DIAS, the Dynamic Intelligent Aggregation Service is a decentralized middleware mechanism that makes locally available in every node of a network system-wide (global) information without involving centralized computational entities. More specifically, DIAS locally computes almost any aggregation function that receives for input numerical values from all of the nodes in a network. These values represent a property of the nodes.
[…] a decentralized mechanism that makes locally available in every node of a network system-wide (global) information without involving centralized computational entities […] e.g. AVERAGE, SUMMATION, MAXIMUM, MINIMUM, COUNT, etc […]
DIAS achieves this abstraction and flexibility by introducing the concept of aggregation memberships. An aggregation membership provides historic information in each node about a computed input value. This information indicates if the aggregated value is new, outdated or duplicate. This distinction guarantees accurate computation of aggregates. It also provides two adaptation strategies that tune performance for a fast update of outdated values or a fast discovery of new input values.
Storing aggregation memberships explicitly is not a scalable and decentralized aggregation approach. Nevertheless, DIAS stores aggregation memberships in probabilistic data structures: the bloom filters. A bloom filter provides large space savings at the cost of false positives. A false positive incorrectly denotes that a membership exists when it actually does not. DIAS is able to detect false positives and, therefore, prevent inaccuracies related to duplicate and outdated information. Experimental evaluation shows the performance trade-offs of DIAS and confirms its high accuracy under different experimental settings.
DIAS is part of my PhD studies and resulted in a journal publication.
- Evangelos Pournaras, Jovan Nikolic, Self-corrective Dynamic Networks via Decentralized Reverse Computations, in the Proceedings of the 14th International Conference on Autonomic Computing-ICAC-2017, Columbus, Ohio, July 2017
- Evangelos Pournaras, Jovan Nikolic, Alex Omerzel, Dirk Helbing, Engineering Democratization in Internet of Things Data Analytics, in the proceedings of the 31st IEEE International Conference on Advanced Information Networking and Applications-AINA-2017, Taiwan, March 2017 © IEEE
- Evangelos Pournaras, Martijn Warnier and Frances M.T. Brazier, Peer-to-peer Aggregation for Dynamic Adjustments in Power Demand, Peer-to-Peer Networking and Applications, Vol. 8, Nr. 2, pp. 189-202, 2015 © Springer
- Evangelos Pournaras, Martijn Warnier and Frances M.T. Brazier, A Generic and Adaptive Aggregation Service for Large-scale Decentralized Networks, Complex Adaptive Systems Modeling, 1:19, 2013 © SpringerOpen
- Evangelos Pournaras, Multi-level Reconfigurable Self-organization in Overlay Services, PhD Thesis, Delft University of Technology, March 2013 (Chapter 5)