San Storage Strategy And Its Management Software - Design And Implementation Of The Backup Server

The expansion of IT applications has resulted in more and more data being produced, making it necessary to manage high-capacity storage systems. If the storage is connected in a disorderly fashion to different types of server, free space is wasted and it becomes difficult to manage. Furthermore, increases in the volume of data make it harder to back up.One idea that has been put forward to solve these problems is the "Storage Area Network" (SAN), which is defined as a network where storage devices and server devices are connected together by fast links such as optical fiber channels. SAN can be used to implement integrated management of storage resource, to reduce the management costs, to improve the speed of data access and the efficiency of storage space, and to increase system flexibility and expandability. The sharing of data and storage resource between different types of server also becomes possible.Storage management software is necessary for effectively managing SAN and plays a more important role in the storage system. Backuping and restoring are one of the core functionalities of software for storage management. During nearly two years from 2002 to now the author has participated in designing and implementing a storage resource management system based on SAN which we co-developed with a Japan company. With the help of ideas and methods from the traditional network management, the author studies deeply into SAN management software' s architecture, analyses the special function requirement of backup/restore software, then makes a backup server to come true. The main contributions included in this paper are the followings:1. As the result of analysis of status quo in the traditional data backuping, the main problems faced with backuping are disclosed and the logical model for backup is proposed.2. The basic tasks and modules of backup softwares are described, as also the relations with other parts of a storage management system.3. Atomic Transaction Group is brought forward and implemented on the basis of operations by conventional backup.4. Hardware information is periodically monitored and fetched from Disk arrays and is updated into locally shared memory.5. Using the disk arrays' infortions stored in the internal data structure implemented by the author and his teammates, commands from users are to be validated. So a more robust logical disk array appears in the user' s view.6. Simulating complex functions with the use of basic commands offered by hardware, users can enjoy a friendly interface which has powerful capabilities rather than simple, primitive commands.