Design And Implementation Of Task-Centric Computing Network Resource Planning System

With the acceleration of the upgrading progress of digital intelligent transformation in various industries,the total amount of social data has an explosive growth.At the same time,the commercialization process is accelerating,and the task needs extensive data transmission and a large number of computing resources to carry out in-depth data analysis and processing based on massive images,videos,physical equipment and network environment.In order to make full use of heterogeneous computing resources in computing network and adapt to increasingly complex functional component structure,task needs to adapt to distributed resources.Componentized tasks represented by augmented reality AR games are formed in computing network.Componentized task consists of different sub-tasks decoupled from various functions and can be deployed to different compute nodes to complete computing tasks.In the computing network,a large number of computing resources distributed on the network can provide more computing nodes close to users to carry computing demands.In this way,task deployment is closer to terminal devices,and data transmission and processing are more timely.However,the computing performance and available resources of heterogeneous resource nodes are significantly different,and the computing resources required by each functional component of a componentized task and the size of the stored data generated during running are also significantly different.The existing computing network resource planning mechanism cannot match the heterogeneous cloud,edge and end computing resources in the computing network environment with the resource requirements of componentized tasks.When the network control plane selects the appropriate computing nodes for the deployment of multiple concurrent componented tasks,subtasks are executed in sequence,resulting in resource conflicts and affecting task quality.Meanwhile,considering that the expansion of network scale will reduce the efficiency of resource planning,and the change of computing capacity of resource nodes will cause the control plane to search for the optimal resource again,which makes the network resources available for task allocation uncertain,so it is necessary to optimize the planning resource scope of concurrent componentized task as much as possible under the condition of meeting these differentiated demands.Therefore,it is of great significance to study the computing network resource planning mechanism for multi-concurrent componentized tasks in the computing network scenario.Firstly,this paper investigates the computing network scenario and its bearer tasks,and studies the research status and existing problems of resource planning algorithm for computing network scenario at home and abroad.On this basis,a method of network resource planning based on reservation strategy for concurrent tasks is proposed.Firstly,a taskcentered algorithm of network resource reservation is proposed to overcome the competition problem in the allocation of network resources for concurrent tasks and to determine the scale of network resources available for task resource allocation.Then,based on the reservation strategy,a method of allocation of computing network resources is proposed,so that the computing network can quickly provide tasks flexibly and dynamically according to the specific task scenarios in reality.The experiment shows that resource allocation within the scope of resource reservation can save about 60.58%to 71.43%of the time,and improve the efficiency of computing network resource planning.On the basis of algorithm research,this paper designs and implements a task-centered computing network resource planning system.The system is designed and developed in accordance with the modular way,which can provide the implementation scheme of integrated computing network planning,provide the task-related computing network resource information concerned by the control plane.In addition,it can visually realize resource monitoring of integrated computing power network,and carry out statistical analysis on task,network and algorithm data,providing a basic platform for the subsequent research on computing power network.In summary,this paper focuses on the study of the computing network resource planning method of multiple concurrent tasks in the computing network scenario,solves the conflicts existing in the process of task competition for computing network resources,improves the efficiency of computing network resource planning,and then designs and implements the corresponding resource planning system based on this,and finally verifies through the experimental simulation and system implementation.