Dsitributed File Systems

1
Dsitributed File Systems

• Prepared By-
• Monika Patel.

2
Outline-

• What is Distributed File System? (DFS)
• Characteristics of DFS
• Transparency
• Concurrent File Updates
• File Replication
• Fault Tolerance
• Security
• IBMs DCE Distributed File Service
• 
  

3
Why we need File System?

• Persistent Data objects need to be stored on a
  long term basis for later use. They must be named
  and saved on a non volatile data Storage device
  such as magnetic disk.
• This data objects are known as a file.
• In any operating system File system is a major
  component and its responsible for the
  organization, storage, retrieval, naming, sharing
  and protection of files.

4

• File systems also take responsibility for the
  control of access to files, restricting access to
  files according to users authorizations and the
  type of access requested
• The term "metadata is often used to refer to all
  of the extra information stored by a file system
  that is needed for the management of files
• The file system is responsible for applying
  access control for files

5
Distributed File System

• Its an implementation of a file system that
  consist of physically distributed storage sites
  but provides a traditional centralized file
  system view for the user.

6
DFS Characteristics-

• Transparency
• Access Transparency-A user can login in at any
  host in the system with uniform login procedure,
  independent of the host.
• In other word program designed to work on
  a local file can work on a remote file without
  modification
• Location Independency-files or groups of files
  can be
• relocated without changing their actual path
  (or file) names.

7
Transparency cont.(DFS introduction
,lasnost.itcarlow.ie/barryp/slides/DFSintro.pdf )

• Mobility transparency- A file may be moved,
  either by systems administrators or
  automatically.
• Performance transparency- programs should
  continue to perform satisfactorily, even under
  heavy load.
• Scaling transparency- file systems can be
  expanded by incremental growth to deal with a
  wide range of loads and network sizes.

8

• Concurrent File Updates-Changes to a file by one
  client should not interfere with the operation of
  other clients simultaneously accessing or
  changing the same file.
• Transaction based application requires that the
  application appears to access the files in
  isolation.
• For any DFS its desirable to support concurrency
  control at transaction level.

9

• File Replication-Several copies of a file's
  contents at different locations enable multiple
  servers to share the load of providing the
  service
• DFS with replication transparency performs atomic
  updates on the replicas and client are not aware
  that there is more then one copy exist.

10

• Fault tolerance-It is essential that the file
  service continue to operate in the face of client
  and server failures.
• Another important design goal is statelessness
  the File server can be restarted and the service
  restored after a failure without any need to
  recover "previous state.

11

• Security- To access control lists, there is a
  need to authenticate client requests so that
  access control at the server is based on correct
  user identities.
• There is also a need to protect the contents of
  request and reply messages with digital
  signatures and encryption of secret data.

12
IBMs DCE Distributed File Service
characteristics (By Agustin Mena III and Carl
Burnett)

• Client Caching-Client caching allows the
  majority of file read/write operations to occur
  locally on the client, thereby avoiding network
  traffic and server resource utilization.
• Maintaining cache Consistency-Cache consistency
  is maintained internally in DFS through the use
  of tokens. Tokens are guarantees from the file
  server that a client can perform certain
  operations on files residing in the client cache.
  
• Security-DFS is tightly integrated with the DCE
  Security Service. DCE Access Control Lists (ACLs)
  protect files in DFS. Communication between DFS
  clients and servers takes place using an
  authenticated RPC mechanism

13
Refrences-

• 1Performance Characteristics of the DCE
  Distributed File Service By Agustin Mena III and
  Carl Burnett.
• http//www.306.ibm.com/software/network/dce/librar
  y/whitepapers/dfsperfhtmlHeader_3.
• 2 http//glasnost.itcarlow.ie/barryp/slides/DFS
  intro.pdf
• 3 Distributed operating systems Algorithms,By
  Randy chow,Theodore johnson.
• 4 Distributed File Systems Pierre Boulet
  Masters Informatique TIIR et IAGL September
  28,2006.

14

• Thank You.