[OpenAFS-devel] .35 sec rx delay bug?

[Niklas Edmundsson]

On Wed, 8 Nov 2006, Ken Hornstein wrote:

>> What would it be that makes packet fragmentation and
>> reassembly so immensely more efficient compared with RX packet
>> handling? Why can TCP fill up a GigE leisurely and RX just gets about
>> half of it sweating a complete CPU?
>
> Fundamentally ... the real issue is that large numbers of smart people
> have spent a lot of time tuning TCP on various platforms for more than
> a decade.  In that same time, Rx has received very little attention.  I
> was able to fill a Gig-E pipe with RxTCP with relatively little work
> (by "little" work I mean tuning I did to RxTCP - RxTCP ended up being a
> lot of work).  All that I need to do now is finish the damn thing and
> get it working in real clients and servers.  Hopefully soon.

When we're at the subject of already tuned solutions: What 
possibilities are there to combine this with the use of sendfile()? 
Experiences from apache and other projects show that there are very 
noticeable effects even when doing sendfile() on small chunks compared 
to the classic read/write-approach.

/Nikke
-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
  Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se     |    nikke@hpc2n.umu.se
---------------------------------------------------------------------------
  "I'm Luke Skywalker.  I'm here to rescue you."
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=