Research And Implementation Of Remote Sensing Information Acceleration Extraction System Based On Microservice And OpenAcc

With the advent of the era of big data,the functions of software systems have gradually diversified.The calculation models of early remote sensing information acceleration extraction systems are mostly single-node parallel,and later developed to use cluster parallel processing architecture for design and development.The current remote sensing information acceleration extraction system faces challenges of limited reusability and scalability,and the convenience of the service,the scalability of the system,and the maintenance of the system are insufficient.The microservice architecture has independent deployment capabilities and can achieve process independence.The microservice architecture is highly scalable,highly available,and has a more intelligent operation and maintenance method.Each service in microservices is cohesive enough to effectively manage each service.Businesses can realize real-time updates and modular release,and can provide services without interruption.Therefore,how to make the remote sensing information acceleration extraction system have stronger service convenience and scalability,and realize the reuse of resources,so as to further improve the development efficiency of the system is the key to the overall efficiency of the system.With the sharp increase in the resolution of remote sensing images and the continuous expansion of data volume,remote sensing-related practitioners are increasingly demanding real-time processing of remote sensing images.CUDA-based remote sensing thematic product algorithm parallel mode has mostly achieved good acceleration performance,but CUDA has the problems of long programming cycle and difficult to master.Using Open Acc parallel can shorten the development cycle,without rewriting the serial algorithm,and can compile and run multiple devices at once.Therefore,how to achieve more efficient data and task parallelization of the algorithm and achieve shorter cycle development,thereby further improving the execution efficiency of the algorithm and the real-time nature of the processing process to meet the needs of remote sensing applications in various industries Key.The main work carried out in this article are:(1)Based on the GPU computing resources in the computing node,two parallel processing modes of Open Acc based on asynchronous multi-stream are proposed.By dividing the calculation data into blocks and adopting asynchronous multi-stream to accelerate the calculation,it effectively shields part of the data transmission and calculation time,thereby improving the parallel processing performance of the algorithm.Taking the normalized vegetation index as an example,the detailed design process of the parallel processing strategy is demonstrated,and the effectiveness of the parallel processing model of Open Acc based on asynchronous multi-flow is verified through numerical experiments.(2)Adopt the design concept of distributed and multi-node collaboration,integrate declarative remote calling component Feign,service gateway Zuul,security authentication component Spring Cloud Security and service fuse Hystrix and other technologies to build a complete microservice framework.Designed and implemented a remote sensing information acceleration extraction system based on microservices and Open Acc,and integrated the core acceleration and remote sensing data display services in the system into the framework.The system is divided into service modules such as remote sensing express,data acceleration processing and remote sensing data display.The Open Acc parallel processing mode based on asynchronous multi-stream is integrated into the data acceleration processing module in the system,and the acceleration effect is visually displayed in the form of a visual chart through the remote sensing data display module.By comparing the experimental results,it is verified that the two parallel processing modes of Open Acc based on asynchronous multi-stream proposed in this paper have better acceleration effect than the traditional processing algorithms.At the same time,through multiple functional tests on each module of the system,it also verified that the remote sensing information acceleration extraction system based on microservices and Open Acc designed in this paper has certain application value.