Radio Link Failure (RLF) detection is essential for maintaining reliable connectivity in 5G networks. However, traditional centralized detection mechanisms often encounter scalability and latency constraints when managing large-scale, geographically distributed infrastructures. To address this challenge, we introduce a Lyapunov-driven federated learning framework that adaptively selects gNodeBs based on both data utility and historical participation. This approach leverages an LSTM-based local model to capture temporal patterns in link performance, thereby enhancing RLF detection. Extensive evaluations on a real-world 5G dataset demonstrate that the proposed method achieves superior performance compared to baseline approaches when detecting rare failure events. By simultaneously prioritizing performance and fairness, this framework offers a scalable solution suited to diverse and dynamic 5G environments.

Robust and efficient design of uniform linear arrays (ULA) with non-uniform weights for multi-user massive MIMO systems is formulated as a constrained min-max optimization problem to minimize the number of antennas while meeting performance target. Chebyshev beamforming weights are shown to be optimal for this problem in line-of-sight (LoS) channels when precise channel state information (CSI) of interfering users is not available. Based on this, explicit closed-form solutions are obtained for sum as well as per-user interference constraints, which reveal the scaling of the minimal number of antennas with target signal-to-interference plus noise ratio and the number of interfering users. Comparison with widely-used uniformweighting arrays highlights the superior scalability and efficiency of the proposed Chebyshev-based design, particularly in complex interference environments.

Integrated Terrestrial-Satellite Backhaul Networks (ITSBN) offer wide coverage and cost effectiveness, especially in remote or underserved areas, where backhaul links are facilitated through a multi-beam mechanism. In this context, a Beam Hopping Plan (BHP) can enhance resource efficiency by dynamically activating beams based on traffic demand, improving coverage flexibility. However, despite its several advantages, ITSBN may encounter congestion challenges during periods of high traffic demand. Content caching may be a solution to alleviate the network load, enabling content delivery at the network edge and reducing backhaul traffic. In this work, we consider ITSBN with caching at local base stations (BSs) and aim to maximize the system throughput. To achieve this, we propose a double-layer iterative algorithm. The outer layer employs our proposed$\varepsilon$-Beam Hopping Plan and Content Selection algorithm ($\varepsilon$-BHPCS), leveraging the matching game approach. This method jointly designs the BHP and selects content carried by each beam, balancing accuracy and execution time through the parameter$\varepsilon \in[0, 1]$. For inner layer, we propose the Alternating Direction Method of Multipliers Based Power Allocation and Caching (ADMM-PAC) scheme, optimizing power across beams and managing caching at BSs in a distributed fashion, using BSs' resources and local information for parallel optimization processes without requiring global information as in a centralized method. Simulations show that$\varepsilon$-BHPCS achieves performance similar to the optimal exhaustive search, while the distributed ADMM-PAC aligns closely with standard centralized interior-point optimization algorithm.

Radio Link Failure (RLF) detection in Radio Access Networks (RANs) is crucial for ensuring seamless communication in 5G networks. Nonetheless, current approaches based on traditional Machine Learning (ML) algorithms fail to find a trade-off between accuracy and computational complexity. Thus, in this paper, we explore advanced and computationally efficient types of transformers, such as Linformer and Performer. These models significantly reduce computational complexity while maintaining strong performance by leveraging different attention mechanisms. Extensive evaluations using a realistic dataset under various percentages of link failures show that Linformer provides a favourable balance between accuracy and training time.

This paper presents a fuzzy set-based methodology to achieve simultaneous path adherence and local collision avoidance in autonomous navigation. It targets scenarios where a global path to the vehicle's destination is already planned but with imperfect knowledge of the obstacles in the environment and their dynamics. Assuming that the vehicle can localize itself and the obstacles around it, the proposed methodology incorporates the global path and the acquired real-time knowledge of the local obstacles by the vehicle to create a comprehensive fuzzy representation of the environment. This fuzzy representation is then utilized to assess the desirability of the vehicle states within an optimization framework that balances global path adherence and obstacle avoidance objectives. The proposed gradient-based solution to this optimization problem navigates the vehicle such that it maintains its global course toward the designated destination while avoiding collision with local obstacles. Extensive simulations on a mobile robot validate the method's efficacy in guiding vehicles along desired paths, maneuvering around obstacles with minimal deviation from the route, and negotiating local minima without oscillations. Additionally, simulation results highlight the proposed fuzzy set-based methodology's low computational complexity, real-time operation capability, and adaptability in handling complex geometries of paths and obstacles.

This paper introduces a novel approach to enhance the stability and efficiency of R-STDP in the context of federated learning. The primary objective is to stabilize the unbounded growth of R-STDP and make it more responsive to real-time changes. The methodology involves integrating R-STDP with Spiking Neural Networks and employing the norm of the neural network model for adjusting weighted aggregation in federated learning systems. The proposed method incorporates a mechanism where weights decay over time, depending on the duration since the agent last published its model. Additionally, the sampling time is dynamically adjusted based on the Euclidean norm, which measures the distance between the weight matrices of the agents and the server. The results demonstrate that the proposed event-triggered federated learning method significantly enhances learning speed and performance. At the same time, the dynamic aggregation interval efficiently reduces communication between the agents and the central server, especially after model convergence. This research presents a significant advancement in federated learning and offers a more stable, responsive, and efficient learning process.

Truck platooning is an innovative logistics approach to lower operational costs, particularly fuel consumption, while addressing contemporary transportation challenges. While recent studies on truck platooning have emphasized platoons’ energy savings, stability, and safety, there has been limited exploration of platoon formation and control. This paper uses optimal control theory to address the dispatching control of trucks with arriving platoons. In particular, trucks arrive at a highway station while platoons arrive alongside it. The station controls the truck holding and dispatching, where trucks are sent out with or without a platoon. Dispatching trucks with an arriving platoon reduces fuel consumption while waiting for a platoon to arrive increases the dwell time (i.e., transportation delay). We assume that an arriving platoon determines the number of trucks (i.e., the batch size) it can accept. Only a single truck can be dispatched if a platoon is absent. Hence, we formulate the dispatching control problem and derive the optimal policy for the discounted costs and the average cost governing the dispatch of trucks alongside platoons. We proved the optimality of threshold policies. Numerical results for the average cost case are presented. They are consistent with the optimal ones.

This paper presents an advanced implementation of the single-snapshot 2D-MUSIC algorithm, enhanced by cross-linear antenna arrays, to improve uplink indoor positioning accuracy using OFDM-based 5G networks. The proposed method adeptly handles multipath interference and non-line-of-sight (NLoS) conditions, demonstrating significant advancements over traditional techniques. Our method not only accurately determines the position of user equipment in two dimensions but also extends to 3D positioning. A novel aspect of our research includes addressing the challenge of unknown time of departure, which often complicates the time of flight calculations necessary for precise localization. By using two strategically placed cross-linear antenna arrays, our system compensates for this uncertainty, providing reliable and precise location estimates even without perfect transmitter-receiver synchronization. Simulation results validate the robustness of our approach, showing exceptional localization accuracy and promising potential for complex IoT applications within industrial settings. Provided the separation between multipath components in either range or angle is sufficiently large, the algorithm is capable of detecting the transmitter with sub-centimeter accuracy.

In this paper we address the challenge of efficiently deploying Virtual Network Functions (VNFs) in network infrastructures. This is particularly crucial when facing resource fragmentation, where available resources are not fully utilized due to the fluctuating allocation and deallocation of virtual network requests. Traditional optimization techniques often fall short in managing the dynamic complexities of VNF placement. To overcome this, we introduce a novel online VNF placement strategy using Deep Reinforcement Learning (DRL) combined with a Reward Constrained Policy Optimization (RCPO). This method leverages the flexibility of DRL and the constraint integration capacity of RCPO, ensuring compliance with performance and resource limitations while minimizing resource fragmentation. The results demonstrate that our DRL-based method surpasses existing methods, resulting in more effective resource management and less resource fragmentation.

This paper tackles an online resource allocation problem with job transfers in communication networks. The system operates in discrete time, where at each time slot, the administrator reserves resources at servers for future job requests at a cost. The specificity here is that the jobs may be transferred between the servers to accommodate the demands best at an additional cost. Moreover, a violation cost is associated with each blocked job request. The goal is then to build an online policy that minimizes the overall cost. We propose a randomized online algorithm based on the exponentially weighted method that learns from the previous job request sequence. We prove that our algorithm enjoys a sub-linear in time regret, which indicates that the algorithm is adapting and learning from its experiences and is becoming more efficient in its decision-making as it accumulates more data. In addition, we test the performance of our algorithm on artificial data and compare it against a reinforcement learning method where we show that our proposed method outperforms the latter.

The recent networking trends driven primarily by the different virtualization technologies, such as Network Function Virtualization (NFV) and Service Function Chaining (SFC) pave the way for next-generation network services. In the 5G and beyond era, such services usually have strict delay requirements and the wider adoption of the distribution of their computational needs across the Edge-to-Cloud continuum is certainly a step in the right direction. However, the majority of the optimization solutions for placing the virtualized services so far focus on server selection, leaving other areas such as the impact of Non-Uniform Memory Access (NUMA) and CPU core selection underexplored. In this work, we herein formulate the problem of placing services as SFCs on an Edge/Cloud infrastructure, as a Mixed Integer Programming (MIP) problem. Then, we propose a heuristic algorithm called “Dynamic numa node Selection through Cores consolidation – DySCo" to solve it, which optimizes the placement in terms of server, NUMA and core selection. To the best of our knowledge, this is the first attempt to optimize network service placement in an Edge-Cloud interplay. Extensive simulation evaluation shows that DySCo is able to perform close to optimal while finding a solution in a real time fashion. Compared to a mix of baselines and modified solutions from the literature to treat this new problem, DySCo reduces on average the deployment cost by 17.53% and the delay by 28.88% for a given SFC.

Task migration to remote servers offers a promising solution to the congestion issue in mobile edge computing systems. Our optimal task offloading design minimizes a system cost function, encompassing offloading and penalty costs. The offloading cost reflects external server resource usage, while the penalty cost accounts for task expiration risk. To optimize the expected cost over a time horizon, we employ Dynamic Programming (DP) and analyze its properties for an optimal offloading policy. “Curse of Dimensionality” of the DP equation poses computational challenges, especially with infinite state space. To mitigate this, we identify crucial policy properties, enabling DP evaluation on a finite state subset. Moreover, we show that the computation of the optimal task offloading decision at a given state can be deduced by leveraging the optimal decision taken at its “adjacent” states. We then provide numerical results to demonstrate parameter impact and validate theoretical findings.

Task offloading in mobile edge computing systems is subject to various random factors including the connection to external servers, new task requests from users, and the availability of local processing services. However, statistical information is often not available in practical scenarios. To tackle the issue, we adopt a Q-learning-based approach that learns the optimal task offloading policy through observations of random events. Traditional Q-learning methods may face challenges such as long training times and high memory usage due to the large state and action space. To overcome this problem, we propose a novel method that leverages the concept of adjacent state sequence. In this type of sequence, we can infer the optimal offloading decision of a system state from other states. This method aims to improve the convergence speed and memory efficiency of the learning model by reducing the number of parameters that need to be learned and stored. Those eliminated parameters instead can be computed via a derived linear expression. We conduct experiments to demonstrate the enhancement of our proposed method compared to the traditional$\mathbf{Q}-$learning in the studied problem.

Low-earth-orbit satellite networks (LSNs) offer an enhanced global connectivity and a wide range of applications such as disaster response and military operations, among others. Each specific application can be represented by a service function chain (SFC) in which each function is considered as a task in the application. Our objective is to optimize the long-term system performance by minimizing the average end-to-end delay of SFC deployments in LSNs. To achieve this, we formulate a dynamic programming (DP) problem to derive an optimal placement policy. To overcome the computational intractability, the need for statistical knowledge of SFC requests, and centralized decision-making challenges, we present a multi-agent Q-learning approach where satellites act as independent agents. To facilitate performance convergence in non-stationary agents' environments, we let agents to collaborate by sharing designated learning parameters. In addition, agents update their Q-tables via two distinct rules depending on selected actions. Extensive experimentation shows that our approach achieves convergence and performance relatively close to the optimum obtained by solving the formulated DP equation.

