Abstract
The Internet of Things (IoT) is becoming real, and recent studies highlight that the number of IoT devices will significantly grow in the next decade. Such massive IoT deployments are typically made available to applications as a service by means of IoT platforms, which are aware of the characteristics of the connected IoT devices–usually constrained in terms of computation, storage and energy capabilities–and dispatch application’s service requests to appropriate devices based on their capabilities. In this work, we develop an energy-aware allocation policy that aims at maximizing the lifetime of all the connected IoT devices, whilst guaranteeing that applications’ Quality of Service (QoS) requirements are met. To this aim, we formally define an IoT service allocation problem as a non-linear Generalized Assignment Problem (GAP). We then develop a time-efficient heuristic algorithm to solve the problem, which is shown to find near-optimal solutions by exploiting the availability of equivalent IoT services provided by multiple IoT devices, as expected especially in the case of massive IoT deployments.
Highlights
The Internet of Things (IoT) has rapidly evolved from a cutting-edge research topic to a real ecosystem affecting people’s everyday life
We focus on on aa massive massive deployment of constrained devices managed by an device has sensor deployment of constrained IoT devices managed by an IoT service platform
Since the deadline d j of service invocations of request j mapped to thing i is larger than the period s j p j with which they arrive at the thing, a sufficient condition for the schedulability of the requests is provided by the well-known rate monotonic bound on the thing utilization, defined as: ai vi = ai 2 − 1 where ai is the number of requests allocated to the thing i, i.e.,: ai
Summary
The Internet of Things (IoT) has rapidly evolved from a cutting-edge research topic to a real ecosystem affecting people’s everyday life. Sensors 2019, 19, 693 devices directly into the cloud infrastructure, which is exploited to interconnect different IoT systems in order to provide a unified interface to end users on one side, and to exploit the capabilities of the cloud, in terms of storage and computation, on the other side. Such architectures are the solid ground for the realization of service-brokering solutions by providing a centralized point of control between smart things and applications.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
