A Database Design And Implementation Of Disaster Recovery Technology

Database plays a key role in enterprises. Some large scale internet companies,banks, governments,telecommunication companies and so on, become more and more dependent on database. However, if these organizations don't take any measure to protect data, the data is likely to be damaged by various factors, nature or human being itself. So the whole service system will be interrupted if the data is destroyed. They will greatly suffer from the pause of computers or networks, even just a little while. At the same time, there are a lot of Trojans and network attacks in computer systems. All these can probably lead to the loss of important data, gained by other people, organizations or companies. It is common that the deletion tool in operating system can't completely delete the data from the disk at all, for it can be recovered by some special tools. In this case, a special tool is needed which can delete the data from the disk forever.Most database management systems provide some basic tools for backup and recovery. But it needs operating manually. Furthermore, the function is also limited because the data is only backuped locally. And in case of disaster, the backup data is likely to be damaged. It will take a long time for recovery, so it can't resume database service immediately. If so, it can't meet the requirement of disaster tolerance. The thesis has analyzed the current research around database disaster tolerance technology. The paper presents: a method of database synchronization for Oracle, a method of network packet redirection based on NDIS and a method of fast and complete destroy for a number of files. Based on the methods mentioned above, the paper has designed and realized service disaster tolerance sub-system and data destroy sub-system in database disaster tolerance system.Specifically, the contributions of the paper include:1. Present a method of database synchronization for Oracle. The method can monitor the change in local database efficiently. Then these changes can be transferred to remote standby database and redone. Through this, it can keep the same pace between the primary database and standby base.2. Present a method of network packet redirection based on NDIS, which can capture the packets to specific IP and port and redirection to another host.3. Present a method of fast and complete destruct for a number of files. Fast destruct can destroy a lot of files in short time locally while complete destruct can destroy them completely.4. Present a method of service switch, which can be used in database disaster tolerance system. When main database crashes, with"a method of network packet redirection based on NDIS,"this method can redirect the packets, which is sent to main database, to standby database. It can sustain the service while customer can't feel the interruption of service.5. Present a method of destroy for database, which is used to delete database files, important logs and temporary files in time after database is out of control in order to protect the key data from letting out.The methods presented in the paper are applied in our database disaster tolerance system. Service disaster tolerance sub-system and data destroy sub-system in database disaster tolerance system are the central sub-systems. Now, this system is widely used, which also enjoys great popularity among customers.