Summary of Storage and File Structure.

Summary

• Several types of data storage exist in most computer systems. They are classified by the speed with which they can access data, by their cost per unit of data to buy the memory, and by their reliability. Among the media avail- able are cache, main memory, flash memory, magnetic disks, optical disks, and magnetic tapes.

• Two factors determine the reliability of storage media: whether a power failure or system crash causes data to be lost, and what the likelihood is of physical failure of the storage devise.

• We can reduce the likelihood of physical failure by retaining multiple copies of data. For disks, we can use mirroring. Or we can use more sophisticated methods based on redundant arrays of independent disks (RAIDs). By striping data across disks, these methods offer high throughput rates on large accesses; by introducing redundancy across disks, they improve reliability greatly. Several different RAID organizations are possible, each with different cost, performance and reliability characteristics. RAID level 1 (mirroring) and RAID level 5 are the most commonly used.

• We can organize a file logically as a sequence of records mapped onto disk blocks. One approach to mapping the database to files is to use several files, and to store records of only one fixed length in any given file. An alternative is to structure files so that they can accommodate multiple lengths for records. There are different techniques for implementing variable-length records, including the slotted-page method, the pointer method, and the reserved-space method.

• Since data are transferred between disk storage and main memory in units of a block, it is worthwhile to assign file records to blocks in such a way that a single block contains related records. If we can access several of the records we want with only one block access, we save disk accesses. Since disk accesses are usually the bottleneck in the performance of a database system, careful assignment of records to blocks can pay significant performance dividends.

• One way to reduce the number of disk accesses is to keep as many blocks as possible in main memory. Since it is not possible to keep all blocks in main memory, we need to manage the allocation of the space available in main memory for the storage of blocks. The buffer is that part of main memory available for storage of copies of disk blocks. The subsystem responsible for the allocation of buffer space is called the buffer manager.

• Storage systems for object-oriented databases are somewhat different from storage systems for relational databases: They must deal with large objects, for example, and must support persistent pointers. There are schemes to detect dangling persistent pointers.

• Software- and hardware-based swizzling schemes permit efficient dereferencing of persistent pointers. The hardware-based schemes use the virtual- memory-management support implemented in hardware, and made accessible to user programs by many current-generation operating systems.