Design And Implementation Of A High-Availability Data-as-a-Service (DaaS) Platform

With the widespread application of technologies such as cloud computing,big data,artificial intelligence,and the Internet of Things,the volume of data has been growing explosively.However,the proportion of data that is actually utilized remains low.In order to improve the utilization of data and accelerate data circulation,many researchers have devoted themselves to the solutions of Data as a Service(DaaS)platforms.However,in current solutions,the focus is mainly on uploading or synchronizing data to a central server and providing access to the data through API interfaces.This approach poses risks of data leakage,high storage costs,and limited service capabilities as it only supports the built-in services of the platform.Therefore,this thesis aims to investigate an open DaaS platform where data remains stored on the data provider’s side.The data provider only needs to connect the metadata of the data to the platform,which takes responsibility for deploying data applications developed by data consumers to execute computational tasks on the data provider’s side.These applications are then published as data services to serve users.The paper also focuses on ensuring the reliability of data services and studying the high availability of the platform.A high availability DaaS platform is designed and implemented.The main research contributions of this thesis include:(1)Analysis of the existing issues in DaaS platforms and the proposal of an open DaaS platform.Based on the requirements of the DaaS platform,this thesis designs and implements the processes of data service creation,publication,and invocation.The platform only requires the integration of metadata from data providers,thus avoiding the data leakage issues present in centralized solutions.(2)Addressing the problem of network incompatibility between data providers and the platform,this thesis designs and implements a network tunnel between them.This includes the design of tunnel routing tables and load balancing,ensuring that requests are evenly distributed to tunnel clients.(3)Meeting the high availability requirements of My SQL clusters,this thesis designs and implements the My SQL-Manager program to monitor the My SQL cluster.It automatically recovers the cluster nodes in case of anomalies and notifies administrators when automatic recovery is not possible.(4)Addressing the high availability requirements of the DaaS platform,this thesis designs and implements the DaaS-Operator program for the automated operation and maintenance of the platform services.It attempts automatic recovery in case of platform service anomalies and notifies administrators when automatic recovery is not feasible.Finally,this thesis constructs a test environment consisting of a three-node management node cluster and two local physical nodes.The platform is tested and analyzed from the perspectives of functionality,high availability,and performance.The test results demonstrate that the DaaS platform achieves basic functionality while exhibiting good availability and performance characteristics.