Poster at 7th ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec),
July 2014. .
Jiaqi Tan, Utsav Drolia, Rajeev Gandhi, Priya Narasimhan
Carnegie Mellon University
Pittsburgh, PA 15213
Mobile personal devices such as smartphones and tablets are ubiquitous today, and they are growing in storage, compute, and sensing capabilities. Collectively, these mobile devices in close physical proximity present a rich pool of both compute/storage resources and personal data. Mobile edge-clouds are clouds comprised entirely of mobile nodes in close physical proximity without any infrastructure support such as back-end compute clouds [1]. Mobile nodes serve as both the compute nodes, and the source of data for mobile edge-clouds. Mobile edge-clouds allow the compute/storage resources and data stored across multiple mobile devices to be pooled to form a single compute resource, and they enable applications across independent mobile devices, particularly when high-bandwidth, low-latency connections to the Inter- net may be degraded (e.g. in massive crowds in stadiums), or unavailable (e.g. during disaster response). However, a key security risk which may prevent users from participat- ing in mobile edge-clouds is that their mobile devices need to execute code from other untrusted edge-cloud nodes [5]. Hence, we propose a system which allows nodes in a mobile edge-cloud to securely execute code from untrusted clients.
KEYWORDS: Mobile Edge-Clouds, Safety Properties, Theorem Proving
FULL PAPER: pdf