Not sure to what depth you want to go, but the basic concept is fairly straightforward. If all goes well, meaning low error rate, you try to keep as many unacknowledged packets in flight as can fit within the round trip delay time. This provides for maximum throughput, because the sender does not have to wait around for each packet to be acknowledged, before sending out another one.

But if errors occur frequently enough, however "enough" is determined, you don't want to have to retransmit a lot of packets. So in the simplest case, you reduce the window size. In a more elaborate implementation, you allow selective packet retransmissions. Now the receiver will have to fill in holes in the incoming stream. This is where protocol design has to get clever.

This is a simple generic description:

http://www.csc.villanova.edu/~bwen/swp.html

I might suggest RFC 793, Section 3.7 for a good description of the sliding windows protocol as used in TCP, with a good discussion of window management.

http://www.rfc-editor.org/rfc/rfc793.txt

Then RFC 7567 goes into the latest IETF thinking on queue management techniques, which are the elaborations of sliding window protocols:

http://www.rfc-editor.org/rfc/rfc7567.txt

These RFCs are quite readable. I think your students would like the presentation.

Not sure to what depth you want to go, but the basic concept is fairly straightforward. If all goes well, meaning low error rate, you try to keep as many unacknowledged packets in flight as can fit within the round trip delay time. This provides for maximum throughput, because the sender does not have to wait around for each packet to be acknowledged, before sending out another one.

But if errors occur frequently enough, however "enough" is determined, you don't want to have to retransmit a lot of packets. So in the simplest case, you reduce the window size. In a more elaborate implementation, you allow selective packet retransmissions. Now the receiver will have to fill in holes in the incoming stream. This is where protocol design has to get clever.

This is a simple generic description:

http://www.csc.villanova.edu/~bwen/swp.html

I might suggest RFC 793, Section 3.7 for a good description of the sliding windows protocol as used in TCP, with a good discussion of window management.

http://www.rfc-editor.org/rfc/rfc793.txt

Then RFC 7567 goes into the latest IETF thinking on queue management techniques, which are the elaborations of sliding window protocols:

http://www.rfc-editor.org/rfc/rfc7567.txt

These RFCs are quite readable. I think your students would like the presentation.

Yes RFC's.. thank you Albert Manfredi, for your kind and detailed response

The data can get lost, reordered or duplicated due to the presence of routers and buffer space over th e unreliable channel in the conventional networks. The data link layer deals with frame formation, flow control, error control, an d addressing and link

management. All such functions will be performed only by data link protocols. The sliding window protocol will detect and correct error if the received data have enough redundant bi ts or repeat a re transmission of data. The paper shows the working of this duplex protocol of data link network.

Please read the attachemt file for more information 

Hi,

   I had taught this portion using "Computer System Architecture" by Morris Mano. This has been explained with a diagrammatic example and a good explanation. Hope this will help you. 

Hi,

      The textbook " Computer Networks", by Andrew S Tanenbaum has the Sliding window protocols - ARQ (Automatic Repeat Request ) topic explained very clearly with example. All the sliding window protocols would have been explained in a simple manner. Hope this will help you. 

Yes Mr Ramalingam I have that book.. I am going to have a look! Thank you

Data communication and networking - Forouzan.  chapter 11, there are a good resources for your question