Comparative Analysis of Distributed and Parallel File Systems' Internal Techniques
share
A file system optimization is the most common task in the file system field. Usually, it is seen as the key file system problem. Moreover, it is possible to state that optimization is dominant in commercial development. A problem of a new file system architecture development arises more frequently in academia. End-user can treat file system performance as the key problem of file system evolving as technology. Such understanding arises from common treatment of persistent memory as slow subsystem. As a result, problem of improving performance of data processing treats as a problem of file system performance optimization. However, evolution of physical technologies of persistent data storage requires significant changing of concepts and approaches of file systems' internal techniques. Generally speaking, only trying to improve the file system efficiency cannot resolve all issue of file systems as technological direction. Moreover, it can impede evolution of file system technology at whole. It is impossible to satisfy end-user's expectations by means of file systems optimization only. New persistent storage technologies can question about file systems necessity at whole without suggestion of revolutionary new file system's approaches. However, file system contains paradigm of information structuring that is very important for end-user as a human being. It needs to distinguish the two classes of tasks: (1) optimization task; (2) task of elaboration a new architecture vision or paradigm. But, frequently, project goal degenerates into optimization task which is meant really elaboration of a new paradigm. End-user expectations are complex and contradictory set of requirements. Only optimization tasks cannot resolve the all current needs of end-user in the file system field. End-user's expectations require resolving tasks of a new architecture vision or paradigm elaboration.
READ FULL TEXT