological connectivity and wavelength conversion in circuit-switched all-optical wavelength-routing networks. A blocking analysis of such networks is given. We first propose an analytical framework for accurate analysis of networks with arbitrary topology. We then introduce a model for networks with a variable number of converters and analyze the effect of wavelength converter density on the blocking probability. This framework is applied to three regular network topologies that have varying levels of connectivity: the ring, the mesh-torus, and the hypercube. The results show that either a relatively small number of converters is sufficient for a certain level of performance or that conversion does not offer a significant advantage. The benefits of conversion are largely dependent on the network load, the number of available wavelengths, and the connectivity of the network. Finally, the tradeoff between physical connectivity, wavelength conversion, and the number of available wavelengths is studied through networks with random topologies.

In this paper, we discuss the problem of establishing multicast connections using self-routing multistage interconnection networks (MINs). We examine the issue of multicasting and propose various schemes to accomplish it. Multicasting using recycling of cells has been suggested in the literature recently. We propose a multicast switch architecture that requires a few simple enhancements over a MIN and describe the process of setting up a multicast connection. The size of the routing tables limit the number of copies per cell that can be generated per pass through the network. We assume that no more than two copies per cell can be generated per pass through the routing network and provide detailed cell-level simulation results for the average delay of a cell. The results show that recycling is a practicable option for multicasting and that the delay does not increase drastically when cells are recycled.