On Fri, Apr 28, 2006 at 06:14:03PM -0400, James Carlson wrote:
> Nicolas Williams writes:
> > Are there apps that want one more confidence than the first scenario
> > provides, but less than 100%?  I don't know, but clearly they can make
> > do without support for the middle two scenarios.
> 
> Yes, I understood what you were getting at.
> 
> The part that I don't see is why validating those middle two layers
> and getting "some" added reliability is really worth anyone's time.
> 
> There are better ways to do this.  For such a hypothetical
> application, a better way would be to introduce a timer-based
> unidirectional ack.  The sender just does write(2) and forgets about
> it.  Once every <timer-interval> seconds, the peer sends back an
> application-layer acknowledgment indicating the newest application
> (sender-specified) sequence number it has correctly processed.

Yes, and this doesn't expose lower layer details to the app, so it keeps
the app more transport independent.

_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to