Research And Implementation Of Massive Database System Integration Management Technique For Cluster Architecture

With rapid development of information technology and internet, a huge amount of data were produced and stored in database. With the increase of the devices and software, the problems in management also increased. Because of the huge scale of hardware device and software service, the cluster structure for massive data process system brought us some problem in management. The cluster management technique has been accepted in the world, there are research organizations and corporations in lots of countries, and many interesting results have been presented. Most of them designed for the runtime state management on one side, and considered insufficiently on the state of whole cluster system. So it is difficult to monitor and configure cluster database system in global environment. The research on management technology for massive data process system is important in both theory and practice.On the basis on analysis of features of the technique on integration management, we design and implement an integration management system for massive data processing cluster. This system based on a lot of subsystem management kit. The administrator could manage whole cluster easily by using the management system.This paper contains following contents:1. We designed a system structure for massive data processing cluster management. Global access control and runtime state view are provided in this structure. The administrator can monitor and configure all resource in hardware and software by using the system.2. On the basis on analysis of user authentication and access control in distribute environment, we designed an integration access control model (IAC). The central access control can be made through using suitable role assignment with IAC.3. We designed an integration model on runtime state of resources in management system. That can integrate lots of management information and display whole cluster state in a single view.4. We analyzed the performance characteristic of local database in multidatabase system, and a performance tuning scheme has been provided. That focuses on some aspect in database, such as storage, instance and SQL statement.We implement an integration management system based on Galaxy Massive Data Management Platform. Lots of management tool such as device management kit, database manager are integrated into a single system and all system resource is under the uniform monitor and management. The availability and manageability are improved by using the system.