Quality of service framework for supporting adaptability and distributiveness in IOT environments

Abstract

Internet of Things (IoT) is one of the fast growing technological paradigm which requires a novel newlineand more optimal solutions in terms of architecture, standards, protocols, infrastructure newlinedeployment, Quality of Service (QoS), Service Level Agreements (SLAs), service provisioning, newlinecross domain and cross platform implementations. At present, IoT involves the techniques and newlinetechnologies for sensing, actuation, communication, computation, networking and storage. In such newlinea dynamic and more emergent technological demanding environment the need for cross layer QoS newlinefunctionalities are essential to address the issues like resources, mobility, security and energy newlinemanagement. From the detailed review of literatures on IoT architectures and QoS newlineimplementations, service provisioning standards it is observed that there is a need for continued newlineresearch and solutions for improving QoS in IoT. newline newlineOne of the approach to address the above challenge(s) in an IoT environment requires an newlineappropriate lathering of functional modules to different layers to meet different QoS requirements. newlineHence a novel cross layer QoS framework supporting adaptable and distributed decision making newlineis proposed in the IoT environment as a cross layer implementation addressing the QoS parameters newlinelike energy, bandwidth, latency, security and more. For implementing the adaptability and decision newlinemaking, it is critical to identify and classify the functional modules as adaptable and non-adaptable newlinemodules. The proposed model does an appropriate lathering of functional modules to the newlinecorresponding IoT layer for QoS improvement. Some of the concepts are applied newlineto ident and newline newlinemutually exclusive with respect to the functions of other layer are layer specific and are non- newlineadaptable . newline newlineThe proposed cross layer QoS framework for IoT which uses computational models like decision newlinemaking, prediction model and security model. These three different model approaches used are newlinebriefed here. newline(i) Many approaches of QoS implementations use task specific algorithms. In this research newlinework the layered architecture uses simple decision support system at perception layer that newlineconsiders the QoS improvement with specific QoS parameters like bandwidth and latency. newline newlineIX newline(ii) Machine Learning (ML) approaches are used as one of the key technologies and computing newlinemethods in the recent works on Quality of Service in the IoT environments for improvising newlineQoS performance and solutions. ML techniques are adopted in most of the domain and newlinetechnology areas. An ML based prediction model is proposed and designed for newlineoptimization of resources for QoS provisioning in the IoT environment. The proposed newlinemulti-layer neural network model (MNN) model for prediction uses Long Term Short newlineMemory (LSTM) learning approach. The prediction model considers energy and newlinebandwidth as QoS parameters by prediction and optimized resources utilization in the IoT newlineenvironment. The proposed model performance is evaluated in a real field implementation. newlineFrom the performance of the prediction model is observed that there is an improved newlinebandwidth and energy utilization there by providing the required QoS in the IoT newlineenvironment. newline newline(iii) Increased adoptions of Internet of Things (IoT) in multiple domain areas continuously newlineincreases the volume of transactions and data from the IoT devices. Securing the data newlineacross the devices, edge nodes, communication layers and applications is more critical. In newlinethe recent there are many attacks which are caused by IoT devices and IoT systems. IoT newlinesystems are also vulnerable for attacks because of increased technology adoptions. newlineProtecting the data and providing security to the data and transactions is one of the critical newlineissue for effective quality of service (QoS) in IoT environment. Through review work newlinetowards the security issues and threats for IoT systems and security solutions is carried out newlinespecifically using machine learning and blockchain technologies for IoT systems. newlineAdaptable IoT framework for security that leverages the blockchain technology and newlineapplicable features is proposed with security as one of the QoS parameter. This includes newlinelayered distributed architecture using blockchain, applying cryptography on edge nodes in newlinethe perception layer, and distributed ledger and consensus on the application nodes. The newlinedesign of the proposed model uses ethereum, an open source public blockchain and results newlineverified for security factors and computing resources. newline newlineThe overall research work is summarized by four sections such as: (i) defining the cross-layer newlineframework and model (ii) categorization of functional modules and (iii) design and implement the newline newlineX newlinemodels using the defined framework and technologies of machine learning and blockchain for newlinespecific and selected QoS parameters. Finally (iv) the results of implementations by ML based newlineprediction model and adoption using blockchain shows the improvement in the QoS. newline