Dynamic Allocation of Service Function Chains under Priority Dependency Constraint

Abstract

Network functions virtualization is a new technology for the future internet that eliminates the dependency of the network function and the hardware requirement. The network functions virtualization provides a successful approach for meeting the increase in demand of the end-to-end (E2E) services with low operational and capital costs. Replacing the network specific purpose hardware (e.g. firewall) with a software implementation of the network functions in which a chain of Virtualized Network Functions (VNFs) can logically connect the end points and provide the desired network services. However, this approach is associated with the challenge of dynamically mapping the predefined VNFs onto the existing substrate network in an optimal way. In this paper, we propose a simple and effective approach for mapping the VNFs with the physical resources in a dynamic service request environment. The algorithm considers the priority dependency between the VNFs as a case of study, with the objective of minimizing the mapping blocking rate.