Fine-grained intercontact characterization in disruption-tolerant networks

2013 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)

Conan

2013

Self Cite

Abstract-Most disruption-tolerant networking protocols available have focused on mere contact and intercontact characteristics to make forwarding decisions. We propose to relax such a simplistic approach and include multi-hop opportunities by annexing a node's vicinity to its network vision. We investigate how the vicinity of a node evolves through time and whether such information is useful when routing data. By analyzing a modified version of the pure WAIT forwarding strategy, we observe a clear tradeoff between routing performance and cost for monitoring the neighborhood. By observing a vicinity-aware WAIT strategy, we emphasize how the pure WAIT misses interesting end-to-end transmission opportunities through nearby nodes. Our analyses also suggest that limiting a node's neighborhood view to four hops is enough to improve forwarding efficiency while keeping control overhead low.

Section: A the κ-Vicinitymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

The strength of vicinity annexation in opportunistic networking

2013 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)

Conan

2013

Self Cite

“…In our work, we make a clear distinction between nodes that are not in direct contact but still have a path connecting them and nodes that have no possibility of communication (i.e., there is no path between them) [11]. A pair of nodes are in favorable intercontact of parameter n when they are exactly at a shortest n-hop distance.…”

Section: Vicinity Formalizationmentioning

confidence: 99%

Examining vicinity dynamics in opportunistic networks

Proceedings of the 8th ACM Workshop on Performance Monitoring and Measurement of Heterogeneous Wireless and Wired Networks

Campista

et al. 2013

Self Cite

Modeling the dynamics of opportunistic networks generally relies on the dual notion of contacts and intercontacts between nodes. We advocate the use of an extended view in which nodes track their vicinity (within a few hops) and not only their direct neighbors. Contrary to existing approaches in the literature in which contact patterns are derived from the spatial mobility of nodes, we directly address the topological properties avoiding any intermediate steps. To the best of our knowledge, this paper presents the first study to ever focus on vicinity motion. We apply our method to several real-world and synthetic datasets to extract interesting patterns of vicinity. We provide an original workflow and intuitive modeling to understand a node's surroundings. Then, we highlight two main vicinity chains behaviors representing all the datasets we observed. Finally, we identify three main types of movements (birth, death, and sequential). These patterns represent up to 87% of all observed vicinity movements.

“…For intercontact patterns, as previously pleaded by Leguay et al, we base our analysis on the pairwise intercontact definition. We rely on our earlier findings that many pairs of nodes, when not in direct contact, remain nearby (within very few hops) [13]. The extension of contact and intercontact to a vicinity notion brings logical variation in previous intercontact and contact analyses.…”

Section: Positioningmentioning

confidence: 99%

Vicinity-based DTN characterization

Proceedings of the Third ACM International Workshop on Mobile Opportunistic Networks

Conan

2012

Self Cite

We relax the traditional definition of contact and intercontact times by bringing the notion of vicinity into the game. We propose to analyze disruption-tolerant networks (DTN) under the assumption that nodes are in κ-contact when they remain within a few hops from each other and in κ-intercontact otherwise (where κ is the maximum number of hops characterizing the vicinity). We make interesting observations when analyzing several real-world and synthetic mobility traces. We detect a number of unexpected behaviors when analyzing κ-contact distributions; in particular, we observe that in some datasets the average κ-contact time decreases as we increase κ. In fact, we observe that many nodes spend a non-negligible amount of time in each other's vicinity without coming into direct contact. We also show that a small κ (typically between 3 and 4) is sufficient to capture most communication opportunities.