Chapter 10. Distributed File Systems

  • 10.1 The Basics

  • 10.2 NFS Version 2

    • 10.2.1 RPC Semantics

    • 10.2.2 Mount Protocol

    • 10.2.3 NFS File Protocol

    • 10.2.4 Network Lock Manager

  • 10.3 Common Internet File System (CIFS)

    • 10.3.1 Server Message Block (SMB) Protocol

    • 10.3.2 Opportunistic Locks

  • 10.4 DFS

  • 10.5 NFS Version 4

    • 10.5.1 Managing State

      • Mandatory Locks

      • Maintaining Order

    • 10.5.2 Dealing with Failure

  • 10.6 Conclusions

  • 10.7 Exercises

  • 10.8 References

Distributed file systems straddle the boundary between what is covered in a course on operating systems and what is covered in a course on distributed systems. They provide common file access to a collection of computers; all computers in the collection can access files as if they were local, even though the files don't reside in their local file systems. The need for such a distributed system grew out of the switch in the 1980s from multiuser time-shared systems to networks of workstations (or personal computers). A community of people using a time-sharing system shared a single file system, all files were accessed identically by programs, and files were shared trivially. Furthermore, the file systems were easily administered — they were all attached to one machine and could be easily backed up. A distributed file system provides all the benefits of the file system on a time-sharing system and opens them up to users of a collection of personal computers.

The history of distributed file systems goes back to the late 1970s and early 1980s. The first commercially ...

Get Operating Systems in Depth now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.