On Jul 9, 2010, at 6:26 AM, Bilal Shahid wrote: >> Do you have any reason to believe that if you implement your own protocol >> (or re-implement existing protocols) you will actually reduce latency? > > It bypasses OS and transport and application layer by going directly to wire > is the reason why I think it will reduce latency.
It might reduce *OS* latency by avoiding some OS layers (the network and transport layer; if you're using TCP sockets, *your code* implements the application layer, there's no application layer in the OS to bypass). However, if you then have to implement the network and transport layer yourself, that latency will come back, if you measure the latency in a consistent fashion (i.e., by counting latency in your code, as well as in the OS code); the only difference is that the latency will be in your code, not in the OS code, but it's still there. _______________________________________________ Winpcap-users mailing list [email protected] https://www.winpcap.org/mailman/listinfo/winpcap-users
