[X-Unix] [OT] TCP ACK...

[Doug McNutt]

At 17:15 +0000 2/2/06, Simon Forster wrote:
>Can someone with a better understanding of the TCP networking  protocol than me tell me what should happen if machine B is receiving  a bunch of data, sending TCP ACKs quite cheerfully and then machine A  doesn't receive one of the ACKs? Who should jump in next? My layman  view is that machine A should resend its previous bunch of data, to  which machine B will again send an ACK - but I may well be wrong. I'm  sure that this is documented but I haven't succeeded in finding an  appropriate link yet - and I don't want to have to spend days  understanding all the minutiae of detail surrounding TCP data flows  before I find a suitable answer.

I just finished a first read of  "The Design and Implementation of the 4.4BSD Operating System. Chapter 13 "Network Protocols" is the best description of UDP and TCP I have seen. McKusick et al, ISBN 0201549794.

But it's about 40 pages of dense reading and I certainly did not absorb it all.

One thing that I do remember is that NAK's and ACK's from the receiving side are delayed purposely so they can be included in a datagram that will surely be sent back quite soon. The reason is to avoid short packets when delay is not critical. The triggering symptom is receipt of a sequence number that isn't one more than the previous one.

References quoted are to RFC 768, 791, 793, 792, 919, 950, 1191
-- 

--> Life begins at ovulation. Ladies should endeavor to get every young life fertilized. <--