| New generation information technology promotes the digital transformation of the whole society.The arrival of the computing era urgently required the deep integration of computing power and network,and the Computing Aware Network is a new network technology proposed by this trend,which is designed to make full use of dynamic distribution of computing resources and meet network needs.Computing Aware Network will make a massive calculation service application to be available and optimal moduli resources in real time.This thesis is designed to achieve a prototype system of computing routing for Computing Aware Network.The system can generate computing routing info table according to the full network real-time computing and network status information,and guided various types of computing services forward to the optimal computing server in accordance with the computing routing table that greatly increases the efficiency of integration.The main work and innovation of the thesis includes two aspects.Firstly,this thesis propose a distributed computing routing control scheme,the control plane collects computing resources and service registration message by network edge nodes,and then advertise these information to other nodes,ultimately using CAN Load-Balancing Routing Algorithm to calculate routing and generate CAN routing table.The data plane is routing the service request flow in accordance with the CAN routing table.Secondly,based on the computing routing control scheme design we implement a prototype system of computing routing with open source platform,the system contains multiple routing function modules such as computing and network status information acquisition,computing status advertisement,computing routing calculation,which determine the optimal computing service node according to various calculation service request flows and utilized the IP tunnel technology to complete the routing forward function.The experimental results show that the prototype system of computing routing can ensure good network performance while completing the research content,achieving optimization of power utilization,and improving the purpose of user satisfaction. |
