Network design and allocation considerations in the Hughes data-flow machine

Enhancement of wireless sensor network lifetime by deploying heterogeneous nodes

Network performance metrics such as available bandwidth and latency are essential to achieve good Quality of Service (QoS) in multimedia streaming. There are unique requirements in network performance metrics for media applications, such as audio conferencing, video streaming, video conferencing, and high-definition (HD) video conferencing. In this paper, we focus on conference call type suggestion based on link quality prediction. The link's quality is classified based on the available bandwidth and latency between two network nodes. We have implemented and compared two of the most popular supervised learning based classification methods, i.e. logistic regression and support vector machine (SVM). We have compared the performance of both methods and their suitability to apply in link quality prediction. The experimental results show that SVM outperforms logistic regression for binary and multiclass classification in terms of accuracy.

This paper quantifies the impact of topological characteristics on the performance of single radio multichannel IEEE 802.11 mesh networks. Topological characteristics are the number of nodes per subnetwork, the hop count, the neighbor node density, the hidden nodes, the number of nodes in the neighborhood of the gateway, and the hidden nodes in the neighborhood of the gateway. Network performance metrics are throughput, fairness and delay. The data mining Support Vector Machine (SVM) model was used to extract the relationships between the network topology metrics and the network performance metrics based on data results obtained through ns-2 simulation of random networks. The results obtained can be used as a basis to design channel assignment algorithms or to aid the deployment and management of single radio wireless mesh networks.

The design and maintenance of the radio access network (RAN) in mobile telecommunications requires meticulous monitoring of network performance to ensure optimal quality and coverage.Drive testing is a prevalent methodology for collecting data on network status across delineated geographical areas.This process involves systematically traversing various routes, including streets and pathways, to assess network performance metrics such as Reference Signal Received Power (RSRP).Although drive testing provides essential insights for identifying regions with inadequate signal quality, it is inherently resource-intensive, involving considerable time and financial expenditures.This paper proposes an innovative optimization methodology aimed at enhancing the efficiency of data collection during drive tests.The proposed approach is organized into four fundamental steps: (1) partitioning the map into smaller sections, (2) selecting critical points within each section, (3) employing map-matching techniques to accurately align these points with actual streets and pathways, and (4) determining the optimal routefor traversing the critical points.A rectangular area of interest is selected and divided into K smaller sub-regions, within which M critical points are identified according to a predefined criterion.These points, which may not initially correspond with the existing street network, are corrected through mapmatching techniques to ensure feasible traversal paths.Lastly, an optimization algorithm is utilized to compute the shortest route that encompasses all identified critical points.The efficacy of the proposed method is assessed through experimental studies that manipulate key parameters K (denoting the number of sub-regions) and M (indicating the maximum critical points per zone).The evaluation emphasizes two critical metrics: the total distance traveled and the success rate in detecting areas with RSRP values below -100 dBm.Results indicate that the proposed approach significantly decreases the distance required for drive testing while achieving substantial coverage of areas with weak signals.For example, in an experiment where the optimized route covered only 18.54 kilometers-equivalent to 34% of the distance of an entire drive test-it successfully identified 70% of regions with poor RSRP.Furthermore, this paper introduces a criterion for assessing the effectiveness of drive test routes, highlighting the balance between route length and coverage of low-signal areas.The findings substantiate that adequate data for network performance analysis can be secured with considerable savings in both cost and time compared to conventional exhaustive drive testing.While this study concentrates on RSRP measurements within 4G networks, the proposed methodology is adaptable for other metrics, including Reference Signal Received Quality (RSRQ).Future research may also examine the integration of alternative data sources, such as satellite imagery, to further refine map partitioning and critical point selection, thereby enhancing the overall efficacy of the proposed method.

In 5G networks, time-series data will be omnipresent for the monitoring and management of network performance metrics. With the increase in the number of Internet of Things (IoT) devices, it is expected that the number of real-time time-series data streams will increase at a fast pace, making forecasting essential for the proactive successful management of the network. In this article, we discuss to use both linear and non-linear forecasting methods, including machine learning, deep learning, and neural networks to improve 5G networks’ management. For this purpose, we design and implement a real-time distributed forecasting framework, used to make simultaneous predictions of different network performance metrics, and with different learning algorithms. By using our framework, we compare the use of forecasting methods in two network scenarios, in a real vehicular network and in a 4G network, representing two different slices in a 5G network. We also integrate our framework in a 5G architecture. Using the best forecasting models assessed previously, we propose a dynamic threshold algorithm for multi-slice management, to ensure that the resources of each slice are updated according to the slices’ needs, while avoiding congestion and saving resources for other slices. The experimental results show that it is possible to forecast the slices’ needs and congestion probability, selecting the best forecasting approach or an ensemble of the best ones, and act accordingly in the network to optimize its management.

