{"title":"Design for a transparent, distributed file system","authors":"D. Mecozzi, J. Minton","doi":"10.1109/MASS.1991.160215","DOIUrl":null,"url":null,"abstract":"At the Lawrence Livermore National Laboratory (LLNL), caching and migration protocols have been designed to integrate distributed UniTree File Management System servers running on separate machines to create a single file system. These protocols allow files to migrate between levels of a storage hierarchy to create a unified distributed storage system. The design provides clients with a single method for accessing files, regardless of file location. File caching provides clients with optimal performance, while file migration enables file servers to optimize their space utilization. The key features of the system include use of unique, location-independent file caching, and a locking mechanism to synchronize access to the system's files and manage conflicts related to multiple copies of the files. A shift in LLNL policy to acquire vendor-supported software prevented the completion of the implementation of this unified storage system. However, the design solves many problems that can occur when providing a transparent distributed file system.<<ETX>>","PeriodicalId":158477,"journal":{"name":"[1991] Digest of Papers Eleventh IEEE Symposium on Mass Storage Systems","volume":"17 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1991-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"[1991] Digest of Papers Eleventh IEEE Symposium on Mass Storage Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/MASS.1991.160215","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 7
Abstract
At the Lawrence Livermore National Laboratory (LLNL), caching and migration protocols have been designed to integrate distributed UniTree File Management System servers running on separate machines to create a single file system. These protocols allow files to migrate between levels of a storage hierarchy to create a unified distributed storage system. The design provides clients with a single method for accessing files, regardless of file location. File caching provides clients with optimal performance, while file migration enables file servers to optimize their space utilization. The key features of the system include use of unique, location-independent file caching, and a locking mechanism to synchronize access to the system's files and manage conflicts related to multiple copies of the files. A shift in LLNL policy to acquire vendor-supported software prevented the completion of the implementation of this unified storage system. However, the design solves many problems that can occur when providing a transparent distributed file system.<>