Grapevine: Efficient situational awareness in pervasive computing environments

Abstract

Abstract-Many pervasive computing applications demand expressive situational awareness, which entails an entity learning detailed information about its immediate and surrounding context. Previous work has largely focused on individual entities' context, in this paper we present Grapevine, a framework for efficiently sharing context information in a localized region of a pervasive computing network, using that information to dynamically form groups defined by their shared situations, and assessing the aggregate context of that group. We provide an implementation of Grapevine and benchmark its performance in a live pervasive computing network deployment.

Trust-Based, Privacy-Preserving Context Aggregation and Sharing in Mobile Ubiquitous Computing

“…To compare and contrast our schemes for context sharing subject to privacy constraints, we implemented the schemes in our Grapevine context framework [8]. Grapevine piggybacks context information (whether individual context or aggregate information) on data packets transmitted in the course of other network (application) traffic.…”

“…We measure the percent error of each node's estimate of the global aggregate. The overheads of our approaches are low (see [8] for a presentation of the overhead of piggybacking context information in Grapevine). The approaches all generate the same amount of extra data except for the Mixed Information scheme, which generates twice the number of piggybacked bits.…”

Spitty Bifs are Spiffy Bits: Interest-Based Context Dissemination Using Spatiotemporal Bloom Filters

“…To evaluate the usefulness of SpTBF, we augmented the Grapevine context dissemination framework [9] to use SpTBF to track interest awareness and to improve context information sharing efficiency. Grapevine provides programmers a simple and efficient library for sharing context and collaboratively forming groups based on that shared context.…”

“…The node can then query its SpTBF to determine whether nearby nodes are interested in receiving context information it has available. Sharing the context of interest is outside the purview of the SpTBF structure; it can be handled by a variety of existing techniques [9]. Fig.…”

“…The SpTBF approach can represent generic types of context and provides consumers the flexibility to control the scope in space and time within which their individual interest is distributed. In this paper, we introduce the SpTBF data structure, demonstrate it in practice by applying it to the Grapevine context dissemination framework [9], and evaluate its performance.…”