Towards Efficient Processing of General-Purpose Joins in Sensor Networks

Abstract

Abstract-Join processing in wireless sensor networks is difficult: As the tuples can be arbitrarily distributed within the network, matching pairs of tuples is communication intensive and costly in terms of energy. Current solutions only work well with specific placements of the nodes and/or make restrictive assumptions. In this paper, we present SENS-Join, an efficient general-purpose join method for sensor networks. To obtain efficiency, SENS-Join does not ship tuples that do not join, based on a filtering step. Our main contribution is the design of this filtering step which is highly efficient in order not to exhaust the potential savings. We demonstrate the performance of SENS-Join experimentally: The overall energy consumption can be reduced by more than 80%, as compared to the state-of-the-art approach. The per node energy consumption of the most loaded nodes can be reduced by more than an order of magnitude.

Efficient Distributed Query Processing

“…When a request is received, a node can probabilistically decide whether the sub-tree contains node(s) that satisfy a given static attribute, and decide whether to forward the request or not. Stern et al (2009) rely on building one tree [7] while Mihaylov et al (2008) build three trees rooted in different parts of the network in order to speed-up the search and to find shorter paths between nodes [8]. The problem of using summaries is that the search is probabilistic, hence confirmation from the destination node(s) is required.…”

“…Stern et al (2009) propose a twophase approach, where summaries are firstly collected from the whole network, then, at the base-station, candidates fitting the query are chosen. In the second phase, only data from chosen candidates are retrieved [7]. The last approach uses pairwise joins, which splits the processing into pairs, and for each pair of sources it finds a node on the path between them that processes data [14].…”

“…All static attributes in a network can be represented as a table, wherein each column represents an attribute and each row represents a node. In current systems, strategies that a node may follow in order to find another node with a given static attribute are: i) flood the whole network with a search query and wait for a response from all nodes [5], [6], ii) use a summary of all static attributes which can probabilistically say whether a given node, or part of the network, has or does not have the specific attribute [7], [8], iii) store information on a set of predefined node [9], [10].…”

Dragon: Data discovery and collection architecture for distributed IoT

“…When a request is received, a node can probabilistically decide whether the subtree contains node(s) that satisfy a given static attribute and decide whether to forward the request or not. Stern et al rely on building one tree [6] while Mihaylov et al build three trees rooted in different parts of the network [7] in order to speed-up the search and to find shorter paths between nodes. The problem of using summaries is that the search is probabilistic, hence the confirmation from the destination nodes is required.…”

“…Stern et al propose a two-phase approach where first summaries are collected from the whole network, then at the base-station candidates fitting the query are chosen. In the second phase only data from chosen candidates are retrieved [6]. The last approach uses pairwise joins which splits the processing into pairs and for each pair of sources it finds a node on the path between them which processes data [10].…”