Design And Implementation Of Distributed Architecture For High-precision Location Services Platform

In today’s intelligent and informative society,the application of location services is becoming more and more widespread,and the demand for application is becoming more and more exuberant.Many provinces,cities and industries have built location service platforms for their own regions and industry sectors,and vigorously develop the location service industry.However,the service scope of these platforms is relatively small,and their service contents and forms are relatively homogeneous.The Bei Dou "One Belt,One Road" high-precision location public service platform(next referred to as high-precision location service platform),on which this research is based,is dedicated to providing high-precision to the whole country and countries along the "One Belt,One Road".The overall construction of the platform will adopt the off-site multi-centre model with one main centre(Xi’an)and multiple sub-centres(Guangxi,Quanzhou,Chengdu,etc.).At present,each centre is relatively independent and data is separated,and there is a lack of systematic research and design in the overall architecture of the platform.The main problems currently faced by the platform include: the inability to schedule resources between centres so that all services need to be deployed,resulting in service redundancy,remote invocation of regional location services resulting in reduced real-time positioning,and the inability to failover when a single data centre fails,resulting in service interruption.All these problems make the platform resources not fully utilized and the overall service efficiency cannot be improved.Therefore,it is of great significance and value to study the integration and interoperability of the main centre and sub-centre system architecture and the mutual support of business services.In this paper,in view of the different business characteristics of high-precision location services and the development layout of off-site multi-centres,research is conducted on how to achieve flexible scheduling of resources and mutual support of business between the main centre and sub-centres.The distributed architecture is designed and optimized in terms of user traffic distribution,load architecture,service governance architecture and data storage and caching architecture,realizing the high concurrency and high availability of the high-precision location service platform and the "off-site multi-live" of the platform business in each centre,solving the problems of resource wastage and business redundancy caused by the inability of the main centre and sub-centres to work together,and forming a set of distributed off-site multi-live architecture for reference by the off-site multi-centre business system on this basis.The main work of this paper is as follows:(1)Research the functional and performance requirements of the distributed high-precision location service platform,and complete the design of the overall functions and the formulation of performance indicators of the platform.In addition,in view of the different characteristics of the platform’s services and the different needs of users,this paper researches and designs three service modes of the platform’s external services: single-user terminal location mode,industry user terminal location mode and single-user platform location mode,which can meet the service needs of various modes such as single data request service of users,long connection data request service of sub-centres and user operators and user access to platform location service.(2)Research and design a distributed off-site multi-live architecture of the high-precision location service platform.To address the problems of user traffic distribution and data consistency in distributed off-site multi-live architecture,this paper researches and designs a user directed access scheme based on intelligent DNS(Domain Name System),which realizes the first level of user traffic distribution and solves the problem of inconsistent data caused by the same user operating in different data centres.To address the problem of business deployment in each centre and improving the concurrency and availability of the architecture,this paper studies the detailed division of business modules and their specific deployment in each centre,and designs a highly available load architecture based on Nginx and Keepalived to improve the concurrency and availability of the platform architecture.To address the problem of difficult governance of many services in the service layer,this paper studies service splitting and deployment in the service layer,and designs the Dubbo distributed service governance architecture based on Zookeeper implementation,which achieves efficient governance of many services and improves the scalability of the service governance architecture.(3)Research and design the data storage and caching architecture of the high-precision location service platform.For structured data storage,this paper designs a database architecture based on Mycat cluster + Keepalived + HAproxy,which improves the scalability and availability of the database architecture;For data frequently accessed and frequently written by users,this paper studies and designs a caching architecture based on Redis,which reduces the pressure on the data server for reading and writing;Finally,for the storage of unstructured massive files,this paper proposes a distributed parallel storage architecture based on the Parastor distributed file system,and studies various aspects of its system architecture,data security and scalability.(4)Based on the functional and architectural design of the platform,this paper deploys and implements the platform functions,and tests the platform functions and performance based on a simulated test environment.The functional test results show that the functions implemented in the platform are consistent with the functional design and can provide normal services to users.The performance test results show that the distributed off-site multi-live architecture can support 10,000 concurrent users with a response time of about 55 ms for a single user in a simulated test environment,and higher concurrency can be supported by improving the performance of the test server,which can meet the performance requirements of the platform;At the same time,each centre can provide real-time PPP(Precise Point Positioning)and network RTK(Real-time Kinematic)services with the same positioning performance,realizing the "multi-site" of platform services in each centre and verifying the effectiveness of the distributed architecture design.