Re: Where we stand and where we are going

[Patrik Fältström <[e-mail]>]

--On 2002-06-27 09.19 -0400 Michael Mealling <[e-mail]> wrote:

> My proposed solution is to limit UDP packet sizes to 512 bytes and put
> packet sequence numbers on them. You still have a connectionless
> interaction but it a) puts the packet size into a realm with a higher
> probability of success and b) allows for a handful of those packets to
> get through. I'm not sure if you need more than that. You can still do
> the "well if  that didn't work I can always do TCP"...
> 
> I really would like to hear some definitive answers on this topic because
> it keeps coming up and a solution would make a heck of a lot of sense.

As soon as you will send more than one packet, you need congestion control
and other things from TCP.

I rather say one can send data in either "one packet" or "many". Many
packets should be over TCP (or something else with congestion control)
while one packet can be over UDP.

That said, any protocol today needs to have negotiation/announcement of
capabilities from both a client and server perspective.

So, client send one packet to server. If server detect the packet was
fragmented, and some data which was wanted was missing, it can send back
"try again over better transport".

My question is: Can not a packet include in the beginning a length
indication in the beginning of the packet so the server can detect if MTU
was too small?

   paf