IoT Resource Access and Management

Introduction

An IoT application seamlessly integrates the virtual world of information with the real world (Uckelmann et al., 2011). However, most existing applications take the form of vertical “silos” (Zorzi et al., 2010), a closely-coupled application pattern lacking effective mechanisms that can support applications in sharing and reusing resources, and interacting with each other. In a vertical silo solution, IoT engineers must bridge the gap between the higher-level application and the underlying technical details manually, and need to be experts in both domains (Bimschas et al., 2011). This is problematic, since application developers are primarily interested in real-world entities (things, places, and people) and their high-level states, rather than in the details of sensors and actuators (Pfisterer et al., 2011). Meanwhile, a large number of existing legacy sensor resources can offer an important contribution to IoT, but are typically locked into closed systems; unlocking these valuable sensor data from closed systems has the potential to revolutionise how we live. To realise this potential, an infrastructure is needed that can connect sensors to the Internet and publish their output in well-understood, machine-readable formats on the Web, thus making them accessible and usable on a large scale (Pfisterer et al., 2011).

Although the concept of IoT was proposed by the MIT Auto-ID Centre in 1999, the relevant theories and IoT applications still take the form of silos. Huge amounts of resources and information are provided by the IoT, and there is a need to build an IoT resource platform that can open and share IoT resources. Several issues are involved in this:

• •

  Resource Access: Dynamic resource access, i.e. realising plug-and-play (PnP) functionality of heterogeneous resources, in which the access system can seamlessly recognise a new resource, and automatically interpret and process the information generated by the resources;

• •

  Resource Management: Resource management and maintenance are highly complicated, due to the heterogeneity, instability and evolution of the IoT;

• •

  Resource Discovery: A very large number of resources provides more capabilities, but also creates issues in terms of resource discovery. The issue of how to find the most appropriate resource across different domains and networks within an acceptable range of time and space is as yet unsolved;

• •

  Resource Utilisation: The real world can be reflected in an increasingly realistic way by the digital world, through an expansion of its range and increased sensing capabilities. Meanwhile, the operations and tasks generated in the digital world are becoming more complicated, and ways in which IoT services can be designed to fully utilise IoT resources should be explored.

Research institutes such as EU FP7, CSIRO have carried out research into resource management architectures for IoT, such as SENSEI, PECES (Villalonga et al., 2010; Haroon et al., 2009), etc. This chapter presents the resource management characteristics involving the access, discovery and management of resources (Haroon et al., 2009; Barnaghi et al., 2011; Han & Crespi, 2017); a resource platform is then implemented based on these resource management characteristics. The resource platform is implemented by the following roadmap:

• •

  It is based on SOA, integrates service component architecture (SCA) and uses the enterprise service bus (ESB). SCA serves as the assembly standard for the service components, while ESB provides protocol translation, message routing, security and management functions for the interactions between components. Hot deployment and cross-platform capabilities are provided by the Open Service Gateway Initiative (OSGi) (OSGI, 2008). These features achieve decoupling between the components and the distributed framework.

• •

  Ontology-based resource models (such as W3C SSN (SSN, 2014) and OWL-S, which are extended to express domain knowledge and business information) act as information bases. The IoT resource platform interprets raw data and automatically promotes observation data to semantic domain information, based on the models.