Truck platooning is a promising technology that can reduce costs (fuel consumption) and enhance the overall transportation productivity. While recent research has focused on platoons' network and stability, few studies have tackled platooning formation and control. This paper uses Reinforcement Learning (RL) to study the dispatching control of trucks with arriving platoons, a problem first proposed in [1]. This work builds on [1] by considering the lack of the cost function and statistical knowledge. In particular, we employ Q-learning to compute the optimal dispatch control policy at a highway hub. Given the unbounded state space of the model, traditional Q-learning may converge slowly or even get stuck in sub-optimal policies. We improve Q-learning by confining the agent to transition in a finite subset of the state space. For this purpose, we use the switching condition property of the optimal policy (derived in [1]), the underlying random walk model, and a sensitivity analysis of the cost function. Our numerical results demonstrate that our Enhanced Q-learning converges significantly faster (up to 97%) in terms of CPU time and number of interactions.

In the 5G/6G era of networking, computational offloading, i.e., the act of transferring resource-intensive computational tasks to separate external devices in the network proximity, constitutes a paradigm shift for mobile task execution on Edge Computing infrastructures. However, in order to provide firm Quality of Service (QoS) assurances for all the involved users, meticulous planning of the offloading decisions should be made, which potentially involves inter-site task transferring. In this paper, we consider a multi-user, multi-site Multi-Access Edge Computing (MEC) infrastructure, where mobile devices (MDs) can offload their tasks to the available edge sites (ESs). Our goal is to minimize end-to-end delay and energy consumption, which constitute the sum cost of the considered system, and comply with the MDs’ application requirements. To this end, we introduce a two-stage Reinforcement Learning (RL)-based mechanism, where the MDs-to-ES task offloading and the ES-to-ES task transferring decisions are iteratively optimized. The proper operation, effectiveness and efficiency of our proposed offloading mechanism is assessed under various evaluation scenarios.

As the automotive industry is developing autonomous driving systems and vehicular networks, attention to truck platooning has increased as a way to reduce costs (fuel consumption) and improve efficiency in the highway. Recent research in this area has focused mainly on the aerodynamics, network stability, and longitudinal control of platoons. However, the system aspects (e.g., platoon coordination) are still not well explored. In this paper, we formulate a platooning coordination problem and study whether trucks waiting at an initial location (station) should wait for a platoon to arrive in order to leave. Arrivals of trucks at the station and platoons by the station are modelled by independent Bernoulli distributions. Next we use the theory of Markov Decision Processes to formulate the dispatching control problem and derive the optimal policy governing the dispatching of trucks with platoons. We show that the policy that minimizes an average cost function at the station is of threshold type. Numerical results for the average cost case are presented. They are consistent with the optimal ones.

Cloud gaming offers excellent potential; however, it presents significant challenges for 5G networks because of its strict requirements for high reliability and low latency. Cloud gaming service deployment can be modeled as a Virtual Network Functions Chain Placement Problem (VNF-CPP), where a service instance is represented by a chain of interconnected Virtual Network Functions (VNFs). Solving the VNF-CPP may result in infeasible solutions when the underlying network infrastructure can not meet the service requirements. In this paper, we propose an automatic feasibility restoration technique and explain how network operators can use it to meet cloud gaming's demands. Our proposed algorithms reveal the origins of the infeasibilities so that network operators can understand the reasons behind them. Moreover, the algorithms suggest the best way to alter the network to regain feasibility by providing realtime elastic resource management. Specifically, two approaches are proposed and evaluated. The first approach is the Irreducible Infeasible Set (IIS) Repair, and the second is the Minimum Cost Redesign. We evaluate the proposed algorithms using two practical use cases: an offline use case, where we need to fix the infeasibility for a bulk of service instances, and an online use case, where we need to fix the infeasibility for a single service instance in realtime. Furthermore, theoretical analysis and results show that the Minimum Cost Redesign method outperforms the IIS Repair method. Results also verify that our algorithms can provide practical solutions for both use cases in realtime.

Recent trends in Network Function Virtualization (NFV) combined with Internet of Things (IoT) and 5G applications have reshaped the network service offering. In particular, Service Function Chains (SFCs) can associate network functions with physical and virtual resources towards providing a complete network service. Concurrently, the management of a continuously expanding network and the fulfillment of the applications’ requirements pave the way for autonomic network solutions. Intent Based Networking (IBN) is a novel paradigm that aims to achieve the automatic orchestration of network services and the assurance of their performance. Accordingly, in this paper, we propose a novel automated network assurance model, based on Model Predictive Control, to guarantee the Quality of Service (QoS) and security requirements of multi-tenant and IBN-enabled SFCs. In this context, corrective decisions are proactively taken, in the form of incoming intent relocations among the SFCs. The results reveal that our model can assure with high probability the application requirements and minimize QoS violations.

Modern networking paradigms like Service Function Chaining (SFC) allow for services to be broken down to a series of ordered and interconnected Virtualized Network Functions (VNFs) that can be hosted in generic servers in EdgeCloud datacenters. Nonetheless, a critical issue arises, when a hardware or software failure occurs and the VNFs of an SFC need to be repositioned, allowing to autonomously bring the system back to its normal operation, a process called self-healing. In this paper, a distributed methodology is proposed that aims to address this challenge, considering the requirements of all involved actors. Specifically, a Reinforcement Learning (RL) based algorithm is proposed that allows to iteratively optimize and determine an SFC healing solution upon a datacenter failure. As a second stage, a revenue-driven resource allocation mechanism is integrated, to resolve the contention for resources in an already functional datacenter that potentially occurs due to the repositioning. Various simulation scenarios prove the efficiency of our proposed resilient healing mechanism.

Pursuit-Evasion (PE) games are regarded as a major platform for game theory. In this kind of game, an agent called an evader tries to escape from another agent called a pursuer. Active Target Defense (ATD) is a derivative of PE games, attracting attention recently. In an ATD game, the evader, often called an invader, strives to capture a moving target. The pursuer, called a defender, tries to intercept the invader. This paper implements the Spike-Timing-Dependent Plasticity (STDP) algorithm to train two Spiking Neural Networks (SNNs) to find a suitable solution for the ATD problem in decentralized situations. One of the SNNs is used to control the invader, while the other controls the defender. The performance is compared with the analytical solution for the pedestrian model. The results showed that an SNN can learn the optimal capture point only using relative velocities and line of sight.

We propose a new approach for proportionally fair sharing of radio access resources (i.e., frequency carriers) in a 5G mobile environment. In particular, we formulate the multi-carrier scheduling problem as a game (played by different carriers) which results in proportionally fair user service rates at its equilibrium. We subsequently present a globally convergent algorithm to find the equilibrium point of the devised game, based on which we develop a frame-by-frame scheduling mechanism. We further discuss how to extend the proposed scheduling scheme to a multi-tenant setting in order to efficiently implement radio access network slicing in a 5G mobile system.

This paper presents a machine-learning algorithm applied to a quadcopter application. We are proposing a fuzzy actor-critic learning (FACL) algorithm. This method enables a pursuer quadcopter to capture an evader quadcopter in the pursuit-evasion (PE) differential game. In this application, the pursuer learns its control strategies by interacting with evader and learning from past experiences. Both the critique and the actor are fuzzy inference systems (FIS). It is assumed that the pursuer knows only the instantaneous position and speed of the evader and vice versa. The FACL will generate the desired trajectory as the input for low-level controllers. Simulation results are presented for the PE differential game to demonstrate the practicality of our machine-learning algorithm.

The considerable increase of multimedia services, such as video-on-demand (VoD) services, is a significant contributor to the total Internet traffic. Software-defined networking (SDN) and information-centric networking (ICN) are two promising technologies that can be combined to facilitate video delivery and to reduce network delays. In this paper, we first formulate the caching decision problem as a 0-1 integer linear programming (ILP) problem. Second, in contrast to existing approaches that solve the formulated ILP problem by assuming all future video requests are known, we consider the impact of the time scale, which transforms the static 0-1 ILP problem into a dynamic problem. By solving the dynamic 0-1 ILP problem, we find more accurate optimal solutions compared to existing approaches. Third, since the formulated 0-1 dynamic ILP problem is NP-hard, we leverage the in-network caching of ICN and the global view of the SDN controller to propose a novel SDN-based caching decision policy. Finally, extensive evaluations are performed, and the results demonstrate that the proposed SDN-based caching decision policy provides solutions that are close to the optimum in substantially less computation time. The SDN-based caching decision policy also outperforms existing practical ICN caching decision policies in terms of the cache hit ratio and the average number of hops, which are directly related to the video delivery latency. Moreover, the SDN-based caching decision policy can substantially reduce the number of generated and broadcasted interest packets, which is a shortcoming of the current ICN.

In this work we present a novel technique for multicast route encoding in SDN using bloom filters and a minimally extended version of OpenFlow. We demonstrate that this technique allows multiple trees to be installed into the network for each multicast group without any overhead in flow table size, thereby significantly improving the forwarding state scalability of multi-tree traffic engineering in SDN. We implement and evaluate this technique using both flow level network simulation and packet level network emulation with Mininet. We demonstrate that the application of our technique imposes only a modest increase in flow setup time, that is on the same order as typical IPTV channel zapping times, and we present several variations of bloom filter construction technique that allow this flow setup time disadvantage to be further mitigated.

Since multimedia services will become fundamental in the upcoming 5G networks, how to improve the user quality of experience (QoE) is becoming a major challenge. In this paper, we integrate the concept of Information-Centric Networking (ICN) to the infrastructure of 5G networks. Due to the in-network caching feature of ICN, proactive caching can be beneficial in 5G networks. More precisely, this paper introduces a novel proactive caching approach (called smart caching) which leverages the non-negative matrix factorization (NMF) technique to predict the future ratings of user preferences on all videos for 5G-ICN networks. To solve the shortcoming of the NMF technique that generates inaccurate predictions for high rated but unpopular videos, we also take video historical popularity into consideration. Thus, the user future demands can be predicted based on the user preferences (i.e. the predicted ratings) and the historical popularity of videos. Simulation results show that the proposed smart caching outperforms existing approaches in terms of hit ratio, average video retrieval delay, and user satisfaction.

In this paper, we study cost-efficient and fair allocation of multiple types of resources in an environment of heterogeneous and self-organizing servers. To address this problem, we formulate an optimization problem which aims at minimizing the operational costs for all servers, while providing fairness across different users. We propose a fully distributed implementation to solve this problem. The proposed mechanism is shown to achieve envy-freeness among different users. Furthermore, we show how it captures the trade-off between cost-efficiency and fairness. We employ numerical experiments to show the effectiveness of our proposed mechanism in reducing operational costs for a geo-distributed data-center.

As devices for the Internet of Things (IoT) are typically battery-powered, energy efficiency is a major challenge for IoT networks. In this paper, we leverage the in-network caching of Information-Centric Networking (ICN) to propose a novel cooperative caching scheme, based on the IoT data lifetime and user request rate, to improve the energy efficiency of IoT networks. By caching IoT data at different nodes (such as content routers, base stations, etc.), IoT devices can stay in sleep mode for a larger portion of time and therefore reduce the overall energy consumption. With the help of an auto- configuration mechanism, the proposed IoT data Lifetime-based Cooperative Caching (LCC) scheme can dynamically adapt to the change of request rate. Extensive evaluations were performed and the simulation results show that LCC outperforms existing schemes in terms of total energy consumption reduction (up to 40%) and the reduction in the average number of hops traversed along the path (up to 20%), which is also directly related to the response time. Keywords- Internet of Things (IoT), Cooperative Caching, Information-Centric Network (ICN).

We first consider the static problem of allocating resources to (i.e., scheduling) multiple distributed application frameworks, possibly with different priorities and server preferences, in a private cloud with heterogeneous servers. Several fair scheduling mechanisms have been proposed for this purpose. We extend prior results on max-min fair (MMF) and proportional fair (PF) scheduling to this constrained multiresource and multiserver case for generic fair scheduling criteria. The task efficiencies (a metric related to proportional fairness) of max-min fair allocations found by progressive filling are compared by illustrative examples. In the second part of this paper, we consider the online problem (with framework churn) by implementing variants of these schedulers in Apache Mesos using progressive filling to dynamically approximate max-min fair allocations. We evaluate the implemented schedulers in terms of overall execution time of realistic distributed Spark workloads. Our experiments show that resource efficiency is improved and execution times are reduced when the scheduler is "server specific" or when it leverages characterized required resources of the workloads (when known).

The virtualization of network functions is promising significant cost reductions for network operators. Running multiple network functions on a standard x86 server instead of dedicated appliances can increase the utilization of the underlying hardware,while reducing the maintenance and management costs of such functions. However, total cost of ownership calculations are typically a function of the attainable network throughput, which in a virtualized system is highly dependent on the overall system architecture - in particular the input/output (I/O) path. In this paper we investigate the attainable performance of an x86 host running multiple virtualized network functions (VNFs) under different I/O architectures: OVS-DPDK, SR-IOV, and FD.io VPP. Running multiple VNFs in parallel on a standard x86 host is a common use-case for cloud-based networking services. We show that the system throughput in a multi-VNF environment differs significantly from deployments where only a single VNF is running on a server.