A GRASP-based network re-optimization strategy for improving RWA in multi-constrained optical transport infrastructures

Several active queue management (AQM) schemes have been proposed to address the bufferbloat issue, a long network delay and performance degradation issue caused by excess buffering of packets in network nodes. The solutions proposed up to now, however, satisfy only a subset of the key network performance metrics, i.e., end-to-end throughput, throughput fairness, and delay, in 802.11s-based multi-radio multi-channel wireless mesh networks (MRMC WMN). In this letter, we propose FD-AQM, a fairness-aware delay-controlled AQM to satisfy all three key network performance metrics simultaneously in the 802.11s-based MRMC WMN. The proposed FD-AQM achieves the goal by adopting a per-queue drop interval which is recalibrated according to the average round-trip time of flows, and a drop decision mechanism which considers average throughput as well as target delay. Through simulations, we show that FD-AQM is superior to other AQM schemes in terms of throughput and fairness.

Wireless sensor network performance metrics for building applications

With the rapid development of Internet of Things (IoT) technology, the number of nodes in wireless sensor networks (WSNs) is explosively increasing, and the scale of network is increased gradually. Traditional single-layer non-clustering network is no longer suitable for current WSNs, which results in high maintenance cost and fast deterioration of network performance. By analyzing the impact of existing static and dynamic clustering schemes on network performance, it is concluded that additional factors need to be considered to improve the overall performance of the network, such as residual energy of nodes, number of neighbor nodes and load balancing. Therefore, an adaptive multi-layer clustering networking strategy based on capability weights is proposed. Based on the real-time changes of each cluster density, node load and residual energy, the node capacity weights are updated dynamically according to the actual network performance, then the cluster heads are renewed adaptively. By comparing the performance metrics in the experiments, proposed strategy can effectively reduce the load of key nodes such as cluster head, and improves the network performance metrics such as average transmission delay, average transmission hops and load balancing.

The Software-Defined Networking technology promises to enhance network performance and cost reduction for service providers by providing scalability, flexibility, and programmability through the separation of the control plane from the data plane. However, the separation between the control plane and data plane in the implementation of SDN presents scalability issues, as the controller has limited computational resources. To address the SDN scalability issues identified, we create a scalable hybrid switching solution using machine learning algorithms. We propose an SDN OpenFlow model switch which collaborates with the traditional switch to represent a scalable framework of Hybrid Routing with Reinforcement Learning (sHRRL). We implement a reinforcement algorithm to randomly explore new routes and discover the most optimal path through the Q-learning algorithm. This primitive and model-free form of reinforcement learning utilizes the Markov Decision Process and Bellman’s equation to reiteratively update Q-values in the Q-table for every transition in the network environment state until Q-function has converged to the best Q-values. The proposed hybrid switching model is benchmarked against the standard SDN OpenFlow switch in terms of network performance metrics, including throughput, packet exchange transmission rates, CPU load, and delay. When statistically comparing simulation results, it is evident that the proposed switching model, incorporating machine learning algorithms, can effectively tackle scalability challenges in the design of SDN controller networks, especially in Data Centre environments where rapid switching speeds are crucial.

We categorized VPN’s impact in to two different aspects; The affecting aspect, and the affected aspect. The affecting aspect of the impact encompasses factors such as security, algorithms, hardware, and software. Whereas the affected aspect is of the likes of network performance. While VPNs have managed to integrate security, one of the affecting aspect of the impact on one hand, on the other hand, VPNs should be regarded as a potential threat to network performance. In this study, for affordability purpose, we choose to use NS-2 simulated test-bed to shed light on the VPN’s performance impact in a network. Considering the most common network performance metrics, throughput and delay; we assessed these performance metrics by means of average and percentage changes theories. The findings emphasize quantitative impact on the TCP/IP throughput than on its counterpart UDP/IP. We finally developed an analytical equation to model this VPN’s performance impact.

