Communication Slice Modeling and Optimization with SARSA Reinforcement Learning

Communication Slice Modeling and Optimization with SARSA Reinforcement Learning: In this document, we present a conceptual model of network slicing, we then formulate analytically some aspects of the model and the optimization problem to address. Next, we propose to use a reinforcement learning SARSA agent to solve the optimization problem and implement a proof of concept prototype highlighting its results.

Modeling and Accomplishing the BEREC Network Neutrality Policy

Modeling and Accomplishing the BEREC Network Neutrality Policy

Future Internet and Smart City with OpenFlow and Machine Learning - Course Poster

Future Internet and Smart City with OpenFlow and Machine Learning - Course Poster

Reconfiguração de Redes de Distribuição de Energia Elétrica Utilizando Aprendizado de Máquina

Reconfiguração de Redes de Distribuição de Energia Elétrica Utilizando Aprendizado de Máquina: Electrical networks are composed of stages of generation, transmission, and distribution of energy. Distribution networks (RD) are an important element of the electricity grid because it provides the effective delivery of energy to end users. The RD’s are subject to failure and their optimization is of fundamental importance in the context of Smart Grids, where it is sought a greater efficiency of the processes involved between the production and distribution of energy. The distribution networks (RD) have topologies and loads of various type. This dissertation proposes a method and algorithm for the reconfiguration of electrical network using machine learning with linear regression and branch exchange algorithm aiming the optimization of RD operation. The method and algorithm proposed do maneuvers in the RD as transfer and load balancing and aiming to increase its level of reliability. The proposal is validated in a test network of IEEE (IEEEbus14) using simulation and testing environment implemented in “R” language and using the Newton Raphson method to calculate the power flow. The solution developed show satisfactory in supporting the decision-making for three reconfigurations of distribution networks in the context of the Smart Grid.

COGCOM Project -- Cognitive Allocation of Network Communication Resources for Edge/Fog/Cloud Computing

COGCOM Project -- Cognitive Allocation of Network Communication Resources for Edge/Fog/Cloud Computing

BAMSim Simulator

BAMSim Simulator

Uma Proposta para a Gerência Autonômica e Escalável de Redes de Computadores

Uma Proposta para a Gerência Autonômica e Escalável de Redes de Computadores

Illustrated Technical Paper - Enhanced Pub/Sub Network Communication

Illustrated Technical Paper - Enhanced Pub/Sub Network Communication: This illustrated technical paper presents the slides describing the contents of the paper 'Enhanced Pub/Sub Communications for Massive IoT Traffic with SARSA Reinforcement Learning'. The talk was presented at the 3rd International Conference on Machine Learning for Networking (MLN'2020), 24 - 26 November 2020 at Paris, France - http://www.adda-association.org/mln-2020. The illustrated technical paper format is intended to complement, enrich and subsidize the technical paper content and contains slides, complementary text and additional and/or focused bibliographic references.

Enhanced Pub/Sub Communications for Massive IoT Traffic with SARSA Reinforcement Learning

Enhanced Pub/Sub Communications for Massive IoT Traffic with SARSA Reinforcement Learning: Sensors are being extensively deployed and are expected to expand at significant rates in the coming years. They typically generate a large volume of data on the internet of things (IoT) application areas like smart cities, intelligent traffic systems, smart grid, and e-health. Cloud, edge and fog computing are potential and competitive strategies for collecting, processing, and distributing IoT data. However, cloud, edge, and fog-based solutions need to tackle the distribution of a high volume of IoT data efficiently through constrained and limited resource network infrastructures. This paper addresses the issue of conveying a massive volume of IoT data through a network with limited communications resources (bandwidth) using a cognitive communications resource allocation based on Reinforcement Learning (RL) with SARSA algorithm. The proposed network infrastructure (PSIoTRL) uses a Publish/ Subscribe architecture to access massive and highly distributed IoT data. It is demonstrated that the PSIoTRL bandwidth allocation for buffer flushing based on SARSA enhances the IoT aggregator buffer occupation and network link utilization. The PSIoTRL dynamically adapts the IoT aggregator traffic flushing according to the Pub/Sub topic's priority and network constraint requirements.

