The Design And Implementation Of HBase Accesss Control System In 4A Unified Management Platform

With the development of Internet,the quantity of information that Internet produced is increasing in geometric order of magnitude.All kinds of data,especially semi-structured data's growth are far more expected than traditional database can handle,and in this case,the big data technology has appeared which now is used by more and more companies.HBase as one of NoSQL database born under the big data backgroud,it is based on Hadoop Distributed File System which produces high performance of IO,as well as easy and highly scalable structure.HBase is also based on MapReduce program model which produces an efficient data mining method on the data itself stored.Many enterprises have used HBase as main database schema.However,with frequently data security accident,the low security level of big data has became the hard point of using HBase.For company in telecom and financial area,it is hardly to use HBase on their project without security,even HBase has such amout of adventages.In the area of enterprise information system security,4A system which has been proved to be a reliable security method.4A system means centralized authentication,account,audit and authorization,and it will relatively restrict the access of hardware and software resource in a company than before.However,high concurrency and throughput often come with those tech of big data such as HBase,and traditional audit tech is no longer matched.This research is based on a company's 4A system,designed and realized an HBase access control extension plugin on its theory.According to the trend of big data security control in recent years,by researching the features and workflow of current distributed data processing software.and combining the tech of HBase and 4A system,this thesis develops and investigates the HBase access control system in 4A unified management platform which means centralized access control on HBase.4A system is able to realize access management of HBase,and can be compatible with old 4A system.It makes HBase easier and safer in ERP system.First,the component improves the process of HBase's authentication,with impersonation proxy function of Hadoop it adapts the third authentication of 4A to the kerberos authentication of HBase and it centralizes authentication,authorization and account of HBase.Security,the component is based on the coprocessor framework of HBase,improves its audit function.It can change the audit strategies dynamically without restarting the whole cluster.Meanwhile,It can get more details of accessed object and accessed resource content for audit.In some occasions,it supports data desensitization by sacrificing some performance which advances the data security performance of HBase.