Research On Energy Efficient And Path Splitted Virtual Network Embedding Algorithms

In recent years,cloud computing related technologies have been widel y used,and the diversity requirements of network services have become increasingly strong.The network function virtualization technology allows specific functions to be implemented on virt ual machines of physical machines to implement diverse services.The virtual network embedding algorithm is one of the ke y issues in.NFV.The embedding algorithms need to determine the deployment location of the virtual network according to the requiremen ts of the virtual network and the resources of the ph ysical network.Allocating an appropriate amount of resources to the service requests includes providing physical nodes to run the network functions and providing mapping paths to deploy the virtual link s.It is a common goal for embedding algorithms hat how to effectivel y deploy virtual networks to maximize the revenues.As we know,activating and running a substrate node come at a certain energy cost,and deploying a service chain requires several open substrate nodes which lead to high energy consumption.Furthermore,the deplo yment of service chains also need to consume bandwidth resource on a substrate network.After embedment of some service chains,the substrate network may consist of man y links with fragmented remaining resources that are too little to be utilized by any other requests.This leads to a situation whereb y the nodes resources can not be fully utilized and are expended quickly,which causes the low resources utilization.Since the resou rces of the physical network are limited,due to their faster consumption,the acceptance ratio remains low.In order to solve the above problems of high energy consumption,low resource utilization and low request acceptance ratio,this paper has carried out the following two aspects:This paper proposes a path splitted and energy-aware virtual network algorithm solved b y a mixed integer linear programming model.The algorihm considers decreasing the energy consumption by reducing the number of open nodes and mapping the service chains to splitted paths to make most of network resources.We propose to divide a service chain into several segments called service chain(SC)segments to maximize the reuse of open nodes for one service chain.Each SC segment ma y be a virtual edge.And we construct an augmented network b y appending these SC segments to substrate network.This ensures that finding a path with nodes and links mapping is equivalent to path selection on augmented network,therefore simplifying dealing with two mappings simultaneously.Then we formulate the embedding service chains on augmented network as a Mixed Integer Linear Program(MILP)with binar y constraints on the nodes and continuous constraints on the flows.We minimize the number of open nod es to save energy and set the flow variable as continuous so that the embedded path is splitted to utilize the fragmented resource on the object function of the program.In order to reduce the difficult y of solving mixed integer programming problems,this paper designs a Service Chain Constrained Min-Cost Flow Algorithm(SC-MCF).We construct a network flow model based on an augmented network to enable the augmented network topology to support the operation of max-cut which is fit to splitted paths in networ k flow model.Firstly,we design the algorithm by using two bidirectional directed edges on network model to represent an undirected edge on augmented network,the bandwidth capacity of each edge on augmented network act as the capacity of each edge on net work flow model.Then we convert a node with on or off state to two nodes in which one presents the come-in node and the other presents the come-out node,and one directed edge from the come-in node to the come-out node with cost weight.We set higher cost weight for closed nodes and lower cost weight for open nodes.Then we propose a SC-MCF Algorithm using cost weight representing the state of nodes to complete the embedment on an augmented path.We find a splitted path with the minimal cost weight to minimize the number of open nodes.In this algorithm,we adopt the Edmonds-Karp algorithm which includes cycle canceling and path augmenting.Finally,we conduct extensive experiments based on two real topologies EasyNet and GrNet.The experiments show that ou r proposed SC-MCF Algorithm decreases the number of open nodes and increases the acceptance ratio while saving the resources in the long run.