Title: Novel decentralized node recovery algorithm in virtualized network environments
Abstract:
Most of the studies in network virtualization focused on solving the
problem of efficient assignment of virtual networks (VNs) in substrate
network (SN). However, none of the existing approaches has considered
the effects of substrate node failure on the operating
VNs. This study focuses on the problem of recovering from
substrate node failure. A novel heuristic algorithm that efficiently
reallocates new resources for the affected VNs after a node failure is
proposed. In this algorithm,
responsible substrate node executes a set of recovery steps to
migrate all the hosted virtual nodes in the failed substrate node in
addition to the virtual paths travel crossing it. The proposed approach
is executed in a distributed manner without any coordination from
the central Infrastructure Provider. Moreover, the proposed approach
efficiently decreases the time delay for recovering the virtual nodes
from any failed substrate node and hence minimizes service
interruption period caused by substrate node failure. This, in turn,
results in increasing the service provider (SP) revenue and decreasing
the paid penalty charges for service violation.