Print Email Facebook Twitter Extreme Chaos: Flexible and Efficient All-to-All Data Aggregation for Wireless Sensor Networks Title Extreme Chaos: Flexible and Efficient All-to-All Data Aggregation for Wireless Sensor Networks Author Chronopoulos, D. Contributor Cattani, M. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department Software Technology Programme Embedded Systems Date 2016-02-16 Abstract Wireless Sensor Networks (WSNs) are networks of sensor devices that perform environmental monitoring applications. Data aggregation is a vital building block for WSNs that enables rapid gather the information held by the network nodes. A continuous study of different approaches to retrieving aggregates is being carried out by the researchers of the field since the emergence of WSNs. This study has resulted in a plethora of aggregation protocols, each with their own advantages and flaws. Recent efforts have given us Chaos, a very powerful tool able to perform network-wide data aggregation. In contrast to its predecessors Chaos performs aggregation with a topology-agnostic communication primitive, embracing pure wireless broadcasts to create extremely fast and energy-efficient aggregation rounds. Although Chaos outperforms other solutions in terms of time and energy efficiency, it also has significant shortcomings. Specifically, in order to operate properly, Chaos imposes strict restrictions on the network composition, which prevent it from adapting to the dynamic nature of WSNs. Additionally, Chaos needs to reserve a portion of space in every transmitted packet whose size is proportional to that of the network; making its scalability questionable on larger networks. In this work we identified the above shortcomings along with their causes, which reside in the core mechanisms Chaos employs, and in the natural phenomena that govern its behaviour. To overcome these shortcomings, we propose a new coordinating data structure based on order-statistics theory, and a flow control technique to avoid congestion. Re-building the core mechanisms of the protocol around these proposals lead to a topology-agnostic primitive that is not only highly efficient (like Chaos), but also more able to cope with the practical challenges and needs of sensor network applications. Our evaluation, performed on two different testbeds, shows that our protocol (i) is more scalable; (ii) is highly tolerant to network dynamics; (iii) allows the retrieval of more types of aggregates; and (iv) reduces the aggregation latency (by up to 15%) while attaining a comparable reliability. Subject WSNData-AggregationNeighbor-DiscoveryChaosSynchronousFlooding To reference this document use: http://resolver.tudelft.nl/uuid:7e7b52b0-cf55-4d29-8809-b6d6357bfe0d Part of collection Student theses Document type master thesis Rights (c) 2016 Chronopoulos, D. Files PDF Extreme_Chaos_DChronopoulos.pdf 765.76 KB Close viewer /islandora/object/uuid:7e7b52b0-cf55-4d29-8809-b6d6357bfe0d/datastream/OBJ/view