Efficient and fair allocation of multiple types of resources is a crucial objective in a cloud/distributed computing cluster. Users may have diverse resource needs. Furthermore, diversity in server properties/capabilities may mean that only a subset of servers may be usable by a given user. In platforms with such heterogeneity, we identify important limitations in existing multi-resource fair allocation mechanisms, notably Dominant Resource Fairness and its follow-up work. To overcome such limitations, we propose a new server-based approach; each server allocates resources by maximizing a per-server utility function. We propose a specific class of utility functions which, when appropriately parameterized, adjusts the trade-off between efficiency and fairness, and captures a variety of fairness measures (such as our recently proposed Per-Server Dominant Share Fairness ). We establish conditions for the proposed mechanism to satisfy certain properties that are generally deemed desirable, e.g., envy-freeness, sharing incentive, bottleneck fairness, and Pareto optimality. To implement our resource allocation mechanism, we develop an iterative algorithm which is shown to be globally convergent. Subsequently, we show how the proposed mechanism could be implemented in a distributed fashion. Finally, we carry out extensive trace-driven simulations to show the enhanced performance of our proposed mechanism over the existing ones.

Information centric networking (ICN) is one of the emerging Internet paradigms proposed to overcome the shortcoming of the current host-centric Internet. With ubiquitous in-network caching, ICN can facilitate content delivery and reduce network delay. In this paper, we propose a novel collaborative caching scheme based on routers' position to cache popular videos on the edge routers which are closer to users. A priori knowledge of videos' popularity is not required as the proposed scheme adapts itself to the user requests. The benefits of our proposed approach are: light-weight, short content delivery time, and reduced network usages and publisher load. We use a simple topology to show how the proposed scheme works. Then, we use a realistic topology with real data traces to evaluate the performance of the proposed scheme. Simulation results show that our scheme outperforms existing schemes in terms of average number of hops and reduced publisher load ratio for both scenarios.

Users of cloud computing platforms pose different types of demands for multiple resources on servers (physical or virtual machines). Besides differences in their resource capacities, servers may be additionally heterogeneous in their ability to service users - certain users' tasks may only be serviced by a subset of the servers. We identify important shortcomings in existing multi-resource fair allocation mechanisms - Dominant Resource Fairness (DRF) and its follow up work - when used in such environments. We develop a new fair allocation mechanism called Per-Server Dominant-Share Fairness (PS-DSF) which we show offers all desirable sharing properties that DRF is able to offer in the case of a single “resource pool” (i.e., if the resources of all servers were pooled together into one hypothetical server). We evaluate the performance of PS-DSF through simulations. Our evaluation shows the enhanced efficiency of PS-DSF compared to the existing allocation mechanisms. We argue how our proposed allocation mechanism is applicable in cloud computing networks and especially large scale data-centers.

We study the broadcast transmission of a single file to an arbitrary number of receivers in a wireless one-hop setting, using random linear network coding (RLNC). In contrast to similar studies, we apply RLNC within segments of the file. In a previous study, we showed that this method can achieve near optimal file transfer completion time while tackling the main drawbacks of RLNC; increased decoding delay and increased storage and complexity requirements at the receivers. Towards that end we developed and evaluated a packet scheduling policy, namely the Least Received (LR) batches policy. In our previous work, we proved the optimality of the LR policy, with regards to the completion time, in systems with two receivers. In this work we will show that the LR policy is optimal, in the same sense, regardless of the number of the receivers.

Network Function Virtualization promises to reduce the overall operational and capital expenses experienced by the network operators. Running multiple network functions on top of a standard x86 server instead of dedicated appliances can increase the utilization of the underlying hardware and reduce the maintenance and management costs. However, total cost of ownership calculations are typically a function of the attainable network throughput, which in a virtualized system is highly dependent on the overall system architecture - in particular the input/ output (I/O) path. In this paper, we investigate the attainable performance of an x86 host running multiple Virtualized Network Functions (VNFs) under different I/O architectures: OVS, SRIOV and FD.io VPP. We show that the system throughput in a multi-VNF environment differs significantly from deployments where only a single VNF is running on a server, while different I/O architectures can achieve different levels of performance.

Network function virtualization along with network service chaining and forwarding graphs envision a reduction in the respective cost that end users, service providers, and network operators are experiencing, while providing complete and high quality services. The allocation of these service chains in a pool of available cloud or data center resources is a challenging problem that can affect the overall performance of the offered network services. Furthermore, a number of challenges associated with the hardware capabilities and the available resources of the cloud infrastructure, along with possible collocation constraints between the components of the service chain, can exponentially increase the complexity of resource allocation. This paper examines how to improve the overall allocation performance of deploying service chains in a cloud environment satisfying server affinity, collocation, and latency constraints. The proposed method is inspired by a partitioning game, where the various components of a service chain are split in a set of partitions executed as virtual machines/containers in appropriate servers. We mathematically prove that a Nash equilibrium exists for our partitioning game corresponding to an optimal solution. By implementing the partitioning game as an iterative refinement process, we also experimentally validate that the proposed algorithm converges to the optimal solution.

Lately, pervasive and ubiquitous computing services have been under focus of not only the research community, but developers as well. Different devices generate different types of data with different frequencies. Emergency, healthcare, and latency sensitive services require real-time responses. Also, it is necessary to decide what type of data has to be uploaded to the cloud, without burdening the core network and the cloud. For this purpose, the cloud on the edge of the network, known as Fog or Micro Datacenter (MDC), plays an important role. Fog resides between the underlying Internet of Things (IoTs) and the mega datacenter cloud. Its purpose is to manage resources, perform data filtration, preprocessing, and security measures. To achieve this, Fog requires an effective and efficient resource management framework, which we propose in this paper. Fog has to deal with mobile nodes and IoTs, which involves objects and devices of different types having a fluctuating connectivity behavior. All such types of service customers have an unpredictable relinquish probability, since any object or device can stop using resources at any moment. In our proposed methodology for resource estimation and management through Fog computing, we take into account these factors and formulate resource management on the basis of fluctuating relinquish probability of the customer, service type, service price, and variance of the relinquish probability. With the intent of showing practical implications of our method, we implemented it on Crawdad real trace and Amazon EC2 pricing. Based on various services, differentiated through Amazon's price plans and historical record of Cloud Service Customers (CSCs), the model determines the amount of resources to be allocated. More loyal CSCs get better services, while for the contrary case, the provider reserves resources cautiously.

We describe a scheduler for multiple servers shared among different packet-flows, where each packet-flow may be served by only a subset of available (preferred) servers. The scheduler allocates tokens to flows in a round-by-round manner, where token allocation to flows at the beginning of each round is weighted max-min fair. We present a packet scheduling scheme where when a server becomes free, it is allocated to serve the HOL packet of an eligible flow with the maximum remaining tokens. The scheduling algorithm is applicable even when the capacity of servers are not known a priori and may vary over duration of a round. Numerical examples are given to illustrate that the scheduler itself is weighted max-min fair.

Network Function Virtualization along with Network Service Chaining envision a reduction in the respective cost that end users, service providers, and network operators are experiencing, while providing complete and high quality services. However, the vast range of available services and the service on-demand model, creates dynamic traffic conditions that necessitates a flexible and automatic network platform to redirect traffic according to network conditions. In this paper, we study the problem of deploying service chains, consisting of a number of virtualized network functions (VNFs), in a SDN enabled data center network, where a random number of users are associated with each service chain. To this end, appropriate resource management algorithms are introduced for the placement of VNFs satisfying server affinity and latency constraints. The interconnection of the VNFs is facilitated by an SDN controller, which periodically recalculates the routing paths to adjust to the dynamic traffic conditions.

Random linear network coding (RLNC) has been shown to efficiently improve the network performance in terms of reducing transmission delays and increasing the throughput in broadcast and multicast communications. However, it can result in increased storage and computational complexity at the receivers end. In our previous work we considered the broadcast transmission of large file to N receivers. We showed that the storage and complexity requirements at the receivers end can be greatly reduced when segmenting the file into smaller blocks and applying RLNC to these blocks. To that purpose, we proposed a packet scheduling policy, namely the Least Received. In this work we will prove the optimality of our previously proposed policy, in terms of file transfer completion time, when N = 2. We will model our system as a Markov Decision Process and prove the optimality of the policy using Dynamic Programming. Our intuition points that the Least Received policy may be optimal regardless of the number of receivers. Towards that end, we will provide experimental results that agree with this intuition.

Bloom filter based multicast is a methodology in which bloom filters are used to implement a source routing approach to multicast forwarding, with the goal of addressing forwarding element memory scalability issues in traditional IP multicast deployments. These techniques face their own scalability issues with large multicast group sizes, or networks with a high degree of node interconnectivity, as the necessary bloom filters may be too large to be practically inserted in packet headers. In this work we contribute a technique by which a centralized network controller may leverage stochastically generated overlay networks to reduce the length of in-packet bloom filters used for multicast packet delivery. We evaluate our technique through simulation of false-positive-free filter generation on representative multicast workloads, and find that it achieves a significant reduction in bloom filter lengths (up to 48% in our best case WAN topology scenario, and up to 89% in our best case regular grid topology scenario). Our technique is appropriate for deployment in software defined networks, where the control plane of the network is logically centralized in a network controller, and our technique may be applied with minimal extensions to forwarding hardware.

Network Coding is a packet encoding technique which has recently been shown to improve network performance (by reducing delays and increasing throughput) in broadcast and multicast communications. The cost for such an improvement comes in the form of increased decoding complexity (and thus delay) at the receivers end. Before delivering the file to higher layers, the receiver should first decode those packets. In our work we consider the broadcast transmission of a large file to N wireless users. The file is segmented into a number of blocks (each containing K packets - the Coding Window Size). The packets of each block are encoded using Random Linear Network Coding (RLNC). We obtain the minimum coding window size so that the completion time of the file transmission is upper bounded by a used defined delay constraint.

In Cognitive Radio (CR), secondary radios are allowed to use the spectrum allocated to primary radios (PR) only if the spectrum is sensed to be unused by the PR users. Cooperative spectrum sensing is a sensing methodology in which secondary users send their detection results to a central fusion node for global decision. It has been shown that for single node detection, SNR wall provides a hard threshold on the sensed SNR for robust signal detection. In this paper, asymptotic performance of cooperative spectrum sensing with Hard Decision Combining (HDC) will be studied under noise uncertainty. It is first shown that under noise uncertainty, SNR wall is yet a substantial constraint on the performance of the cooperative spectrum sensing when applying HDC. We then show that k-out-of-n fusion rule with finite k (e.g., “OR” rule) does not result in robust detection of the PR signal, even for SNR values greater than SNR wall. Finally, it is shown that for SNR values greater than the SNR wall, the PR signal can be robustly detected by the k-out-of-n fusion rule with finite n - k (e.g., “AND” rule).

Media content in its digital form has been rapidly scaling up, resulting in popularity gain of cloud computing. Cloud computing makes it easy to manage the vastly increasing digital content. Moreover, additional features like, omnipresent access, further service creation, discovery of services, and resource management also play an important role in this regard. The forthcoming era is interoperability of multiple clouds, known as cloud federation or inter-cloud computing. With cloud federation, services would be provided through two or more clouds. Once matured and standardized, inter-cloud computing is supposed to provide services which would be more scalable, better managed, and efficient. Such tasks are provided through a middleware entity called cloud broker. A broker is responsible for reserving resources, managing them, discovering services according to customer's demands, Service Level Agreement (SLA) negotiation, and match-making between the involved service provider and the customer. So far existing studies discuss brokerage in a narrow focused way. In the research outcome presented in this paper, we provide a holistic brokerage model to manage on-demand and advance service reservation, pricing, and reimbursement. A unique feature of this study is that we have considered dynamic management of customer's characteristics and historical record in evaluating the economics related factors. Additionally, a mechanism of incentive and penalties is provided, which helps in trust build-up for the customers and service providers, prevention of resource underutilization, and profit gain for the involved entities. For practical implications, the framework is modeled on Amazon Elastic Compute Cloud (EC2) On-Demand and Reserved Instances service pricing. For certain features required in the model, data was gathered from Google Cluster trace.

