| Summary: | This paper focuses on the problem of interoperability and resource sharing in wireless sensor networks (WSNs) running under the Internet of Things (IoT) and cyber-physical systems (CPSs). Considering the scale of IoT/CPS projects, conventional WSN virtualization techniques remain incapable because of the hardware/software constraints and heterogeneity. To this end, in this paper, an agent-based server system approach, which improves the resource sharing between heterogeneous WSNs in IoT/CPS providers, is proposed. In line with this approach, a software agent framework is introduced. With the help of the framework, called Firat Virtual WSN framework (FVWSN), the clients can move the commands/queries and data fusion/aggregation algorithms, which use on their local networks, to the provider side and run them remotely or automatically. This process is carried out by logical agent entities, called virtual nodes, which are created with the help of FVWSN. In this way, since the client evaluation mechanism is performed at a closer point to the shared resources, a shorter response time can be achieved in time-critical applications. The most important features that differentiate the developed agent framework from other agent-based technologies are that it is semi-autonomous and uses a specific resource selection/allocation algorithm. With the improvement that FVWSN provides for IoT/CPS WSN providers, it is possible to achieve a shorter response time and allow more client applications to share the same limited WSN resources. In this paper, first, the analysis and the necessity of the proposed system are discussed. Then, the system is simulated in the OPNET Modeler platform to make comparisons with well-known conventional WSN resource sharing mechanisms. Finally, the physical comparison tests of the system are carried out on an OpenStack-based cloud system, and the success of the system is shown.
|
|---|