We present Virtual Resources: a software architecture to resolve the tension between effective development and efficient operation of Internet of Things (IoT) applications. Emerging IoT architectures exhibit recurring traits: resource-limited sensors and actuators with RESTful interfaces at one end; full-fledged Cloud-hosted applications at the opposite end. The application logic resides entirely at the latter, creating performance issues such as excessive energy consumption and high latencies. To ameliorate these, Virtual Resources allows developers to push a slice of the application logic to intermediate IoT devices, creating a continuum between physical resources and Cloud-hosted applications. With Virtual Resources, for example, developers can push processing of sensed data to IoT devices close to the physical sensors, reducing the data to transmit and thus saving energy. We describe the key concepts of Virtual Resources and their realization in a CoAP prototype atop resource-constrained devices. Experimental results from cycle-accurate emulation indicate that Virtual Resources enable better performance than Cloud-centric architectures, while retaining the RESTful interaction pattern. For example, energy consumption in representative scenarios improves up to 40% and control loop latencies reduce up to 60%.
Virtual resources for the Internet of Things
MOTTOLA, LUCA
2015-01-01
Abstract
We present Virtual Resources: a software architecture to resolve the tension between effective development and efficient operation of Internet of Things (IoT) applications. Emerging IoT architectures exhibit recurring traits: resource-limited sensors and actuators with RESTful interfaces at one end; full-fledged Cloud-hosted applications at the opposite end. The application logic resides entirely at the latter, creating performance issues such as excessive energy consumption and high latencies. To ameliorate these, Virtual Resources allows developers to push a slice of the application logic to intermediate IoT devices, creating a continuum between physical resources and Cloud-hosted applications. With Virtual Resources, for example, developers can push processing of sensed data to IoT devices close to the physical sensors, reducing the data to transmit and thus saving energy. We describe the key concepts of Virtual Resources and their realization in a CoAP prototype atop resource-constrained devices. Experimental results from cycle-accurate emulation indicate that Virtual Resources enable better performance than Cloud-centric architectures, while retaining the RESTful interaction pattern. For example, energy consumption in representative scenarios improves up to 40% and control loop latencies reduce up to 60%.File | Dimensione | Formato | |
---|---|---|---|
azzara15vres.pdf
accesso aperto
:
Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione
762.67 kB
Formato
Adobe PDF
|
762.67 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.