Το αντικείμενο της διατριβής αυτής είναι η μελέτη της διαστρωματικής (cross-layer) προσέγγισης ανάπτυξης ασύρματων δικτύων κατανεμημένης λειτουργίας με απαιτήσεις επικοινωνίας πραγματικού χρόνου και περιορισμένους διαθέσιμους πόρους. Επιπλέον βασικό στόχο αποτελεί και η σχεδίαση, πρόταση αντίστοιχης αρχιτεκτονικής η οποία στοχεύει στη βέλτιστη διαχείριση διαθεσίμων δικτυακών πόρων σε καταστάσεις συμφόρησης του δικτύου και κατά συνέπεια στην βελτίωση της απόδοσης αυτού. Μέσα από μελέτες στα πλαίσια της διατριβής αναδεικνύεται η σημασία του φαινομένου συμφόρησης ως κύριου παράγοντα σπατάλης δικτυακών πόρων καθώς και οδήγησης του δικτύου σε κατάσταση κορεσμού με αρνητική επίπτωση σε όλες τις παραμέτρους απόδοσης του δικτύου. Στόχος, λοιπόν, της προτεινόμενης επικοινωνιακής αρχιτεκτονικής είναι η αποφυγή του φαινομένου συμφόρησης έτσι ώστε το δίκτυο να οδηγείται δυναμικά σε ένα σταθερό σημείο απόδοσης (όσο αυτό είναι δυνατό) το οποίο θα επιτρέπει στο δίκτυο να αποδίδει όσο το δυνατόν καλύτερα αποφεύγοντας τη ίδια στιγμή άσκοπη σπατάλη πόρων. Βασική παράμετρος στη σχεδίαση αποτελεί η συμβατότητα τόσο ως προς σημαντικό εύρος διαθέσιμων πρωτοκόλλων σε διάφορα επίπεδα, όσο και ως προς τη δυνατότητα συνύπαρξης κόμβων που ενσωματώνουν την προτεινόμενη αρχιτεκτονική με κόμβους χωρίς αυτήν. Η προτεινόμενη αρχιτεκτονική υλοποιήθηκε στα πλαίσια γνωστού και αξιόπιστου δικτυακού εξομοίωση. Αξιολόγηση της υλοποίησης αυτής μέσα από μεγάλο αριθμό πειραμάτων έδειξε επίτευξη του στόχου καθώς το φαινόμενο της συμφόρησης αντιμετωπίστηκε σε όλες τις περιπτώσεις με σημαντικά οφέλη στην απόδοση του δικτύου και διαχείριση των πόρων. Επιπλέον, λόγω του μεγάλου βαθμού παραμετροποίησης αποτελεί ιδανική βάση για μελλοντικές προεκτάσεις.

The exponential growth of distributed artificial intelligence training workloads has created unprecedented challenges in network performance management and optimization. Traditional monitoring approaches fail to adequately address the dynamic and complex communication patterns inherent in distributed AI systems. This paper presents a comprehensive machine learning-based framework for network performance monitoring and prediction specifically designed for distributed AI training environments. Our approach leverages advanced feature engineering techniques to capture multi-dimensional network performance metrics, including latency variations, bandwidth utilization patterns, and communication bottlenecks. We develop and evaluate multiple machine learning models, including gradient boosting machines, neural networks, and ensemble methods, to predict network performance degradation before it impacts training efficiency. Experimental evaluation on real-world distributed training scenarios demonstrates that our framework achieves 94.7% prediction accuracy for network latency spikes and reduces training time by 23.4% through proactive performance optimization. The proposed monitoring architecture provides real-time insights into network behavior patterns, enabling adaptive resource allocation and communication scheduling optimization. Our contributions include a novel feature extraction methodology, a comprehensive performance prediction model, and a scalable monitoring framework that can be deployed across heterogeneous distributed computing environments.

Microgrids help to achieve power balance and energy allocation optimality for the defined load networks. In this paper, the focus is to enhance the intelligence of the microgrid network using a multi-agent system and validation is using network performance metrics such as delay, throughput and jitter. Network performance is analyzed for the medium-scale microgrid using two IEEE test systems, i.e., IEEE 34 and IEEE 39. In this paper, Bellman–Ford algorithm is incorporated to calculate the shortest path to a given destination. The algorithm is defined for the distributed nature of the microgrid. From this model, researchers have achieved up to 30% improvement in the network performance of a microgrid.

In wireless sensor networks (WSNs), preserving energy requires utmost attention, as they are highly resource constrained. Clustering is commonly considered as one of the efficient energy conservation techniques. In a clustered WSN, due to the nature of operation, cluster heads (CHs) near the sink bear the major share of data forwarding compared with the CHs far away from the sink, resulting in an energy hole problem. First, we have analyzed the optimization of network lifetime by balancing the energy consumption among different CHs. To meet the requirement of optimization of network lifetime, we have devised a routing-aware optimal clustering strategy. Further, keeping the network model in mind, we identify Archimedes' spiral, based on which a deployment function is proposed for distributing a member node and a CH. The performance of the optimal clustering strategy is evaluated in terms of energy balance and network lifetime. Simulation results show that our scheme provides satisfactory network performance in terms of end-to-end delay and throughput. Finally, all the results are compared with two competing schemes that confirm our scheme's supremacy in terms of both design performance metrics as well as network performance metrics.