Task requirement aware pre-processing and Scheduling for IoT sensory environments

In Internet of Things (IoT) sensory environment Wireless Sensor Networks (WSNs) are connected to the Internet through gateways and this gives birth to many real time sensor based applications. Applications from Internet querying for Spatio-temporal information within WSN may require query decomposition. Decomposition of queries may result in tasks having similar functional and Quality of Service (QoS) requirements. Thus pre-processing the tasks results in decreased number of task executions within WSN as compared to executing tasks individually. Moreover prior knowledge of sensor nodes' residual energy can help in deciding if a task can be scheduled for its successful completion. Existing energy monitoring protocols in WSNs incur considerable amount of message traffic and energy. Thus an Energy Monitoring System (EMS) that works on reduced message traffic and yet provide the correct energy information will be of great value. This paper proposes Task Requirements Aware Pre-processing and Scheduling (TRAPS) mechanism comprising a task pre-processor, EMS and scheduler within the gateway. The scheduler allocates the best available sensor nodes to the incoming tasks while meeting their QoS requirement. Task pre-processor and EMS are tested individually to measure the key performance metrics that include number of tasks entering WSN and energy information prediction accuracy respectively. Further the performance evaluation of TRAPS is extended to study its effects on various network parameters. It is found that TRAPS outperforms the findings of recent research on task scheduling in IoT sensory environments in terms of the identified network performance metrics.