Qatar project to leverage IoT
Connected devices at home can be tapped for their spare power
The Carnegie Mellon University in Qatar (CMU-Q) recently attained a major milestone when Sannan Tariq, a senior in the Computer Science program at CMU-Q, presented a paper in Paris on computational offloading at the International conference, “Cloudification of the Internet of Things”.
Tariq and CMU-Q alumna Hend Gedawy are the two main authors, supervised by Abderrahmen Mtibaa and Khaled Harras, who created the first testbed for flexible edge computational offloading across the Internet of Things (IoT), dubbed Cumulus.
While other researchers have created testbeds using a pool of the same devices, the CMU-Q team used a heterogeneous pool of devices including IoT sensors, mobile devices and laptops, in order to increase flexibility and real-world applicability.
With the proliferation of devices like smartphones, tablets, laptops and game consoles, as well as Internet of Things (IoT) devices like sensors, smart watches and smart appliances, there are massive computational resources that lie dormant in any given home at any given time.
Researchers worldwide are studying how to leverage these idle computing devices to address the large resource requirements of mobile applications that might transcend the capabilities of a single device. While traditional cloud computing offers virtually unlimited computing capacity through data centers, there are drawbacks to sending data to a remote location, including security and privacy concerns. Offloading these tasks to nearby devices is a viable option that could transform the way we compute.
As researchers turn to computational offloading, they will require a stable environment for testing. Cumulus is a flexible, distributed platform for offloading computation. The Cumulus prototype consists of three main components: an experimentation testbed consisting of a task dispatcher as the central entity, a pool of IoT devices that includes Android smartphones, tablets, Raspberry Pis and Intel Edisons, and a controller which manages the cloud of devices.
Cumulus implements a hierarchical architecture allowing worker devices to function either as standalone devices, or as part of IoT clouds that interact with the system via a controller. The heterogeneity of the platform was tested by using different operating systems, device types, device compute and communication capabilities.
The team believes that using a heterogeneous pool of testbed devices gives users more flexibility to carry out experiments in diverse environments.