On Wed, 30 Jul 2003, Kallol Biswas wrote: > Someone asked the following questions, I know there are many gurus on > the net: > > > on winsock: > > 1. How does an application determine send buffer when it issues send > call ? It need not determine, but if it wishes it can do so using getsockopt(SO_SNDBUF) > > 2. Is there relation between send buffer and size of tcp receive window > as sent by the receiving end ? The other ends receive buffer has a direct relation to the receive window advertised, Our send buffer size has no relation, but it does not make much sense to have a very large send buffer if the other end always advertises a small receive window. > > 3. If, in a packet trace, I see sender who puts a push bit after every > 64K, does that mean that the sender uses a 64K send buffer > when he issued send call ? Not really, but it might be so. Normally TCP sets the PSH flag on whenever it can send everything the application has asked it to. > > 4. Is the use of push bit in send call optional ? BSD sockets API do not provide any way for the application to ask TCP to set PSH flag on. > > 5. Why would you do that in a non-interactive application ? (i.e. bulk > transfer: printing). Same as above. TCP takes this decision not the app. tomar > > > > - > : send the line "unsubscribe linux-net" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
