This question appears in Mumbai University > Distributed System Subject

Marks: 5 M

Year: Dec 2015

1 ) Null Buffer (No Buffer)

• No place to temporarily store message.

• One of following implementation strategy is used.

• Message remains in senders address space and execution of send is delayed until receiver finishes current receive.

• When receiver finishes, act is sent to sender kernel, then sender is unblocked and send is executed.

• Message is discarded & timeout mechanism is used to resend message after timeout period.

       Fig. Logical Path of MSG Transfer is directly from sender’s address   space to Receiver Address Space.

2] Single-Message Buffer:

Synchronous communication use single msg buffer because atmost, one msg may be outstanding at a time.

3] Unbounded Capacity Buffer:

In the asynchronous mode of communication since sender does not wait for receiver to ready, there may be several pending msgs, so unbounded capacity buffer is used.

4] Finite Bound Buffer (Multiple Msg):

Unbounded Capacity of buffer is practically impossible so in practice systems using asynchronous mode used finite bound buffer also known multiples msg buffer.

• With Finite bounds, buffer overflow is the problem.

• Thus can be deal with one of the following problems.

Unsuccessful Communication :

Msg Transfer simply fails when there is no more buffer space, send return error message to sending process.

Flow Controlled Communication :

You use flow control the sender is blocked until receiver accepts some messages, thus creating space in buffer for new msgs.