The problem of joint resource allocation and relay selection is studied for bidirectional LTE-advanced relay networks. The bidirectional communication between user equipment (UE) and eNodeB (eNB) is performed via direct transmission, co-operative relaying (CoR), or a combination of network coding (NC) and CoR (NC/CoR). In this paper, an enhanced three-time-slot per cycle time-division duplexing (TDD) scheme is proposed for LTE-Advanced frame architecture to accommodate a hybrid transmission scheme. More specifically, we formulate the problem of joint resource assignment, relay selection, and bidirectional transmission scheme selection as a combinatorial optimization problem with the objective to maximize the total product of backlog and rate (back-pressure principle). Two approaches are considered to solve our combinatorial optimization problem. First, a graph-based framework is proposed in which the problem is transformed into a maximum weighted Clique problem (MWCP). In addition, our problem is also transformed into a three-dimensional assignment problem (3DAP) which is solved using a hybrid ant colony optimization (ACO) algorithm. Using simulations, it is concluded that the hybrid transmission scheme outperforms all conventional nonhybrid schemes. Moreover, the simulation results confirm that while the two proposed solutions provide similar results, the ACO algorithm is faster due to its lower complexity.

In this paper we propose an approach for applying traffic load balancing to multicast traffic through real-time link cost modification in a software defined network (SDN) controller. We present an SDN controller architecture supporting traffic monitoring, group management, and multicast traffic routing. An implemented prototype is described, and this prototype is used to implement shortest path multicast routing techniques which make use of the real-time state of traffic flows in the network. This prototype is evaluated through experimentation in Mininet emulated wide area networks. Evaluation is presented in terms of resulting network performance metrics focusing on the distribution of traffic flows. Our results demonstrate that real-time modification of links costs produces statistically significant improvements in traffic distribution metrics, with an average improvement of up to 52.8% in traffic concentration relative to shortest-path routing. This indicates that SDN enables the use of real-time modification of link cost functions as an effective technique for implementing traffic load balancing for multicast traffic.

Location privacy preservation in multi-channel cognitive radio mobile ad hoc networks (CR-MANETs) is a challenging issue, where the network does not rely on a trusted central entity to impose privacy-preserving protocols. Furthermore, even though the multi-channel CR-MANETs have numerous advantages, utilization of multiple channels degrades the location privacy, by disclosing more information about CRs. In this paper, location privacy is studied for cooperative spectrum sensing (CSS) in multi-channel CR-MANETs.We first quantify the location privacy. Then, we propose a new privacy-preserving distributed cooperative spectrum sensing scheme for multi-channel CR-MANETs. We design a new anonymization method based on random manipulation of the exchanged signal-to-noise ratio (SNR). Afterwards, a coalitional game-theoretic distributed channel assignment is proposed to maximize location privacy and sensing performance over each channel in the network. Simulation results show that the proposed scheme can enhance sensing performance and location privacy over multiple channels.

In this paper, we study the problem of optimal resource allocation in relay-assisted bidirectional LTE-Advanced networks. The bidirectional network consists of User Equipment (UE), Base Station or eNodeB (eNB) and a Relay Node (RN). We model the network by using a Two-Way Relay Channel (TWRC) in which UE and eNB can choose between different transmission schemes: direct transmission, pure CoR (CoR scheme), or via the combination of Network Coding and Cooperative Relaying (NC/CoR scheme). In this paper, an enhanced three-time-slot per cycle time-division duplexing (TDD) transmission scheme is proposed for LTE-Advanced frame architecture to accommodate the hybrid transmission scheme. We formulate the joint problem of subcarrier assignment, relay selection, and bidirectional transmission scheme selection as a combinatorial optimization problem to maximize the system total product of backlog and rate (backpressure principle). The problem is then transformed into a three-dimensional assignment problem which is solved using a hybrid ant colony optimization (ACO) algorithm. The simulation results show that our optimization algorithm significantly outperforms conventional schemes for both SISO and MIMO systems.

Software-Defined Networking has been gaining increasing attention in both the research and the industry communities. Separating the control and data planes has brought many advantages such as greater control plane programmability, more vendor independence, possibility of network virtualization, lowered operational expenses, etc. SDN deployments are possible in a variety of contexts: Enterprise networks, WANs and networks connecting data centers. The realization and operation of an SDN requires the inclusion of a central controller. SDN however raises several concerns for WANs including performance limitations due to the larger propagation delays of control information to and from the controller, the increased work load due to a larger number of network elements and concerns as to performance impacts related to the controller placement. This paper attempts to address some of these issues by examining the effects of using source routing as an alternative to hop by hop in an Internet2 SDN-based WAN production deployment. Our analysis shows that source routing can bring significant gains in SDN convergence performance in WAN environments and reduce the sensitivity in performance due to controller placement.

In cooperative wireless networks, relay nodes are employed to improve the performance of the network in terms of throughput and reliability. However, the presence of malicious relay nodes in the network may severely degrade the performance of the system. When a relay node behaves maliciously, there exists a possibility that such a node refuses to cooperate when it is selected for cooperation or deliberately drops the received packets. Trust establishment is a mechanism to detect misbehaving nodes in a network. In this paper, we propose a trust establishment method for cooperative wireless networks by using Bayesian framework. In contrast with the previous schemes proposed in wireless networks, this approach takes the channel state information and the relay selection decisions into account to derive a pure trust value for each relay node. The proposed method can be applied to any cooperative system with a general relay selection policy whose decisions in each cooperative transmission are independent of the previous ones. Moreover, it does not impose additional communication overhead on the system as it uses the available information in relay selection procedure. Copyright © 2012 John Wiley & Sons, Ltd.

The rising complexity of network anomalies necessitates increased attention to developing new techniques for detecting those anomalies. The majority of current network and security monitoring tools utilize a signature-based approach to detect anomalies. This approach must be complemented with other methods to widen the coverage and speed of anomaly detection. In recent years, a great deal of effort has been spent on studying network traffic anomaly detection techniques by security researchers. Those techniques include the statistical analysis technique referred to as PCA (Principal Component Analysis), clustering and Wavelet-based spectral analysis of network traffic. This paper makes three key contributions to advance the state of the art in network traffic anomaly detection. First, we study the effectiveness of PCA and Wavelet algorithms in detecting network anomalies from a labeled data set known as Kyoto2006+ - providing a useful baseline for future researchers. Second, we propose a novel anomaly detection approach based on a hybrid PCA-Haar Wavelet analysis methodology. The hybrid approach uses PCA to describe the data and Haar Wavelet filtering for analysis. Finally, we study the impact of applying the techniques solely to flow-based traffic summary data to detect network anomalies. The experimental results demonstrate an improved accuracy of the hybrid approach in comparison with the two algorithms individually.

We consider a problem of supplying electricity to a set of N customers in a smart-grid framework. Each customer requires a certain amount of electrical energy which has to be supplied during the time interval [0, 1]. We assume that each demand has to be supplied without interruption, with possible duration between ℓ and r, which are given system parameters (ℓ ≤ r). At each moment of time, the power of the grid is the sum of all the consumption rates for the demands being supplied at that moment. Our goal is to find an assignment that minimizes the power peak - maximal power over [0, 1] - while satisfying all the demands. To do this first we find the lower bound of optimal power peak. We show that the problem depends on whether or not the pair ℓ, r belongs to a “good” region G. If it does - then an optimal assignment almost perfectly “fills” the rectangle time × power = [0, 1] × [0, A] with A being the sum of all the energy demands - thus achieving an optimal power peak A. Conversely, if ℓ, r do not belong to G, we identify the lower bound A̅ > A on the optimal value of power peak and introduce a simple linear time algorithm that_almost_perfectly arranges all the demands in a rectangle [0, A/A̅] × [0, A̅] and show that it is asymptotically optimal.

We consider a problem of scheduling electric power demands in a smart-grid framework. Our model consists of n energy requirements {Ai, ℓi, ri}ni=1, needed to be scheduled in time interval [0,1]. Here Ai is the amount of energy, while ℓiand n are respectively, the left and right constraints on the length of the time period, during which Aihas to be supplied without interruption. The triples are assumed to be i.i.d. random vectors, with A distributed according to some general distribution G, and pair (ℓ, r) distributed uniformly in the region {0 ≤ ℓi≤ r ≤ 1}. Our goal is to find a scheduling policy minimizing the power peak - maximal power over [0,1] - and/or the operational convex cost of the system while satisfying all the demands. The problem becomes very complicated as the number n of demands increases. To address this issue, we consider an asymptotic approach, in which the average amount of energy in each demand is inversely proportional to n, thus keeping the total scheduled amount stable. In this paper we first introduce lower bounds for both types of costs and then introduce a scheduling algorithm, asymptotically optimal in the sense that its cost converges to a corresponding lower bound almost surely, as n increases to infinity. Moreover, the algorithm is on-line (each demand is scheduled at the time its parameters become known) and has fully linear running time.

Cooperative relaying (CoR) and network coding (NC) are two promising techniques for improving the performance of next generation LTE-Advanced networks. In this paper, a dynamic resource allocation scheme is proposed in relay-assisted LTE-Advanced networks which consists of three nodes: User Equipment (UE), Base Station or eNodeB (eNB) and an intermediate Relay Station (RS). In such a network, LTE UEs and eNB can choose between different transmission schemes: direct transmission, pure CoR (CoR scheme) using an intermediate relay station, or via the combination of Network Coding and Cooperative Relaying (NC/CoR scheme). We study the achievable rate regions for direct transmission, CoR and NC/CoR and it is noticed that NC/CoR does not always achieve better performance than CoR or direct transmission. Therefore, a hybrid transmission scheme with adaptive resource allocation is proposed to dynamically choose the best transmission scheme among NC/CoR, CoR or direct transmission. The proposed hybrid scheme determines the best transmission strategy and the optimal resource allocation decision at each bidirectional transmission time-frame based on the system channel information as well as the queue length information. The simulation results show that the hybrid algorithm significantly outperforms conventional schemes for both SISO and MIMO systems.

One of the fundamental components in cognitive radios (CRs) is spectrum sensing. For sensing the wide range of frequency bands, CRs need high sampling rate analog to digital converters (ADCs) which have to operate at or above the Nyquist rate. The high operating rate constitutes a major implementation challenge. Compressive sensing (CS) is a method that may overcome this problem. Sub-Nyquist rate can be used for CS recovery algorithms such as ℓ1-minimization. While boundary information of all frequency sub-bands is available, a more efficient recovery algorithm based on ℓ2/ℓ1-minimization can be used instead of ℓ1-minimization. In cognitive radio systems, network coding could be used for primary users (PUs) to increase packet transmissions. Furthermore, network coding provides a structure for vacant sub-bands of spectrum and makes the spectrum more predictable. Using this information that network coding provides us, we combine ℓ1-minimization and ℓ2/ℓ1-minimization algorithms with network coding for compressive spectrum sensing. Our methods require reduced signal sampling rate and result in improved false alarm (FA) and missed detection (MD) probabilities for idle band detection.

The problem of assignment of K identical servers to a set of N symmetric parallel queues is investigated in this paper. The parallel queueing system is considered to be time slotted and the connectivity of each queue to each server is varying randomly over time and following Bernoulli distribution with a given parameter. Each server is capable of serving at most one packet per time slot (if it is connected and assigned to a queue). At any time slot, each server can serve at most one queue and each queue can be served by at most one server. We assume that the service of a scheduled packet by a connected server fails randomly with a certain probability. The packet arrival processes to the queues are assumed to be i.i.d. and follow Bernoulli distribution with a fixed parameter. For such a symmetric system, i.e., with the same arrival, connectivity and service failure parameters for all the queues, we show that Maximum Weighted Matching (MWM) server assignment policy is delay optimal. More specifically, using stochastic ordering and dynamic coupling techniques we prove that MWM minimizes, in stochastic ordering sense, a broad range of stochastic cost functions of the queue lengths including total queue occupancy (or equivalently average queueing delay).

In this paper, we present a probabilistic analysis of iterative node-based verification-based (NB-VB) recovery algorithms over irregular graphs in the context of compressed sensing. Verification-based algorithms are particularly interesting due to their low complexity (linear in the signal dimension n). The analysis predicts the average fraction of unverified signal elements at each iteration ℓ where the average is taken over the ensembles of input signals and sensing matrices. The analysis is asymptotic (n → ∞) and is similar in nature to the well-known density evolution technique commonly used to analyze iterative decoding algorithms. Compared to the existing technique for the analysis of NB-VB algorithms, which is based on numerically solving a large system of coupled differential equations, the proposed method is much simpler and more accurate. This allows us to design irregular sensing graphs for such recovery algorithms. The designed irregular graphs outperform the corresponding regular graphs substantially. For example, for the same recovery complexity per iteration, we design irregular graphs that can recover up to about 40% more non-zero signal elements compared to the regular graphs. Simulation results are also provided which demonstrate that the proposed asymptotic analysis matches the performance of recovery algorithms for large but finite values of n.

