Sockets - send()

Fairlight fairlite at fairlite.com
Thu Oct 30 19:04:19 PDT 2008


Four score and seven years--eh, screw that!
At about Thu, Oct 30, 2008 at 09:32:25PM -0400,
Brian K. White blabbed on about:
> >> So, _usually_ 1000 bytes of payload data are going to be small
> >> enough that nothing ever ends up needing to break the packet
> >> into smaller packets. 1000 bytes from you at the application
> > 
> > Negative.  I've seen breaks at far smaller (e.g., 304 bytes, or even 84
> > bytes) with MTUs of no less than 1456 along the whole chain.
> 
> You can't say I'm wrong and then correct me by saying exactly the
> same thing.

Maybe I wasn't emphatic enough.  There's a first for everything.  :) :) :)

Your "_usually_" is a complete misnomer.  I've seen regular and repeatable
packet fragmentation at -well- under the MTU in various environments.  It's
not just that it sometimes fragments them, it's that it happens a lot
dependant on how the client and server are written, the platform's network
stack, the network and system latencies, etc.  Not just a one-off here or
there, but a whole series of fragmented transmissions to the degree you
could pretty much bank on fragmentation well below the MTU threshhold.

> The rest after this point was indeed all valid.
> And good to know. Thanks.

Yuppers.  Thanks back.

Sorry if I was more abrasive than usual.  Rough couple of weeks.  Still
recovering from some stuff and I'm pretty testy.  I recognise that.  Sorry.

mark->
-- 
"I'm not subtle. I'm not pretty, and I'll piss off a lot of people along
the way. But I'll get the job done" --Captain Matthew Gideon, "Crusade"


More information about the Filepro-list mailing list