Smart City Drivers and Challenges in Energy and Water Systems

Smart City Drivers and Challenges in Energy and Water Systems: The IEEE Smart Cities Initiative brings together the IEEE’s broad array of technical societies and organizations to advance the state of the art for smart city technologies for the benefit of society and to set the global standard in this regard by serving as a neutral broker of information amongst industry, academic, and government stakeholders. These smart city technologies draw upon expertise in several functional domains including: ● Sensors and Intelligent Electronic Devices ● Communication Networks & Cyber Security ● Systems Integration ● Intelligence & Data Analytics ● Management & Control Platforms Together, this functional expertise serves to achieve the mission of the IEEE Smart Cities initiative: 1. To be recognized as the authoritative voice and leading source of credible technical information and educational content within the scope of smart cities identified below. 2. To facilitate and promote both the collaborative and individual work of its Member societies regarding smart city technology. To that end, the IEEE Smart Cities initiative has identified several application domains in which to apply its expertise. These are: ● Smart energy systems ● Smart water systems ● Smart mobility systems ● Smart healthcare systems Each of these systems has generally developed in its own right in response to the needs and context of the domain. Each faces its own set of drivers and challenges. And yet, as each of these sys...

Smart City Drivers and Challenges in Urban-Mobility, Health-Care, and Interdependent Infrastructure Systems

Smart City Drivers and Challenges in Urban-Mobility, Health-Care, and Interdependent Infrastructure Systems: The IEEE Smart Cities Initiative brings together the IEEE’s broad array of technical societies and organizations to advance the state of the art for smart city technologies for the benefit of society and to set the global standard in this regard by serving as a neutral broker of information amongst industry, academic, and government stakeholders. These smart city technologies draw upon expertise in several functional domains including: ● Sensors and Intelligent Electronic Devices ● Communication Networks & Cyber Security ● Systems Integration ● Intelligence & Data Analytics ● Management & Control Platforms Together, this functional expertise serves to achieve the mission of the IEEE Smart Cities initiative: 1. To be recognized as the authoritative voice and leading source of credible technical information and educational content within the scope of smart cities identified below. 2. To facilitate and promote both the collaborative and individual work of its Member societies regarding smart city technology. To that end, the IEEE Smart Cities initiative has identified several application domains in which to apply its expertise. These are: ● Smart energy systems ● Smart water systems ● Smart mobility systems ● Smart healthcare systems Each of these systems has generally developed in its own right in response to the needs and context of the domain. Each faces its own set of drivers and challenges. And yet, as each of these sys...

FAB LABs: Estímulo à Inovação, Usando a Fabricação Digital

FAB LABs: Estímulo à Inovação, Usando a Fabricação Digital: The analyzis of several collaborative manufacture spaces – Maker Spaces – like Fab Labs, Hackerspaces, and Techshops, and its social and economic impact in the environments where they are inserted will be fulfilled. The advantages of these spaces in the insertion of individuals in the world of technique, without discrimination of gender, qualifications, training, race or religion will also be analyzed. The insertion of women in these spaces, and the training of adults and youngsters who did not follow studies at the university will be highlighted. The goal of this work is to study the creation of collaborative spaces for innovation in the exploration of the creative capacities of their communities, focused on spaces of digital manufacturing like Fab Labs. The results leading to innovation created in these spaces, as well as its various stakeholders will be described. Collaborative learning and building a community enhance the sharing of experience and knowledge that leverage the creation of new inventions with innovation potential. The utilization of digital manufacturing technologies, available in these spaces, together with the knowledge available and shared in the network where they belong, make possible to quickly verify the acceptance of the solutions conceived and measure the public interest as well as its critiques and suggestions of improvement. Some examples of advantages and solutions promoted by these spaces will be presented. In only one year, Ford Motors impro...