In this paper, we present a new approach for the analysis of iterative node-based verification-based (NB-VB) recovery algorithms in the context of compressed sensing. These algorithms are particularly interesting due to their low complexity (linear in the signal dimensionn). The asymptotic analysis predicts the fraction of unverified signal elements at each iterationlin the asymptotic regime wheren→∞. The analysis is similar in nature to the well-known density evolution technique commonly used to analyze iterative decoding algorithms. To perform the analysis, a message-passing interpretation of NB-VB algorithms is provided. This interpretation lacks the extrinsic nature of standard message-passing algorithms to which density evolution is usually applied. This requires a number of nontrivial modifications in the analysis. The analysis tracks the average performance of the recovery algorithms over the ensembles of input signals and sensing matrices as a function ofl. Concentration results are devised to demonstrate that the performance of the recovery algorithms applied to any choice of the input signal over any realization of the sensing matrix follows the deterministic results of the analysis closely. Simulation results are also provided which demonstrate that the proposed asymptotic analysis matches the performance of recovery algorithms for large but finite values ofn. Compared to the existing technique for the analysis of NB-VB algorithms, which is based on numerically solving a large system of coupled differential equations, the proposed method is more accurate and simpler to implement.

In this paper, we present a new approach for the analysis of iterative node-based verification-based (NB-VB) recovery algorithms in the context of compressive sensing. These algorithms are particularly interesting due to their low complexity (linear in the signal dimension n). The asymptotic analysis predicts the fraction of unverified signal elements at each iteration ℓ in the asymptotic regime where n → ∞. The analysis is similar in nature to the well-known density evolution technique commonly used to analyze iterative decoding algorithms. To perform the analysis, a message-passing interpretation of NB-VB algorithms is provided. This interpretation lacks the extrinsic nature of standard message-passing algorithms to which density evolution is usually applied. This requires a number of non-trivial modifications in the analysis. The analysis tracks the average performance of the recovery algorithms over the ensembles of input signals and sensing matrices as a function of ℓ. Concentration results are devised to demonstrate that the performance of the recovery algorithms applied to any choice of the input signal over any realization of the sensing matrix follows the deterministic results of the analysis closely. Simulation results are also provided which demonstrate that the proposed asymptotic analysis matches the performance of recovery algorithms for large but finite values of n. Compared to the existing technique for the analysis of NB-VB algorithms, which is based on numerically solving a large system of coupled differential equations, the proposed method is much simpler and more accurate.

In this paper, we characterize the stability region of multi-queue multi-server (MQMS) queueing systems with stationary channel and packet arrival processes. Toward this, the necessary and sufficient conditions for the stability of the system are derived under general arrival processes with finite first and second moments. We show that when the arrival processes are stationary, the stability region form is a polytope for which we explicitly find the coefficients of the linear inequalities which characterize the stability region polytope.

Cooperative communication makes use of the broadcast nature of the wireless medium where adjacent nodes overhear the message transmitted by the source and assist in the transmission by relaying the overheard message to the destination. Although cooperative communication brings in significant benefits, it also raises serious security issues to wireless mobile networks. For example, there exists a possibility that a node refuses to cooperate when it is selected for cooperation or deliberately drop the received packets. In wireless mobile networks with cooperative communication, trust management is an important mechanism to monitor such networks for violations of security. In this paper, we propose a trust management method for wireless mobile networks with cooperative communications. Conventional Bayesian methodology is insufficient for the cooperative communication paradigm, as it is biased by the channel conditions and relay selection decision processes. Therefore, we modify the conventional trust management method by incorporating not only the relay selection policy but also the dynamic wireless channel conditions among the source, relays and destination. Simulation results are presented to show the effectiveness of the proposed scheme.

In this work, we investigate the optimal key generation problem for a threshold security scheme in mobile ad hoc networks. The nodes in these networks are assumed to have limited power and critical security states. We model this problem using a closed discrete-time queuing system with L queues (one per node) randomly connected to K servers (where K nodes need to be contacted to construct a key). In this model, each queue length represents the available security-related credits of the corresponding node. We treat this problem as a resource allocation problem where the resources to be allocated are the limited power and security credits. We introduce the class of Most Balancing Credit Conserving (MBCC) policies and provide their mathematical characterization. We prove, using dynamic coupling arguments, that MBCC policies are optimal among all key generation policies; we define optimality as maximization, in a stochastic ordering sense, of a random variable representing the number of keys generated for a given initial system state.

Availability of multiple channels and multiple radio interfaces can lead to substantial improvements in the performance of wireless access networks. The optimal allocation of available channels to the users with multiple interfaces however is not a trivial problem especially in a dynamically varying system. In this paper, we will consider the problem of channel and radio interface management in multi-channel multi-interface wireless access networks. In our model, there is a set of orthogonal channels which is shared among several users. Furthermore, each user is equipped with a fixed number of radio interfaces through which it can communicate with the access point. We also consider exogenous stochastic packet arrivals for each user which may be queued for future transmissions. We introduce a general modelling for multi-channel multi-interface wireless access networks for which we propose a throughput optimal channel/interface allocation policy that stabilizes the system for all the arrival rates strictly inside the stability region. We show that the optimal policy determination is equivalent to finding the maximum weighted matching in a bipartite graph at every time slot. Finally, we characterize the stability region for specific cases of the proposed model. Simulation is used to compare the performance of the optimal policy with some other policies in terms of average total queue occupancy.

The problem of estimating the positions of the sensors in a wireless sensor network is commonly known as the wireless sensor localization problem and has been formulated as a relaxed semidefinite programming problem assuming inter-sensor distance measures corrupted by additive Gaussian noise. In this paper, we assume received signal strength measurements under a spatially correlated lognormal shadowing pathloss model and formulate the corresponding non-convex maximum likelihood distance estimator. We apply a Taylor approximation to the objective function, and then relax the problem to a semidefinite program. The localization performance of the approximation is analyzed and is shown to be satisfactory in the case the shadowing covariance is unknown, and excellent when known.

We investigate an optimal scheduling problem for a discrete-time system of two parallel queues with infinite capacity, sharing two symmetrical servers. This model can be used to study a variety of scheduling problems in wireless networks. At any time slot, a queue can be served by one or two connected servers; the queue-server connectivity is assumed to be random and modeled by a two-state Markov chain. The arrivals to each queue are assumed to be independent and identically distributed. A scheduled packet completes service successfully with a given probability. Otherwise, it has to be retransmitted in a later time slot. The optimal scheduling policy is defined as the server allocation policy that minimizes, in a stochastic ordering sense, the total number of packets in the system. We prove, using a dynamic coupling method, that a "Most Balancing" policy, a policy that attempts to balance the lengths of the two queues, is optimal. We also compare the performance of the optimal policy to that of a few other policies via simulations.

Network capacity region of multi-queue multi-server queueing system with random connectivities and stationary arrival processes is studied in this paper. Specifically, the necessary and sufficient conditions for the stability of the system are derived under general arrival processes with finite first and second moments. In the case of stationary arrival processes, these conditions establish the network capacity region of the system. It is also shown that AS/LCQ (Any Server/Longest Connected Queue) policy stabilizes the system when it is stabilizable. Furthermore, an upper bound for the average queue occupancy is derived for this policy.

Inter-cell interference management is one of the most important issues in next generation wireless networks. As opportunistic beamforming induces more fluctuations in the channel, it can be used as an effective means for inter-cell interference management. In this paper, we present a new scheme to decrease inter-cell interference by utilizing the information of user locations along with the inter-base station cooperation in a system that employs opportunistic beamforming. The performance of the proposed scheme is analyzed and compared with existing schemes. Simulation and analytical results are presented to show the effectiveness of the proposed scheme.

The propagation speed of fast scanning worms and the stealthy nature of slow scanning worms present unique challenges to intrusion detection. Typically, techniques optimized for detection of fast scanning worms fail to detect slow scanning worms, and vice versa. In practice, there is interest in developing an integrated approach to detecting both classes of worms. In this paper, we propose and analyze a unique integrated detection approach capable of detecting and identifying traffic flow(s) responsible for simultaneous fast and slow scanning malicious worm attacks. The approach uses a combination of evidence from distributed host-based anomaly detectors, a self-adapting profiler and Bayesian inference from network heuristics to detect intrusion activity due to both fast and slow scanning worms. We assume that the extreme nature of fast scanning worm epidemics make them well suited for extreme value theory and use sample mean excess function to determine appropriate thresholds for detection of such worms. Random scanning worm behavior is considered in analyzing the stochastic time intervals that affect behavior of the detection technique. Based on the analysis, a probability model for worm detection interval using the detection scheme was developed. Simulations are used to validate our assumptions and analysis.

Detection of slow worms is particularly challenging due to the stealthy nature of their propagation techniques and their ability to blend with normal traffic patterns. In this paper, we propose a distributed detection approach based on the generalized evidence processing (GEP) theory, a sensor integration and data fusion technique. With GEP theory, evidence collected by distributed detectors determine the probability associated with a detection decision under a hypothesis. The collected evidence is combined to arrive at an optimal fused detection decision by minimizing a cumulative decision risk function. Typically, malicious traffic flows of varying scanning rates can occur in the wild, and the difficulty in detecting slow scanning worms in particular can be exacerbated by interference from other traffic flows scanning at faster rates. Our proposed detection technique uses a window-based self adapting profiler to filter detected malicious traffic profiles with scanning rates greater than the low scanning rates we are interested in. Experiments on a live test-bed are used to demonstrate behavior of the technique.

We investigate an optimal scheduling problem in a discrete-time system of L parallel queues that are served by K identical servers. This model has been widely used in studies of emerging 3G/4G wireless systems. We introduce the class of Most Balancing (MB) policies and provide their mathematical characterization. We prove that MB policies are optimal among all work conserving policies; we define optimality as minimization, in stochastic ordering sense, of a range of cost functions of the queue lengths, including the process of total number of packets in the system. We use dynamic coupling arguments for our proof. We also introduce the Least Connected Server First/Longest Connected Queue (LCSF/LCQ) policy as an approximate implementation of MB policies. We conduct a simulation study to compare the performance of several work conserving policies to that of the optimal one. In the simulations we relax some of the mathematical assumptions we required for the analytical proofs. The simulation results show that: (a) in all cases, MB policies outperform the other policies, (b) randomized policies perform fairly close to the optimal one, and, (c) the performance advantage of the optimal policy over the other work conserving policies increases as the channel connectivity decreases.

In this paper, we propose a practical implementation of a location identification approach. In this approach, only the basic assumptions that are realistic in most types of ad hoc/sensor networks are required, which means that this system is implementable in most kinds of ad hoc/sensor networks. By introducing the idea of cell-based location identification method into the system, some drawbacks to the TDOA (time difference of arrival) system are resolved. More importantly, by employing the new theory, the location estimation accuracy of the proposed system is improved without costing extra resources. Finally, simulations show that the proposed Hybrid TDOA (HTDOA) location estimation system performs better in different environments compared with the current TDOA method.

The IEEE 802.17 resilient packet ring standard is a medium access control (MAC) protocol for metro-area ring networks. RPR supports spatial reuse which allows multiple nodes to send traffic at the same time. RPR employs a distributed fairness algorithm to maintain fairness among all nodes in accessing the ring. The performance of the RPR fairness algorithm depends on a number of configurable parameters. In this paper, we provide some recommendations for practitioners on how to configure the fairness algorithm in RPR to achieve a high performance while meeting fairness constraints throughout the ring. The performance metric is the achieved throughput which is translated into revenue for a service provider. We develop criteria which are used to set the parameters. We also study the effect of different parameters on the performance of the fairness algorithm through simulation results.

The emerging IEEE 802.15.4 standard is designed for low data rate, low power consumption and low cost wireless personal area networks (WPANs). Video transmission over such networks is considered an issue since video traffic demands high data rates. In this paper, the TES (transform-expand-sample) methodology is used to model low rate MPEG4 video. The performance of a surveillance video application is evaluated over wireless Zigbee networks. A rate control algorithm (RC-VBR) adapted to MPEG4 variable bit rate (VBR) video coders is studied over Zigbee networks. The algorithm avoids unpredictable rate variations of the VBR coding and removes the coding delay in constant bit-rate (CBR) coders. A region of interest (ROI) encoding is added to the rate control algorithm in order to capture the important parts of the frame which is suitable for IEEE 802.15.4 (Zigbee) networks. Zigbee networks will enable a large number of applications for surveillance networks. The ns-2 simulator is used to test and validate the MPEG4 real video and modeled video over ad hoc Zigbee networks and to test the different algorithms.

In this paper, a discrete stochastic dynamic programming model for the high speed downlink packet access (HSDPA) downlink scheduler when using finite-state Markov channel (FSMC) model is presented. The model then solved numerically (using value iteration) to find the optimal policy for the case of two users sharing the same cell using a 3-state channel. The optimal policy is the one that maximizes the system throughput while maintaining a level of fairness between users. Simulation is used to study the performance of the resulted optimal policy and compared with Round Robin (RR) scheduler. The effects of the allowable number of channel states and the fairness factor on the optimal policy performance were also studied.

