[gPXE] any negative implications of setting TCP window size to (65536 - 4) ?

[H. Peter Anvin]

On 02/17/2010 12:39 PM, Glenn Brown wrote:
> If you don't support out-of-order reception, I don't see why a TCP stack 
> needs to allocate storage for the whole window.  If packets are 
> processed as fast as they are received (in order) then it's enough to 
> keep a single buffer posted at all times.
> 
> So, why not advertise a dynamic window size and keep the small 
> buffering?  Doing so would prevent the window from throttling transfers 
> on high round-trip-time connections.

The lack of out-of-order reception drastically increases the likelihood
that you will have to do an expensive recovery, especially over WAN
links.  This problem really needs to be addressed.

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.