Research And Implementation Of State Management Mechanism For Computing Force Networking

In recent years,with the continuous update and development of the Internet technology,more and more emerging services have emerged,which lead to the numbers of devices in the network grow exponentially.Existing traditional networks tend to establish a centralized computing power resource pool to support the computing needs of the business.But in the emerging application scenarios,most services have low latency,high bandwidth,and requirements for lightweight and dynamic computing.The traditional method cannot fully meet these requirements,which will lead to network congestion and the user experience satisfaction will drop.Therefore,it is necessary to select suitable computing service nodes according to business characteristics,network costs and network conditions to solve the problem of insufficient devices computing power resources.The integration of network and computing power has become a new trend,and the Computing Force Networking emerge as the times require.Based on the distributed Computing Force Networking,this paper studies the perception method and implementation scheme of compute state information and network state information.The main work and the research innovations of this paper are as follows:(1)Based on the distributed Computing Force Networking architecture,we have designed and implemented the perception and notification mechanism of computing power status.The computing power status of the computing service node describes the ability of the current service node to provide services for user requests.Obtaining the computing power status information of the computing service node can help to select the more appropriate node to provide services for the users.The perception of computing power status includes the computing power registration process and the notification process of computing power status information.We have added computing power perception function in the edge management system to update and manage computing power status information.Based on the extended BGP protocol,the whole network can realize the notification of the computing power status information.(2)Design and implement the network status perception mechanism.In the Computing Force Networking,the computing power routing entry node sends network status detection packets regularly to each computing power routing exit nodes to obtain the network status information.After obtaining the network status information on each path,the information is normalized and updated to the network status table.The edge management system has the function of detecting network status,and the network status information is managed uniformly in the edge management system.The computing power routing entry node can obtain network status information on a specific path as needed.(3)Design and build a computing power routing simulation topology prototype.Design the front-end display interface.The distributed Computing Force Networking can perceive the compute status information and network status information in the network.The computing power routing node announces the compute status information to the nearest network nodes,and spreads to the whole network.In order to simulate the real network as much as possible,the computing power routing simulation topology is designed based on the distributed Computing Force Networking architecture.Based on the ubuntu system,we use the lightweight virtualization platform docker to build it.Also,we realized the automatic configuration of the simulation topology.After the connectivity test of the constructed computing power routing simulation topology,the effectiveness of the computing power status and network status awareness mechanism is verified.In order to display the entire Computing Force Networking simulation topology as well as computing power status data and network status data more clearly and intuitively,we designed and implemented a front-end display interface.It includes the Computing Force Networking topology view and the performance data view.It also presents abstract simulation network topology and data.