🚀 go-pugleaf

From article <C76A6E12D11F801957@HEARNVAX.nic.SURFnet.nl>, by REIJS@SURFNET.NL:
> Hello all of you,
> 
> 
> I have a theoratical and also pratical question. When using the
> packet-size negotiation of X.25 what is intended to happen?
> Because X.25 is a DTE-DCE protocol (on all levels), is the packet
> size negotiation theoreticaly only for the local interface between DTE
> and DCE, or is the interface at the remote side also involved (that is what
> pratis is telling me).
> 
> Furthermore I thought that the network could help in different packet size
> at the remote and local side (by using buffers and M-bits it can help
> de DTE's in communicating).
> 
> So my question is: Is pacjet-size negotiation (the same for window
> and throughput) an end-to-end facility in theory and in practice.

In theory it is not an end-to-end facility, as the original idea of
X.25 was to be only an interface protocol, not a full network protocol,
so there was never the thought of having X.25 throughout the net from
end to end.

In practise i have not seen any X.25 switch may it be working on
level 2 or level 3, that uses different packet sizes on the two
ends of a X.25 connection, i.e.: i have not seen any switch doing
assembly or reassembly of X.25 packets. Possibly some switches
(esp. those at the PSDN) can do this, but the configuration has been
selected so as not to do it, possibly because there is a performance
overhead associated with doing so. In theory only an IWU could
really do these things, but sometimes it's difficult to decide
if a vendor has really a level 2 switch or an IWU.


      Toerless Eckert       | /C=de/A=dbp/P=uni-erlangen/OU=informatik/S=eckert
Imagine it is war and your tv set is broken | eckert@informatik.uni-erlangen.de