A Methodological Approach to Model CBR-Based Systems

A Methodological Approach to Model CBR-Based Systems

Federated Persistent Data Repository Approach for FIBRE multi-CMF Testbed

Federated Persistent Data Repository Approach for FIBRE multi-CMF Testbed: A Network for Experimentation (NfExp) is typically a regional, national, or specialized international testbed like GENI, PlanetLab, and OFELIA that are developed to support the experimentation of new protocols, services and applications mostly in the context of Future Internet (FI). In effect, a Network for Experimentation allows an experimenter to configure, instantiate, run, and collect measurements of distributed experiments on top of a real testbed managed by a Control and Monitoring Framework (CMF). Currently available CMFs developments are directly or indirectly service-oriented architectures focused, as an example, on wired or wireless networks research domains. A new approach for the Network for Experimentation considered in this paper is the FIBRE multi-CMF testbed. In FIBRE, various existing CMFs are integrated to achieve a broader set of supported applications and service domains. Among the technical challenges involved in the multi-CMF approach, this paper describes the approach adopted for the persistent data repository necessary to store the monitored and collected experimental data.  An iRODS-based federated monitoring persistent data repository proposal is presented for the FIBRE multi-CMF testbed. The solution adopts a standardized monitoring data format and representation whenever possible. It is a transparent data storage facility among multiple native DBMS, with persistent monitoring data access and retrieval, and a high-level monitoring data integration.

Network Self-Management Pseudo-Predictability based on Case-Based Reasoning

Network Self-Management Pseudo-Predictability based on Case-Based Reasoning: Network self-management, among other objectives, looks for solutions (systems, frameworks and/or strategies) in which the human intervention is drastically reduced in the analysis and planning of new network states, which in turn, is based on a set of SLAs (Service Level Agreement) definitions and traffic profile. The self-management system/framework has to find a solution meeting SLA and traffic requirements within a reasonable computational time (execution time) in order to achieve a near real time operation. In general, it is necessary to adopt approaches based on reasoning and learning in self-management systems in order to avoid and/or reduce human intervention [1]. As such, self-management systems/frameworks do require the storage of network states. In this context, it is important to identify critical states and, if possible, to predict when they will occur in order to facilitate the self-management process. In this paper, a proposal for pseudo-predictability based on Case-Based Reasoning (CBR) is presented. The pseudo-predictability CBR method considers previous network states to infer future network state scenario. It was developed for the self-management framework defined in [2] and, in general, further promotes the scalability of self-managed systems.

An Instrumentation and Measurement Architecture Supporting Multiple Control Monitoring Frameworks

An Instrumentation and Measurement Architecture Supporting Multiple Control Monitoring Frameworks

A (R)evolução das Redes sem Fio

A (R)evolução das Redes sem Fio: A evolução e implícita revolução causada pela utilização das redes sem fio é discutida através de uma apresentação de suas tecnologias mais relevantes e seus domínios de aplicação atuais e futuros.

Use of Artificial Intelligence Techniques in the Implementation of a Proactive Management System for ATM Networks

Use of Artificial Intelligence Techniques in the Implementation of a Proactive Management System for ATM Networks

Avaliação de Protocolos Multicast

Avaliação de Protocolos Multicast: With the development of new computer network applications based on TCP/IP protocols we need to improve the performance of these protocols to match the new application demands. In order to provide these enhancements, research and developments on TCP/IP is under way for new mechanisms to reduce the use of bandwidth as much as possible. Accordingly, a new concept called multicast traffic has been developed to transmit data from one location to many others. This paper evaluates some multicast protocols that are being developed currently with respect to the routing mechanism focusing on the average delay time, the behavior of the routing tables, the use and overloading of the links. This allows to choose the most adequate protocol given the requirements of each topology of the network and the resources available.

Simulação de Protocolos Multicast

Simulação de Protocolos Multicast