Adaptation strategies for self-management of tree overlay networks

Abstract

Self-management of tree overlay networks for distributed applications is the challenge this paper addresses. Eight local adaptation strategies are introduced based on which autonomous self-organized agents establish connections that build and maintain a tree topology. Quantitative and qualitative experimental evaluation illustrates and compares the effects of adaptation strategies in the resulting tree topologies according to a defined self-organization goal and four metrics: connectedness, connectivity, instability and robustness. This paper concludes that further applicability of adaptation strategies in other self-organization goals and topologies is promising.

A generic and adaptive aggregation service for large-scale decentralized networks

“…User ranking aggregation in recommender systems (Garcin et al 2009), is based on a finite and often restricted number of options for a user to rank an element. In applications of demand-side energy management (James et al 2006;Pournaras et al 2010), aggregate information about a finite number of alternative demand options improve the stability of the Smart Power Grid.…”

“…Robust tree overlays are a flexible methodology to compute a wide range of aggregates but require topology self-management (Pournaras et al 2010) in decentralized environments. Communication and storage complexity can be higher than the aggregation itself.…”

Measuring and controlling unfairness in decentralized planning of energy demand

“…EPOS coordinates demand by letting agents interact over self-organized tree topologies [15]. For the same tree topology, 10 instances are generated, in each of which the agents are positioned 9 Unfairness is also minimized as shown in this paper.…”

Decentralized Planning of Energy Demand for the Management of Robustness and Discomfort

“…EPOS decreases computational complexity to O(l c ), where c is the number of children per agent for a c-ary tree. Fault tolerance can be provided with self-organization mechanisms such as AETOS [12] that builds and maintains reconfigurable tree topologies in dynamic distributed environments.…”

“…Each coordination phase of EPOS runs for 10 different 3-ary tree topologies. Each topology is built by the AETOS overlay service [12]. AETOS self-organizes agents in different random positions for each tree topology to capture the effect of topological positioning.…”