In this paper, we present an analytic model and methodology to determine optimal scheduling policy that involves two dimension space allocation: time and code, in high speed downlink packet access (HSDPA) system. A discrete stochastic dynamic programming model for the HSDPA downlink scheduler is presented. Value iteration is then used to solve for optimal policy. This framework is used to find the optimal scheduling policy for the case of two users sharing the same cell. Simulation is used to study the performance of the resulted optimal policy using round robin (RR) scheduler as a baseline. The policy granularity is introduced to reduce the computational complexity by reducing the action space. The results showed that finer granularity (down to 5 codes) enhances the performance significantly. However, the enhancement gained when using even finer granularity was marginal and does not justify the added complexity. The behaviour of the value function was observed to characterize the optimal scheduling policy. These observations is then used to develop a heuristic scheduling policy. The devised heuristic policy has much less computational complexity which makes it easy to deploy and with only slight reduction in performance compared to the optimal policy according to the simulation results.

The nodes with shared-queues at the CDMA2000 data network have inherent rate mismatch and a single level of service. A core node, the packet control function (PCF) is processing power limited. Whereas a feeding node, the packet data serving node (PDSN), has superior processing metrics. A close-loop backpressure solution was proposed at the literature to efficiently control the PCF queue by utilizing PDSN free buffer space during PCF congestion events. Differentiated services (DiffServ) have been designed for Internet to provide multiple levels of network services. We propose a model for providing service differentiation at the CDMA2000 data networks. The model aims to provide service differentiations comparable to the traditional DiffServ model. The proposed improvement is in providing those services under the CDMA2000 structure of tandem nodes with large rate mismatch and a constraint of maintaining the complexity level of the processing limited core node (the PCF). The model uses a combination of backpressure and DiffServ techniques. The backpressure mechanism is used to push congestion from the PCF to the PDSN edge node where superior treatment of differentiated services can be provided to the traffic. The model differentiates services in terms of the relative access to the output link bandwidth and provides distinct handling to traffic at the multiple RED physical queues. It enables differentiation in terms of the achieved relative throughput, packet loss rate, delay, and jitter. We demonstrate the solution's robustness with various traffic scenarios and system topologies. We show that our architecture is effective in providing bandwidth differentiation, as well as throughput, packet drop rate, and average delay preferences.

A common 'finite-burst' mode of wireless links scheduling in CDMA2000 has been shown to cause occupancy oscillations at the bottleneck shared-queue resulting with large overflow-based closely clustered bursts of data-packets drops. The CDMA2000 gateway (PDSN) node is constructed with superior service-rate compared to its downstream core node (PCF). The rate mismatch further allows for traffic load variations and subsequent congestion at the core node. The use of a Xoff/Xon feedback flow control in CDMA2000 was proposed at the 3GPP2. We evaluate the 3GPP2 backpressure proposal for protecting the bottleneck queue during heavy congestion conditions. We devise an adaptive-Xoff/Xon for tandem queues, which extends the traditional Xoff/Xon to provide threshold adaptation according to overflow prediction. The adaptive-Xoff/Xon complements the RED AQM at the bottleneck node, creating a hybrid flow-control model for tandem nodes. Experimental results show that the hybrid flow-control model eliminates packet discards due to overflow at the bottleneck queue, improves throughput, and lowers the overall data packets drop volume. Packets show to not experience backpressure-based delay variation while traversing the tandem queues. An associated cost is low volume of feedback control packets. Larger tandem-queues' average-delays are observed, which result with lower power-function. Hence, degraded combined performances are delivered by the 3GPP2 proposal for backpressure.

Ring Ingress Aggregated with Spatial Reuse (RIAS) has been proposed as a fairness model for Resilient Packet Rings (RPR) where available bandwidth is allocated among traffic flows in max-min sense. In this paper, we show that in a feedback controlled system such as RPR, this may cause severe under-allocation for long-haul flows and may also result in degradation of total ring throughput. To overcome this problem, we introduce the concept of Equal Opportunity (EO) fairness. Application of this model for bandwidth management in RPR is studied and a new stateful distributed algorithm to calculate per-destination fair rates is developed. The proposed scheme is studied both analytically and through simulations. The simulation results show that using the proposed algorithm, intra-station fairness among local flows at every station is significantly improved. In some cases, imbalances up to 55% between long and short flows are completely removed. An improvement of up to 20% in total ring utilization is also observed.

Resilient Packet Ring (RPR), which is standardized as IEEE 802.17, is a new Medium Access Control (MAC) protocol for high-speed metro-area ring networks. RPR supports spatial reuse and, therefore, maintaining fairness among different nodes in accessing the ring bandwidth is a challenging task in RPR. In order to achieve fairness among nodes, a fairness algorithm is employed at each RPR node. When a node is not given access to the ring, it becomes congested. In this case, the congested node calculates a fair rate which is advertised to all upstream nodes contributing to congestion. When upstream nodes receive the control message, they limit the rate of their injected traffic to the advertised fair rate. Consequently, the congested node will be able to add its traffic to the ring. In this paper, we develop an analytical model for fair rate calculation in the standard RPR Conservative Mode fairness algorithm in the parking-lot scenario. This model can be used to evaluate the performance of the fairness algorithm. We investigate this problem in two cases. First, we assume that the link propagation delay is zero. Then, we consider the link propagation delay and develop a more realistic model. The fair rate equations are derived for both cases. We verify the accuracy of our model by simulation results. Furthermore, we use the developed model to study the impact of various parameters on convergence of the fair rate.

Abstract — Content inspection technology is promising to address the growing demands of network security and contentaware networking. The component of a network device responsible for content inspection is called Content Inspection Engine (CIE). In content inspection processing, both headers and payload of packets are parsed to determine their content and classify them based on administrative policies. Due to the inclusion of complex payload processing in content inspection, a large amount of processing is required for each data packet, making content inspection a primary bottleneck in high performance routers that support gigabit link capacities. Therefore, there is a need for solutions that can inspect packets quickly with reasonable amount of storage requirements. We describe an architecture based on high performance state machines which provides an efficient solution to this problem by processing multiple characters per state transition. In this architecture the CIE is implemented on a Two-Level State Machine (TLSM). The TLSM implementation exploits the dependencies of policy rules to compress the policy and reduce memory requirements. Using TCAM units in the TLSM for performing fast multiple-character matching operations, makes a wire-speed content inspection possible. We also propose to use a new criterion called worstcase throughput as an appropriate metric for speed evaluation of CIEs. It is shown that this criterion can be efficiently calculated by applying existing algorithms for the Minimum Weight to Time Ratio problem, to a graph-based model of the functionality of CIEs. I.

We investigate a threshold-based wavelength allocation scheme in order to support fairness and service differentiation in WDM unidirectional ring networks. In a ring network, classes of calls with smaller hop-counts (i.e., the number of hops used from source to destination), experience lower blocking rates than ones with greater hop-counts. In this paper, a multi-threshold (MT) wavelength allocation scheme is proposed to provide equal blocking probabilities experienced by different classes. Simulation results compare the performance of our proposed scheme, with that of complete sharing and complete partitioning schemes.

Resilient packet ring (RPR), which is standardized as IEEE 802.17, is a new medium access control (MAC) protocol for high-speed metro-area ring networks. RPR supports spatial reuse and, therefore, maintaining fairness among different nodes in accessing the ring bandwidth is a challenging task in RPR. In order to achieve fairness among nodes, a fairness algorithm is employed at each RPR node. When a node is not given access to the ring it becomes congested. In this case, the fairness algorithm calculates and advertises a fair rate to all upstream nodes contributing to the congestion point. Consequently, the congested node will be able to add its local traffic to the ring. In this paper, we develop an analytical model for fair rate calculation in the standard RPR fairness algorithm in the parking lot scenario. We first ignore the link propagation delay and model the system using a non-linear discrete-time low-pass filter. We, then, consider the link propagation delay and develop a more realistic model. We verify our model by simulation results and analyze the convergence of the fairness algorithm. Further, the effect of various parameters on the convergence time is investigated. Finally, we determine the low-pass filter coefficient in order to ensure that the convergence time of the algorithm in the parking lot scenario is within its minimum range.

We investigate a threshold-based wavelength allocation scheme in order to support fairness and service differentiation in WDM unidirectional ring networks. A ring network can handle different classes of traffic streams, which differ by their hop-counts (i.e., the number of hops used from source to destination). We assume that for each class of traffic, call interarrival and holding times are exponentially distributed. In such a network, classes of calls with smaller hop-counts, experience lower blocking rates than ones with greater hop-counts. In this paper, a multi-threshold wavelength allocation scheme is proposed to provide equal blocking probabilities experienced by different classes. A recursive simulation-based algorithm is designed to numerically compute the optimal thresholds. Simulation results compare the performance of our proposed scheme, with that of complete sharing and complete partitioning schemes

The increasing need for enhanced data services in cellular networks requires sharing of scarce wireless channel resources among the mobile users by dynamic channel assignments. It has been shown in previous works that such a scheme causes queue management problems in the buffers shared by multiple mobiles, e.g. the input buffer at base station controller (BSC), when their rates are increased after a period of lower aggregate data rate in radio links. Traditionally buffer management techniques like random early detection (RED) are used for a single buffer only. In this paper, we extend the RED algorithm to an aggregated queueing algorithm (AQuA) that simultaneously regulates the queueing discipline in both the shared and individual link buffers so that buffer overflow problems after aggregate link rate increase do not occur. Our algorithm relies on standard information on queueing backlog in link buffers available at BSC to perform buffer management in a unified manner with shared buffer. Furthermore, we demonstrate that our approach provides significantly greater determinism in packet transit delays over BSC and greater throughputs and similar levels of fairness as RED mechanism in shared buffers in conjunction with unregulated link buffers

Resilient packet ring (RPR) is a new medium access control (MAC) protocol for high-speed ring networks. It supports spatial reuse and, therefore, maintaining fairness among different nodes is a challenging task in RPR. To ensure fairness among nodes, a fairness algorithm is employed at each RPR node. In case of congestion, the fairness algorithm advertises a fair rate to all upstream nodes contributing to the congestion. In this paper, we develop an analytical model for fair rate calculation in the standard RPR fairness algorithm in the parking lot scenario. We first ignore the link propagation delay and model the system using a nonlinear discrete-time low-pass filter. We, then, consider the link propagation delay and develop a more realistic model. We verify our model by simulation results and analyze the effect of various parameters on the convergence time. Finally, we determine the low-pass filter coefficient to ensure that convergence time of the algorithm is within its minimum range

One of the main requirements in packet ring networks is to provide fairness in bandwidth allocation among the ring nodes. Each node must receive a fair share of the ring bandwidth and should not starve for an extended period of time. Due to the particular architecture of the packet rings, fairness models such as max-min fairness and proportional fairness are not suitable for these networks. Ring ingress aggregated with spatial reuse (RIAS) is a proposed model for packet rings. However, it lacks generality and intuition. In this paper, a new fairness model for ring networks called ring ingress aggregated max-min (RIAMM) fairness, is proposed. This model is invariant of the source behavior. We have analyzed conditions as well as feasibility criteria for this model. Considering resilient packet ring (RPR) as a particular case, we have studied the effect of source behaviors and fairness algorithms. Three main source behaviors, namely, MF (maximally feasible), FEP (feasible equal partitioning), and SC (single choke) are studied. We show that the FEP source behavior can result in a throughput loss of up to 17%, when traffic disparity exists. It is also shown that fairness algorithms with a slow convergence can result in permanent unfairness during a congestion period.

In order to support high data rate requirements and effectively manage scarce wireless resources, additional bandwidth channels are quite frequently allocated and taken away from mobile stations in 3G wireless data networks. A TCP sender connected to the mobile, on seeing ACKs coming at a faster pace after additional bandwidth allocation, turns overtly optimistic and injects data into the network in a more bursty manner that might be excessive for an intermediate router, thereby leading to loss of multiple packets and subsequent prolonged recovery and periods of underutilization. We characterize this problem using an analytical model for losses based on a continuous flow approximation as well as an extensive simulation setup. We also illustrate how bandwidth oscillations create more severe congestion than an increase in the number of users to the extent that even the RED algorithm is unable to check the sharp growth of queues. As a result, multiple packets are lost in a droptail fashion. We further demonstrate the dependence of congestion due to bandwidth allocation on the time during which mobiles' rates are increased and observe the degradation in performance for typical load scenarios.

