https://pdos.csail.mit.edu/6.824/notes/l-frangipani.txt

6.5840 2023 Lecture 13: Frangipani Frangipani: A Scalable Distributed File System Thekkath, Mann, Lee SOSP 1997 why are we reading this paper? cache coherence distributed transactions distributed crash recovery and the interaction among the three what's the context? the authors' research lab programming, text processing, e-mail, &c lots of collaboration UNIX workstations in offices most file access is to user's own files need to potentially use any file on any workstation user/user collaboration one user logging into multiple workstations solution: a network file system looks to users/apps just like an ordinary local-disk filesystem but workstations talk to server to get at files this was a common scenario when the paper was written the state of the art in 1997: AFS and NFS [workstations; one AFS server] MIT Athena uses AFS: /afs/.../rtm RPC interface is system calls: create(), mkdir(), readdir(), read(), &c one central server, ex

6.5840 2023 Lecture 13: Frangipani Frangipani: A Scalable Distributed File System Thekkath, Mann, Lee SOSP 1997 why are we reading this paper? cache coherence distributed transactions distributed crash recovery and the interaction among the three what's the context? the authors' research lab programming, text processing, e-mail, &c lots of collaboration UNIX workstations in offices most file access is to user's own files need to potentially use any file on any workstation user/user collaboration one user logging into multiple workstations solution: a network file system looks to users/apps jus