14th International Conference on Distributed Computing Systems (1994)
June 21, 1994 to June 24, 1994
R. Panadiwal , Sch. of Comput. & Math., Deakin Univ., Geelong, Vic., Australia
A.M. Goscinski , Sch. of Comput. & Math., Deakin Univ., Geelong, Vic., Australia
In this paper we demonstrate that the typical problem of loss in performance due to disk I/O can be efficiently dealt with by using separate strategies for the storage of file data and for the data structures required for file management. Furthermore, to make the design very reliable, stable storage is provided. The design of the disk service allows the contents of a file to be distributed among more than one disk drive. Striving for reliability has also generated a new file facility architecture where a separate interface is provided for the transaction service. The claim that using transaction semantics file operations in not only database applications but also in system programming can be made resilient against system and media failure. In order to further improve performance the design of the caching module takes into consideration all the aspects of basic file and transaction services.<
data structures, file servers, network operating systems, transaction processing
R. Panadiwal and A. Goscinski, "A high performance and reliable distributed file facility," 14th International Conference on Distributed Computing Systems(ICDCS), Pozman, Poland, 1994, pp. 116-123.