Resilient packet ring (RPR) is being devised as part of IEEE 802.17 standard, where fairness in bandwidth allocation among ring nodes, efficiency in resource utilization, and a low computational complexity are the main requirements. Although recent efforts have improved the performance of the RPR fairness algorithms to have acceptable steady-state behavior, we demonstrate that current algorithms suffer from extreme unfairness and throughput loss in some dynamic traffic scenarios. In this paper, we address the bandwidth management in RPR. First, we propose a general fairness model for packet rings. Then, a new algorithm for bandwidth management in RPR called virtual queuing (VQ) is introduced. We study the fairness properties of VQ algorithm both analytically and with simulation results. Compared to the RPR standard fairness algorithms that suffer from a throughput loss of up to 28% in some cases, the throughput loss with VQ is less than 2%. Comparing to another algorithm, called distributed virtual-time scheduling in rings (DVSR), VQ has a lower computational complexity and a better performance in a dynamic traffic environment. We show that the average throttled rate of the head node in a congestion span can be up to 80% for DVSR With VQ, it is less than 4% in all cases.

Resilient packet ring (RPR) is an emerging standard for the construction of local and metropolitan area networks. The priority queue (PQ) algorithm, which is recommended as the scheduling scheme for RPR always gives priority to the transit buffer. Using this scheduling scheme in single transit buffer RPR, the high priority traffic, such as video packets waiting to access the ring at congested node in the transmit buffer will suffer large delays and unsteady delay jitters. In this paper, we propose a new scheduling scheme for RPR to improve the quality of service for video traffic transmission. The proposed scheduling scheme alternately selects packets from the single transit buffer and the high priority transmit buffer using deficit round-robin (DRR) algorithm. If there is no packet in the above two buffers, low priority transmit buffer is then served. We investigate the system performance for transmission of MPEG-4 encoded bitstream as high priority traffic in an RPR network in scenario that all traffic is forwarded to a common node. We report end-to-end delay and delay jitter for I, P and B frames of several encoded video streams. Simulation results show certain improvement on overall delay and delay jitter performance for all types of video frames, especially for I frames which are the most important frames for video reconstruction.

The resilient packet ring (RPR) IEEE 802.17 is a new medium access control (MAC) protocol for high-speed ring networks. In order to achieve fairness among nodes, a fairness algorithm is employed at each RPR node. In this paper, we develop an analytical model for fair rate calculation in the standard RPR aggressive fairness algorithm in parking lot scenario. Our approach is to model the fair rate calculation by a non-linear discrete-time low-pass filter. We verify our model by simulation results and analyze the convergence of the fairness algorithm. Further, the effect of various parameters on convergence time is investigated.

This paper investigates the problem of optimal wavelength allocation and fairness control in all-optical wavelength-division-multiplexing networks. A fundamental network topology, consisting of a two-hop path network, is studied for three classes of traffic. Each class corresponds to a source-destination pair. For each class, call interarrival and holding times are exponentially distributed. The objective is to determine a wavelength allocation policy in order to maximize the weighted sum of users of all classes (i.e., class-based utilization). This method is able to provide differentiated services and fairness management in the network. The problem can be formulated as a Markov decision process (MDP) to compute the optimal allocation policy. The policy iteration algorithm is employed to numerically compute the optimal allocation policy. It has been analytically and numerically shown that the optimal policy has the form of a monotonic nondecreasing switching curve for each class. Since the implementation of an MDP-based allocation scheme is practically infeasible for realistic networks, we develop approximations and derive a heuristic algorithm for ring networks. Simulation results compare the performance of the optimal policy and the heuristic algorithm, with those of complete sharing and complete partitioning policies.

This paper presents our work on developing an architecture for multicasting real-time MPEG4 over IP networks that provide service differentiation. In particular, this work is targeted at assured forwarding (AF) style services. This work is an attempt to find a simple solution to the problem of multicast congestion control of real-time traffic by exploiting the service differentiation capabilities of AF networks. Our architecture assumes loss differentiation in the network and assumes the network's ability to provide explicit congestion notification messages to the sender. We do not consider policing/shaping at the edge routers. Rather, we consider a more general case where packet marking and flow control are provided at the senders. For this network model, we built an end-to-end architecture and developed a rate-adaptation algorithm that can operate in both unicast and multicast applications with a minor modification. The simulation results show how the rate-adaptation algorithm accommodates different receivers with different networking capabilities and provides receivers with different levels of quality by taking advantage of the queue management capabilities of the AF service. We test how the architecture scales to a large number of receivers, how multiple multicast sessions interact, and how it interacts with TCP.

We investigate a call admission control (CAC) mechanism to provide fairness control and service differentiation in a WDM network with grooming capabilities. A WDM grooming network can handle different classes of traffic streams which differ by their bandwidth requirements. We assume that for each class, call interarrival and holding times are exponentially distributed. Using a Markov decision process approach, an optimal CAC policy is derived to provide fairness in the network. The policy iteration algorithm is used to numerically compute the optimal policy. Furthermore, we propose a heuristic decomposition algorithm with lower computational complexity and very good performance. Simulation results compare the performance of our proposed policy, with that of complete sharing and complete partitioning policies. Comparisons show that our proposed policy provides the best performance in most cases. Although this approach is motivated by WDM networks, it may be deployed to determine the optimal resource allocation in many problems in wireless and wired telecommunications systems.

Fair bandwidth sharing among traffic flows with different characteristics in Differentiated Service (DiffServ) networks is the focus of the current research. This paper examines and enhances an algorithm developed to enforce fairness among disparate TCP flows in the assured forwarding (AF) service in DiffServ. equation based marking (EBM) was introduced (M. El-Gendy and K. Shin (2002)) to enforce fairness in AF by monitoring existing network conditions used in marking decisions. The estimation of packet losses by the algorithm is integral to marking. The loss rates of different connections were demonstrated to converge hence enforcing a fair marking regardless of the metrics of individual flows. In this paper, EBM is analyzed for fairness and enhanced by implementing a more efficient technique for loss rate estimation. Comparison is made between EBM and the enhanced technique with results showing appreciable improvements in the maintenance of fairness. Furthermore, a service definition required by QoS standards is met with the implementation of the additional algorithm to EBM.

We investigate a call admission control (CAC) mechanism to provide service differentiation and fairness control in a WDM network with grooming capabilities. A WDM grooming network can handle different classes of traffic streams which differ by their bandwidth requirements. We assume that for each class, call interarrival and holding times are exponentially distributed. Using a Markov decision process approach, an optimal CAC policy is derived to provide service differentiation in the network. The policy iteration algorithm is used to numerically compute the optimal policy. Furthermore, we propose an heuristic decomposition algorithm with lower computational complexity and very good performance. Simulation results compare the performance of our proposed policy with that of complete sharing and complete partitioning policies.

We present an efficient IP packet forwarding technique and its architecture. One forwarding table is decomposed into two balanced smaller sub-forwarding tables by a novel splitting rule. Therefore, an IP lookup can be converted into a pair of small sub-lookups. The output of an incoming packet can be determined by comparing the information, attached to the matching sub-prefixes of both sub-lookups. The sub-lookups and information comparison can perform in parallel. Our approach not only speeds up the Best Matching Prefix (BMP) search, but also reduces storage space at the same time.

In this paper we propose a Virtual Round Robin (VRR) gateway algorithm to enforce per-flow fair bandwidth allocation by keeping per-flow information. This mechanism achieves reasonably fair bandwidth allocation and is easily amenable to high-speed implementations. It uses a single FIFO queue with probabilistic drop-on-arrival. In our simulation study, we compare the performance VRR with other two algorithms, Random Early Detection (RED) and Flow Random Early Drop (FRED) (D. Lin and R. Morris, 1997). Our simulation results show that VRR outperforms RED and FRED in wide variety of scenarios.

This paper outlines an optimal dynamic wavelength allocation in all-optical WDM networks. A simple topology consists of a 2-hop path network with three nodes is studied for three classes of traffic where each class corresponds to different source-destination pair. For each class, call interarrival and holding times are exponentially distributed. The objective is to determine a wavelength allocation policy in order to maximize the weighted sum of users of all classes. Consequently, this method is able to provide differentiated services in the network. The problem can be formulated as a Markov decision process to compute the optimal resource allocation policy. It has been shown numerically that for two and three classes of users, the optimal policy is of threshold type and monotonic. Simulation results compare the performance of the optimal policy, with that of complete sharing and complete partitioning policies.

In this paper, we consider parsimonious Gauusian and table (heavy-tailed) models, which best capture the self-similarity of aggregate packet traffic in broadband networks. Using the effective bandwidths theory, we extend the recent results on stable self-similar-driven queues with infinite buffer to the finite buffer case that model routers/switches more accurately. Large deviations results are extended from the large buffer regime to the many sources limiting regime. Unfortunately, in the stable case, traditional large-deviations formulae degenerate into not very helpful asymptotic results, unlike the Gaussian case. This has a negative impact in engineering considerations (e.g., connection admission control, buffer management, statistical multiplexing gains), with respect to those results, and leads to alternative solutions, e.g. empirical/numerical and simulation techniques.

Wavelength routed all-optical networks require that continuous wavelengths be established from source to destination nodes if no wavelength converters exist in the network. The requests to establish these lightpaths can be blocked depending on the availability of the wavelengths. It has been discovered that lightpaths of no longer length suffer higher blocking probability than those of shorter lightpaths, which is known as fairness problem. The traffic classification and service (ClaServ) method is introduced to optimize the fairness problem as well as reduce the traffic blocking probability. Simulation results for a 4/spl times/4 mesh-torus network and a NSFNET topology show that the ClaServ method can greatly reduce the blocking probability for average network traffic requests.

This paper presents a summary of our work on developing an architecture for transporting real-time traffic (MPEG4 video in this paper) in IP networks that provide service differentiation. We target our architecture at assured forwarding (AF) style services. This architecture assumes loss differentiation in the network and the network's ability to provide ECN messages to the sender. We did not consider policing/shaping at the edge routers. Rather we considered a more general case where marking and flow control are provided at the senders. For this network model, we developed a rate adaptation algorithm that can operate in both unicast and multicast applications with a minor modification. The simulation results presented in this paper represent the multicast case. The results show how the rate adaptation algorithm accommodates different receivers with different networking capabilities and provides them with different qualities by taking advantage of the queue management capabilities of the AF service. We also show the results of testing this architecture with different AF queuing mechanisms, namely RIO and WRED.

In sparse wavelength conversion networks only a few nodes support wavelength conversion. The optical paths in the network consist of a group of segments where each segment independently must meet the wavelength continuity constraint when setting up lightpaths across them. In this paper, we propose a distributed control algorithm called first-available that can efficiently be used to assign wavelengths in networks with sparse wavelength conversion. The wavelength reservation protocol described is a backward reservation protocol. In previous research it has been found that backward reservation algorithms do not offer much improvement in the case where optical converters are used. First-available was compared to other backward reservation algorithms such as first-fit and random and was shown to outperform those in the case of sparse wavelength conversion. Also, compared to the case of no conversion in the network the use of the first-available algorithm in combination with using converters gives a lower average blocking probability. In previous papers, we have outlined a method called OBGP to support lightpath setup and management. We have used OBGP to implement and simulate the first-available algorithm in OPNET. From our simulation results we also collected nodal statistics, and based on these we studied where should be the optimal placement of the converters using the first-available algorithm.

This paper proposes a new algorithm for predicting audio packet playout delay for voice conferencing applications that use silence suppression. The proposed algorithm uses a hidden Markov model (HMM) to predict the playout delay. Several existing algorithms are reviewed to show that the HMM technique is based on a combination of various desirable features of other algorithms. Voice over Internet protocol (VoIP) applications produce packets at a deterministic rate but various queuing delays are added to the packets by the network causing packet interarrival jitter. Playout delay prediction techniques schedule audio packets for playout and attempt to make a reasonable compromise between the number of lost packets, the one-way delay and the delay variation since these criteria cannot be optimized simultaneously. In particular, this paper will show that the proposed HMM technique makes a good compromise between the mean end-to-end delay, end-to-end delay standard deviation and average packet loss rate.

This paper evaluates the suitability of Backward Explicit Congestion Notification (BECN) for IP networks. The BECN mechanism has previously been used in non-IP networks, but there has been limited experimental investigation into the application of the BECN scheme as congestion control mechanism in IP networks. In this paper, we consider an enhanced algorithm for BECN which uses Internet Control Message Protocol (ICMP) Source Quenches for backward congestion notification in IP networks and undertake comparative performance evaluation of Random Early Detection (RED), Explicit Congestion Notification (ECN) and our enhanced BECN mechanism using both long-lived TCP bulk transfers and short-lived Web traffic workloads. Our results show that for Web traffic workloads, BECN offers only slight improvement in transfer delay while average goodput for bulk transfers is no worse than that of ECN. For paths that have a high bandwidth delay product our results show that not only can BECN offer significant improvement in average goodput for bulk transfers over the ECN mechanism, but packet drops and transfer delay for short-lived Web traffic connections are also comparatively reduced. Additional observations show that on such paths TCP (NewReno) with RED can offer higher goodput for bulk transfers compared to ECN. We investigate the overhead due to Source Quenches in a BECN capable network and find that for scenarios considered in this paper it does not significantly impact performance of BECN.

