File Sharing Semantics , File Caching Scheme

What is file sharing semantics ?

It defines how multiple user interact with the same file in distributed system

Its important because files are replicated , cached

What are the types of file sharing sementics ?

• Unix Semantics

• Session Semantic

• Immutable File Semantic

What is Unix Semantics ?

• It follows the traditional unix file behaviour

• All clients always see the latest changes. 
  
  Ex: User A writes x = 10 then B reads immediately x = 10
  
  
  Adv: Good for systems where latest data matters
  Disadv : Slow

What is session semantics ? 

• In this user sees its own changes first 

• Other user may see old data until system syncs 

Example:

• You edit a file on your cloud drive → you see your changes instantly.

• Your friend opening the file at the same time may see the older version.
  
  Adv: Fast 
  Disadv: Not suitable when all client want to see immeidate changes

Immutable File Semantics

• In this once file is created it can’t be modified

• If we want to update the file then it will create the new file

• Consistency is guaranteed

Adv: Very easy to maintain consistency
Disadv: Requires more storage

What is file caching ?

It is when a client keeps a copy of a file or part of a file locally so taht it deosn’t need to fetch from server every time.

This makes the accessing faster

What are the main file caching schemes?

1. Write Through caching 

2. Write Back caching

3. Read Only caching 

4. Callback Scheme

What is write through caching ?

• Any updates to the file is immediately send to the server

• In this local cache and server are always exactly the same

Example : You type in a google docs → It saves instantly online

Every changes goes straight to the server

What is write back caching ?

In this changes first saved locally in cache

Later , Updates are send to server all at once or when file is closed

Example : You edit a dowload file → it only uploads when you click it save

Work locally first and then send changes later

What is read only caching?

• In this only read operation are cached

• Any write must be done directly to the server , not locally
  
  Example :
  You open a PDF from the cloud → Cache locally so next read is fast
  Editing the PDF still requires server access

Cache only for reading , write goes straight to server

What is Callback Scheme

It is a method where server keep the track of everyone who has a copy of a file

If anyone changes the file , the server will tell you that your file is outdated , update it

Ex: AFS

It simply tells when someone changes the file and tell that your file is outdated