Site Loader
Rock Street, San Francisco

Concurrency/Time factor – In distributed system computers
are connected through network and are set by their own time. The only method of
communication between any programs is only messages and that also depends on
the time coordination. What if the clients time is different? if a client is
sitting in a different part of the world, how accurate is the time
synchronisation going to be for the distributed system? This is difficult for
distributed system to synchronize process due to lack of shared memory and a
global clock.  Usually when a system synchronizes
processes running at different computers its concurrently access a shared
resource but with Distributed system it is safer to let one single process to
access the resource at a time. That is when the access is atomic. So, in this
case only one process is allowed to execute at one time.

Independent Failure – If there is any type of network
failure between a client or components of the distributed system then how is it
possible to fix this problem when the other client might not be aware of the
issue immediately. The components of Distributed system are built in way that
the it doesn’t stop other processes and leaves the system running. The
development of distributed system is for users that needs computational power
at low cost. For example, when a group of members wants to communicate with
each other and share expensive resources between them.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Error Handling –  With
distributed system its difficult to handle when a problem occurs, while message
passing communication model provides high flexible communication ability,
programming such model can get difficult to handle as more factors needs to be
to be considered. This is where Remote Procedure Call comes in, it extends the
uniprocessor call mechanism (a single central processing unit that is used to
execute computer tasks) to transfer control and data over a communication
network. A RPC can fail for two reasons, computer failure or communication failure.
For example:

Message is lost: when a remote server is slow, the program that’s
supposed to invoke the remote procedure once, ends up doing this procedure more
than once, suspecting message lost. This may result more than one execution of the
procedure.

Client machine crashes: Client machine crashes after invoking
the 

Post Author: admin

x

Hi!
I'm Dora!

Would you like to get a custom essay? How about receiving a customized one?

Check it out