1

While reading Issues in TCP Vegas by Richard J. La, Jean Walrand and Venkat Anantharam (available online here), I stumbled across the following sentence (Part 3 - Rerouting, page 4):

When the propagation delay of connection i is d, the expected number of backlogged packets of the connection is w - rd, where w is connection i's window size and r the flow rate.

If the number of backlogged packets of the connection means the number of packets present into the network, that is the number of sent packets that didn't arrive yet, and if the flow rate is the number of packets that can be put into the network per unit of time, then I don't really understand why w - rd would be the number of backlogged packets.

In fact, if w - rd is the number of packets present into the network, then rd is the number of packets that have left the network but whose acknowledgment didn't arrive yet, and this seems to be false to me because I understand rd as the maximal number of packets that can be present in the network at any time.

Can anybody prove why this formula is working ?

  • Did any answer help you? If so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin Apr 1 '18 at 22:39
2

I'm affraid your assumption is wrong. backlogged packets are packets passed to the TCP stack (i.e. with write() syscall) but not sent to the network yet.

w is the peer's window size, or the buffer sizes. rd (or BDP) is indeed the maximum number of bytes that can be sent without tcp ack.

so w - rd is the number of bytes that can't be sent on the wire if present in the sender tcp stack (i.e. backlogged).

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.