We propose a novel architecture for multicasting adaptively-encoded layered MPEG4 over a QoS-aware IP network. We require such a network to (1) support priority dropping of packets in time of congestion; (2) provide congestion notification to the multicast sender. For the first requirement, we use RED's extension for service differentiation. It recognizes the priority of packets and drops lower priority packets first. We couple RED (random early detection) with our proposal for the second requirement which is the adoption of backward explicit congestion notification (BECN). BECN provides early congestion notification at the IP layer level to the video sender. BECN detects upcoming congestion based on size of the RED queue in the routers. The MPEG4 adaptive-encoder can change the sending rate and divide the video packets into lower priority packets and high priority packets. Based on BECN messages from the routers, a simple flow controller at the sender sets the rate for the adaptive MPEG4 encoder and also sets the ratio of the high priority and low priority packets within the video stream. We use a TES model for generating the MPEG4 traffic that is based on real video traces. Simulation results show that combining priority dropping, MPEG4 adaptive encoding and multicast BECN: (1) improves bandwidth utilization; (2) reduces the time to react to congestion and hence improves the received video quality; (3) maintains graceful degradation in quality with congestion and provides minimum quality even if congestion persists.

In this paper, we present a study of layered video fairness on IP networks. Our study is based on simulation. We investigate some issues that have direct impact on fair allocation of bandwidth between layered video and TCP, in particular: (a) congestion control mechanisms employed by layered video transfer protocols; for this part we studied the interaction of RLM with TCP; (b) the effect of the distribution of video traffic across layers in layered multicast video; (c) the effect of VBR video on fairness to TCP. We show that fairness is affected by all the above factors. We also show that fairness of layered video comes at the expense of instability of the video quality and poor link utilization. We conclude by discussing the performance of layered video protocols in general and recommendations on the design of video transfer systems on IP networks.

This paper proposes a novel acoustic echo cancellation structure intended for multiple participant, full-duplex, hands-free voice over Internet protocol (VoIP) conferencing. A synthetic stereo image is generated through the use of spatialization functions, which are used to assist the near-end user in distinguishing between far-end talkers. The proposed synthetic stereo structure, which uses a single echo canceler per spatial region, is compared to structures with a single echo canceler per channel and to stereophonic echo cancelers. The proposed structure does not suffer from correlation of the reference signals, which can cause system misconvergence in structures that use a single canceler per channel. Correlation of the reference signals is eliminated since monaural signals are transmitted at the far-end and static spatialization functions are used at the near-end to allocate each participant to a spatial region.

We model and simulate stochastic traffic based on two established statistics: marginal distribution and autocorrelation function. The objective or this paper is two-fold: firstly to investigate the issue of modelling input network traffic in an automated way using the QTES methodology. In order to capture both the autocorrelation structure and the marginal distribution simultaneously and secondly to examine the use of different traffic models such as transform expanded sample (TES) models, spatial renewal process, distorted Gaussian model in the output queueing analysis of a system.

We analyze the accuracy of four methods for measuring effective bandwidths (EBs). We point out advantages and drawbacks of the four estimators. We find that for finite time realizations of a process the measured effective bandwidth differs considerably from its analytical counterpart. We also show that increasing the trace length has little impact on the accuracy of the measurements. We explain this behavior as a consequence of the intrinsic characteristics of the observed trace. We subsequently introduce the notion of "empirical effective bandwidth" (EEB) as a measure of performance tailored to the actual values. We derive properties of the EEB which capture its behavior in the parameter space and we contrast these properties with the ones obtained for analytical effective bandwidths. Finally, we comment on the use of EEBs in the context of connection admission control.

This paper proposes a synthetic stereo acoustic echo cancellation structure with microphone array beamforming for multiple participant voice over Internet protocol (VoIP) conferencing. The proposed acoustic front-end combines a synthetic stereo structure, acoustic echo canceler and microphone array beamformer. Synthetic stereo is used to add spatial information to signals from multiple participants that facilitates distinguishing between participants. Microphone array beamforming is then combined with the single canceler per spatial region echo cancellation structure to reduce noise and reverberation for near-end speech pickup and to cancel echo.

The IETF Differentiated Services Working Group has recently standardized the assured forwarding (AF) per hop behavior (PHB). RFC 2597 recommends that an active queue management (AQM) technique be used to realize the multiple levels of drop precedence required in the AF PHB. The most widely used AQM scheme is RED (random early detection). There are several ways to extend RED to a multi-level RED (MRED) algorithm suitable for the AF PHB. This work compares two possible MRED implementations and their ability to protect lower drop precedence traffic: WRED (weighted RED) and RIO (RED with in/out). Based on an empirical study, this paper makes the following key contributions: firstly, the results show that for ON-OFF traffic, RIO is better than WRED in protecting packets marked for treatment with lower drop precedence. Secondly, for short-lived flows, RIO achieves higher transactional rates than WRED. Thirdly, for bulk transfer, RIO and WRED achieve comparable long-term throughput. Finally, this paper also reports the results of experiments with 3 different models for setting of WRED and RIO parameters. We recommend the "staggered" model as best suited to achieve the requirements of the AF PHB.

We propose a minimum cost method for traffic shaping in the context of QoS-based networks. Given the user's desired QoS and the network's resource availability, our procedure determines the least-cost parameters for a traffic shaper which still guarantees access to the network whilst satisfying the QoS constraints. We illustrate our scheme using on-off sources and formulate the QoS constraints by effective bandwidths.

The objective of this paper is to investigate a modeling methodology called QTES (quantized transform-expand-sample) which can be used to model network traffic taking into consideration both the marginal and the autocorrelation function of the empirical data. An effort is made towards an algorithmic procedure rather than a heuristic search, thereby largely automating QTES modeling.

In this paper, DS/CDMA S-ALOHA systems with finite buffer capacity and finite user population are analyzed using the tagged user approach (TUA) proposed in Wan and Sheikh. The data are assumed to arrive in the user buffer in the form of a message containing one or more packets. An arriving message is accepted into the user buffer if the buffer has enough space to hold the whole message, otherwise it is rejected. The queueing analysis for the system is developed in this paper. The analysis is verified by simulation.

In this paper we propose a new technique for access control in an outdoor CDMA cell supporting multimedia traffic. The proposed scheme controls the flow of the traffic at both packet and call level. Delay scheduling at packet level is employed to minimize the variance of intra-cell interference while a call admission control based on a modified equivalent bandwidth technique exploits the minimized interference and reduces the call blocking probability. It is shown that the proposed scheme allows efficient integration of voice and low bit rate video in a CDMA cell with imperfect power control.

An efficient and easy technique to generate fractional Gaussian noise traffic based on the spatial renewal process is developed and demonstrated. The synthetically generated trace reproduces the desired marginal, autocorrelation and Hurst parameters well. The model is particularly suitable for use in the discrete-event simulation of queueing systems involving VBR compressed video and aggregated LAN traffic.

This paper proposes a theoretical model and an ITU-T H.323 compliant empirical measurement method for the acoustic round trip delay (ARTD) of a voice over Internet protocol (VoIP) conference. The empirical measurement method is able to compute the complete acoustic round trip delay that the user perceives. The acoustic round trip delay is experimentally measured by slicing the real-time signal path and inserting test hooks on the Carleton University VoIP platform. Current VoIP systems cannot measure the acoustic round trip delay. An acoustic domain verification test is used to check the ARTD empirical measurements for accuracy.

This paper discusses an ATM congestion control mechanism that introduces a leaky bucket control scheme based on fuzzy logic principles. Network congestion is described linguistically by introducing a fuzzy rule and appropriate fuzzy variables, and is treated mathematically via fuzzy set manipulations. With the application of fuzzy logic the complex mathematical treatment of classical feedback control is avoided, and the "hard" bound effect in the traditional LB is also eliminated in favor of "soft" bound membership functions. In order to evaluate the effectiveness of the fuzzy LB, the performance of the ATM network with a non-fuzzy adaptive LB mechanism is also investigated. Network parameters which affect the performance are identified and optimized for the two control schemes and a comparison of the two approaches is carried out under the same network condition and optimal parameters. Finally, the performance of the fuzzy LB and adaptive LB are also evaluated under self-similar traffic load. The performance analysis in this paper is based on simulation combined with numerical optimization method. Our results indicate that the fuzzy leaky bucket mechanism leads to significant improvement to the system performance.

This paper deals with the problem of admission control when different media with different quality of service requirements (QOS) are present in a CDMA (code division multiple access) environment. We propose an adaptive access control method based on the equivalent bandwidth theory and the cell load measurement. Our algorithm allows efficient integration of speech and low bit rate video in an outdoor CDMA cellular network with imperfect power control.

Examines the transmission of variable bit-rate (VBR) H.261 video over a mixed traffic (video/inter-LAN) integrated services frame relay (FR) network. We introduce a modified H.261 codec that produces VBR output and show that parsing of the video bit stream at group of blocks (GOB) boundaries produces variable length FR packets well suited to the network. We demonstrate that GOB-level video traffic requires a more sophisticated statistical model of the resulting data stream than the frame-level models. The transform expand sample (TES) method is used to obtain an accurate model of the autocorrelation and the marginal probability distribution of the bit-rate variations at the GOB level. A simple and effective methodology is introduced for capturing the periodic components that are present in the GOB-level autocorrelation. The methodology is extended to permit simulations of VBR codecs in which the codec quantization step size is adjusted in response to prevailing network conditions. We show that the quality of service requirements of VBR video can be met by using the FR backward explicit congestion notification (BECN) facility in conjunction with a modified H.261 codec whose rate is controlled by the congestion notification. We also show that the performance of the control mechanism is significantly influenced by a subset of network threshold and codec control parameters which are identified using 2/sup k/ factorial analysis techniques. We obtain optimal ranges of values for these parameters using mean field annealing. Variable quantization rate control may be more effective for this purpose than variable frame rate control.

Variable bit rate (VBR) compressed video is expected to become one of the major loading factors in high-speed packet networks such as ATM-based B-ISDN. However, recent measurements based on long empirical traces (complete movies) revealed that VBR video traffic possesses self-similar (or fractal) characteristics, meaning that the dependence in the traffic stream lasts much longer than traditional models can capture.In this paper, we present a unified approach which, in addition to accurately modeling the marginal distribution of empirical video records, also models directly both the short and the long-term empirical autocorrelation structures. We also present simulation results using synthetic data and compare with results based on empirical video traces.Furthermore, we extend the application of efficient estimation techniques based on importance sampling that we had used before only for simple fractal processes. We use importance sampling techniques to efficiently estimate low probabilities of packet losses that occur when a multiplexer is fed with synthetic traffic from our self-similar VBR video model.

Several services to be supported by emerging high-speed networks are expected to result in highly bursty (autocorrelated) traffic streams. A typical example is variable bit-rate (VBR) compressed video. Therefore, traffic modeling and performance evaluation techniques geared towards autocorrelated streams are extremely important for the design of practical networks.The TES (Transform - Expand - Sample) technique has emerged as a general methodology for modeling autocorrelated random processes with arbitrary marginal distributions. Because of their generality and practical applicability, TES models can be readily used to accurately characterize bursty traffic streams in ATM networks.Although TES models can be easily implemented for simulation studies, the need still exists for analytical results on the performance of queueing systems driven by autocorrelated traffic. Of particular interest are the tails of the waiting time distribution in queues driven by TES-modeled bursty traffic. Such tail probabilities, when they become exceedingly small, may be difficult to obtain via conventional simulation.In order to extend existing results, based on Large Deviations theory, to TES processes, the main difficulty is posed by the continuous state-space of the TES time-series. In this paper, we develop a general result concerning exponential bounds for the waiting time under continuous state-space Markov arrivals. We apply this result to TES/GI/1 queues, show numerical examples, and compare our bound with simulation results. Accurate estimates of extremely low probabilities are obtained by employing fast simulation techniques based on importance sampling.

The performance of a multicast switching system with a cyclic priority input access scheme is evaluated by means of an analytic model which may treat two &tinct multicast switches in a unified way. The performance measure is the packet delay at the input buffer of the system. The analysis is subsequently performed using matrix-geometric techniques and the numericai results are compared with simulations.