[comp.protocols.tcp-ip.ibmpc] IP in netbios datagrams.

KCM@BYUADMIN.BITNET (Kelly McDonald) (05/11/88)

I'm in favor of the generic name that Dick mentioned to standardize IP
in netbios, but I believe that the netbios name limit is 16 bytes and
the one he proposes could have a maximum of 18 bytes.  I believe that
the IP address format is unique enough that labeling it with "IP:" is
not necessary. I would suggest that we simply stick to the "nnn.nnn.nnn.nnn"
format (in ascii characters) as the netbios name that is used to send
and receive datagrams.  I would assume that we all agree that we should
simply place raw IP packets in the netbios datagram without any further
encapsulation?

karl@TRWIND.IND.TRW.COM (Karl Auerbach) (05/11/88)

> I would assume that we all agree that we should simply place raw
> IP packets in the netbios datagram without any further encapsulation?

I don't agree:
	1) Some netbios implementations do name discovery for EACH
	   outgoing datagram -- this tends to triple the net traffic,
	   and much of this is usually broadcast.
	2) Other netbios datagrams broadcast netbios datagrams -- without
	   name discovery the sender does not know whether the target name
	   is a group or unique name.

In both these cases, the b'cast activity places a burden on every node
of the net.

			--karl--