A Holistic Architecture for the Internet of Things, Sensing Services and Big Data David Tracey, Cormac Sreenan Dept. Of Computer Science, University College Cork, Cork, Ireland Abstract—Wireless Sensor Networks (WSNs) increasingly enable applications and services to interact with the physical world. Such services may be located across the Internet from the sensing network. Cloud services and big data approaches may be used to store and analyse this data to improve scalability and availability, which will be required for the billions of devices envisaged in the Internet of Things (IoT). The potential of WSNs is limited by the relatively low number deployed and the difficulties imposed by their heterogeneous nature and limited (or proprietary) development environments and interfaces. This paper proposes a set of requirements for achieving a pervasive, integrated information system of WSNs and associated services. It also presents an architecture which is termed holistic as it considers the flow of the data from sensors through to services. The architecture provides a set of abstractions for the different types of sensors and services. It has been designed for implementation on a resource constrained node and to be extensible to server environments. This paper presents a ‘C’ implementation of the core architecture, including services on Linux and Contiki (using the Constrained Application Protocol (CoAP)) and a Linux service to integrate with the Hadoop HBase datastore. Index Terms—Wireless Sensor Networks, Tuple Space, Information Model, Protocols, Cloud Computing, Big Data I. INTRODUCTION Wireless Sensor Networks (WSNs) are being enabled by the increasing availability of sensors and advances in wireless technologies, hardware and the use of IP for connecting resource constrained devices. The use of micro IP stacks (and IPv6 over Low power Wireless Personal Access Networks (6LowPAN) [1] has enabled constrained devices to connect to the Internet in a so called “Internet of Things” (IoT). Definitions of IoT generally share the idea that it relates to the integration of the physical world with the virtual world of the Internet [2]. IoT is characterised by an interconnected set of individually addressed and constrained (possibly autonomous) devices in a distributed system, with sensing/active devices for physical phenomena, data collection, and applications using sensing, computation and actuation. There could potentially be billions of such devices connected across the Internet with predictions of 50 to 100 billion devices being connected to the Internet by 2020 [3]. WSNs have a (possibly large) number of devices with sensing capabilities, limited processing capability and wireless connectivity (allowing nodes to be deployed close to the phenomenon being observed) to other sensor or gateway nodes. WSN nodes exist to sense a particular entity, collect (and possibly parse or aggregate) the data and send the data to one or more destinations and ultimately to an application across a range of areas, e.g. environmental monitoring, surveillance and healthcare. Such deployments are usually dedicated and proprietary or specialized to optimise one particular aspect such as lifetime. The availability of increased storage and processing power at a lower cost with greater bandwidth has enabled a range of Cloud Computing services. In terms of IoT, this allows more sources of data to be collected and for the data to be held for a longer time and to be processed by powerful cloud based applications and Big Data techniques, e.g. HBase and MapReduce. Big Data can be characterised by the 3 ‘Vs of Volume (size of the data), Variety (range in type and source of data) and Velocity (frequency of data generation) [4]. The constrained nature of WSN nodes in terms of processing power, memory and energy consumption makes it difficult to enable WSNs to be more easily deployed, developed and integrated with new Internet based services. A key challenge is to enable WSNs to become extensions of the Internet infrastructure, to take full advantage of Cloud and Big Data services [5] and be universally available, rather than isolated and relatively small islands of sensor networks. To address this challenge, this paper presents a set of architectural requirements, a resulting layered architecture and abstractions for the data exchange roles taken by services on WSN nodes and in the Cloud, supported by a novel protocol. It also evaluates an initial implementation of the architecture. The remainder of this paper is organised as follows. We discuss prior work in section II and present a set of architectural requirements to meet the challenge above in Section III. Section IV presents the architecture, including its service abstractions, object library and introduces the message protocol. Sections V and VI present an initial implementation and evaluation of the architecture and its HBase integration. The paper concludes in Section VI. II. EXISTING AND EMERGING FRAMEWORKS This section outlines the current frameworks and approaches used in the Internet of Things, WSN software, Cloud Integration and Big Data. A recent survey shows that only 13 of 28 WSN systems surveyed have actually been implemented on hardware rather than run in simulators [6] and that there is still an absence of broad abstractions, which we propose later. Hence applications are often bound to a particular WSN technology and not easily portable as the 2013 13th IEEE/ACM International Symposium on Cluster, Cloud, and Grid Computing 978-0-7695-4996-5/13 $26.00 © 2013 IEEE DOI 10.1109/CCGrid.2013.100 546
A Holistic Architecture for the Internet of Things, Sensing Services and Big Data
Apr 25, 2023
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Related Documents
