Distributed File Systems

1. Distributed File Systems

2. File System Modules • Directory • File Module • Access Control • File access • Block • Device

3. File System Operations • Open • Create • Close • Read • Write • Seek • Stat • Link • Unlink

4. Distributed File System Requirements Transparency • Access • Location • Mobility • Performance • Scaling

5. Reqmts contd.. Concurrent File Updates • Changes to a file by one client should not interfere with the operations of other clients simultaneously accessing or changing the same file Done by -File Locking -Record Locking

6. Reqmts cond.. File Replication • Represented by several copies of its contents at different locations Provides - Sharing the load of Multiple Servers - Fault tolerance

7. Reqmts Contd.. Hardware and OS heterogeneity • Design of Service Interface should be implemented such that client and server software can be implemented for different operating system

8. Reqmts contd Fault tolerance • Service should continue in the face of client and server failures • Server protocol can be designed with idempotent operations • Stateless servers

9. Reqmts contd Consistency Unix offer one-copy update Semantics • As if a single copy of the file existed • Propagation delay

10. Reqmts contd Security • Based on Access Control List • Authenticate Client Requests • Digital Signature • Encryption

11. Reqmts contd Efficiency A DFS should provide a service that is comparable with or better than, local file systems in performance and reliability

12. File Service Architecture • Flat file service • Directory service • Client Module

13. Flat file service • Concerned with implementing operations on the contents of the file • UFID – Unique File Identifier • Each file has a unique UFID • Generates a new UFID when a request to create a file is received

14. Directory Services • Provides a mapping between text names for files and their UFID’s • Text name  DS UFID • Functions • Generate Directories • Add new file to directories • To obtain UFID

15. Client Module • Integrating the operations of a FFS and DS under a singel interface • Holds info about n/w locations of FFS and DS processes • Caching of recently used file blocks • Increases performance

16. FFS INTERFACE • Read(FileId, i ,n) • Write(FileId,i,n) • Create() • Delete(FileId) • GetAttribute(FileId) • SetAttribute(FileId)

17. Properties • Repeatable Operations • Idempotent Operations Except Create • Stateless Servers • can be restarted after failures without any need for client and server to restore any state

18. Access Control • An access check is made whenever a name is converted to a UFID and results are encoded in the form of a capability which is returned to the client for submission with subsequent request • A user id is submitted with every request and access checks are performed by the server

19. Directory Service Interface • Lookup(Dir, Name) • AddName(Dir,Name, File) • UnName(Dir, Name) • GetNames(Dir,Pattern)

20. Heirarchical File System • No of directories arranged in tree structure • Files at leaves and directories at other nodes • The root is a directory with a UFID • A function that gets UFID of a file given its pathname • File Attribute – a type field

21. File Grouping • A collection of files located in a server • Representation of UFID includes a file group identifier component. • File group Identifier should be unique • Generated by concatenating • 32 bit IP address of the host • With a 16 bit integer derived from the date