You are here

Scalability-aware mechanism based on workload prediction in ultra-peer networks


Nabila Chergui, Tahar Kechadi, Salim Chikhi

Publication Type: 
Refereed Original Article
Abstract Ultra-peers networks are emergent architectures in large-scale distributed computing environments. Controlling the workload is essential since the system scales up rapidly and accommodates a dynamic change in the number of users, resources, etc. Thus, developing an advanced model based on scalability aspects is a necessity, since it is an important and critical issue when designing such systems. In this paper, we present a scalability-aware approach for ultra-peers networks where each ultra-peer behaves like an ecosystem, in which, we prevent the presence of bottleneck in these structured information systems, and help each ultra-peer to find and stay in a steady state. We make use of neural networks in conjunction with the queueing theory to understand the behaviour of each ultra-peer, by first estimating its future workload, then to take decision on whether the next period will cause a bottleneck situation or not. After that we propose solutions to allow each ultra-peer to scale with the growth of the network size. The effectiveness of the design on scalability is evaluated using synthetic as well as realistic workloads for a number of different scenarios. Results show that the ultra-peer has successfully supervised its state to scale with the growth of the system size
Digital Object Identifer (DOI): 
Publication Status: 
Date Accepted for Publication: 
Friday, 13 January, 2017
Publication Date: 
Journal of Peer-to-Peer networks and Applications, Springer, 2017, doi 10.1007/s12083-017-0542-z
National University of Ireland, Dublin (UCD)
Open access repository: