Thomas Clouser

Tiara: A Self-stabilizing Deterministic Skip List

We present Tiara -- a self-stabilizing peer-to-peer network maintenance algorithm. Tiara is truly deterministic which allows it to achieve exact performance bounds. Tiara allows logarithmic searches and topology updates. It is based on a novel sparse 0-1 skip list . We rigorously prove the algorithm correct in the shared register model. We then describe its extension to a ring and incorporation of crash tolerance.

Tiara: A self-stabilizing deterministic skip list and skip graph

We present Tiara-a self-stabilizing peer-to-peer network maintenance algorithm. Tiara is truly deterministic which allows it to achieve exact performance bounds. Tiara allows logarithmic searches and topology updates. It is based on a novel sparse 0-1 skip list. We then describe its extension to a ringed structure and to a skip-graph.

Concurrent face traversal for efficient geometric routing

We present a concurrent face routing CFR algorithm. We formally prove that the worst case latency of our algorithm is asymptotically optimal. Our simulation results demonstrate that, on average, the path stretch, i.e., the speed of message delivery, achieved by CFR is significantly better than by other known geometric routing algorithms. In fact, it approaches the shortest possible path. CFR maintains its advantage over the other algorithms in pure form as well as in combination with greedy routing. CFR displays this performance superiority both on planar and non-planar graphs.

Fast Geometric Routing with Concurrent Face Traversal

We present a concurrent face routing CFR algorithm. We formally prove that the worst case latency of our algorithm is asymptotically optimal. Our simulation results demonstrate that, on average, CFR significantly outperforms the best known geometric routing algorithms in the path stretch: the speed of message delivery. Its performance approaches the shortest possible path. CFR maintains its advantage over the other algorithms in pure form as well as in combination with greedy routing; on planar as well as on non-planar graphs.

Emuli: model driven sensor stimuli for experimentation

We describe Emuli - a method of replacing sensor data with a network-wide model of stimuli events. Sensor readings are generated on demand from the modeling data stored at each device. This approach allows for both repeatable and variable experimentation with a network of physical devices for existing and planned sensing modalities. We illustrate the approach with (i) a light sensor and (ii) a hypothetical range sensor used in a tracking application.

DRIFT: Efficient Message Ordering in Ad Hoc Networks Using Virtual Flooding

We present DRIFT - a total order multicast algorithm for ad hoc networks with mobile or static nodes. Due to the ad hoc nature of the network, DRIFT uses flooding for message propagation. The key idea of DRIFT is virtual flooding - a way of using unrelated message streams to propagate message causality information in order to accelerate message delivery. We describe DRIFT in detail. We evaluate its performance in a simulator and in a wireless sensor network. In both cases our results demonstrate that the performance of DRIFT exceeds that of the simple total order multicast algorithm designed for wired networks, on which it is based. In simulation at scale, for certain experiment settings, DRIFT achieved speedup of several orders of magnitude

Void traversal for efficient non-planar geometric routing

Geometric routing provides a scalable and efficient way to route messages in ad hoc networks if extensive routing information is unavailable. Such algorithms require a planar graph to guarantee message delivery. The routing techniques for such guarantee usually center around the traversal of planar faces of the graph. However, in realistic wireless networks existing planarization methods, if at all applicable, tend to require extensive local storage or result in suboptimal route selection. In this paper we study an alternative approach of translating the algorithms themselves to be able to route messages over voids in non-planar graphs. We prove sufficient memory requirements for such translations. We then translate several well-known planar geometric routing algorithms and evaluate their performance in both static and mobile networks.