[net.lan] anybody running multiple protocols on a single cable?

gamiddleton@thunder.UUCP (Guy Middleton) (05/28/85)

Sorry if this is a dumb question.
It it possible to run both (for instance) XNS and TCP/IP on a single
ethernet?  How about DECNET as well?  Is anybody doing this?
-- 
	Guy Middleton, Lakehead University, Thunder Bay, Ont.
	..{allegra,clyde,decvax,utcsrgv}!watmath!thunder!gamiddleton
	..ihnp4!mmm!thunder!gamiddleton

mark@tove.UUCP (Mark Weiser) (06/02/85)

In article <222@thunder.UUCP> gamiddleton@thunder.UUCP (Guy Middleton) writes:
>Sorry if this is a dumb question.
>It it possible to run both (for instance) XNS and TCP/IP on a single
>ethernet?  How about DECNET as well?  Is anybody doing this?
>-- 
>	Guy Middleton, Lakehead University, Thunder Bay, Ont.
>	..{allegra,clyde,decvax,utcsrgv}!watmath!thunder!gamiddleton
>	..ihnp4!mmm!thunder!gamiddleton

At Maryland we run XNS and TCP/IP on the same cable via the same interface
via hacks to 4.2bsd.
	-mark
-- 
Spoken: Mark Weiser 	ARPA:	mark@maryland	Phone: +1-301-454-7817
CSNet:	mark@umcp-cs 	UUCP:	{seismo,allegra}!umcp-cs!mark
USPS: Computer Science Dept., University of Maryland, College Park, MD 20742

ry@cadre.ARPA (Russell J. Yount) (06/03/85)

In article <222@thunder.UUCP> gamiddleton@thunder.UUCP (Guy Middleton) writes:
>Sorry if this is a dumb question.
>It it possible to run both (for instance) XNS and TCP/IP on a single
>ethernet?  How about DECNET as well?  Is anybody doing this?
>-- 


	Yes, it is quite possible to run mutiple protocols on the same
ethernet. We run both IP, CHAOSNET, and PERQ protocals on our local ethernet.
As long as all the different protocols are using different ethernet packet
types with the possible exception of a common ARP (address resolution
protocol) coexistance on same ethernet is easy.

	  Beware however some Vendors seem to claim any ethernet packet
type as there own (this is a very poor choice and not to spec.) Also
some vendors protocals tend to use broadcast packets way to much, this may
affect preformance of other machines on same ethernet by keeping their
interfaces busy (and cases heavy DMA/interrupt loads) when packets are
not used on that machine anyway and just thrown out.



	

bob@ulose.UUCP ( Bob Bismuth ) (06/03/85)

> Sorry if this is a dumb question.
> It it possible to run both (for instance) XNS and TCP/IP on a single
> ethernet?  How about DECNET as well?  Is anybody doing this?

No it's not such a dumb question - computer sales people and marketing
people frequently have the misconception that only 1 protocol type can
be used under the ethernet container protocol. This quite often gets
passed onto ethernet users.

You can have several protocol types in use on a single ethernet at once.
That is a design feature/goal of the ethernet "standard". Here at CADMUS
we run both TCP-IP and our own distributed file system protocol (UNISON)
concurrently. UNISON is completely independent from TCP-IP.

At the winter UNIFORUM, DEC was running both TCP-IP and DECnet over the
building ethernet. Of course, the only systems responding to the DECnet
protocol type were their ULTRIX systems and their token VMS system ...

	--  Bob
	    (decvax!ulose!bob)
	    
	    (Note, UNISON is a trademark of CADMUS and ULTRIX and VMS
	     are trademarks of Digital Equip. Corp.)

henry@utzoo.UUCP (Henry Spencer) (06/04/85)

Given that multiple protocols can co-exist if their type numbers don't
clash, does anybody have a list of in-use type numbers?  The mumbles
in my old copy of the Ethernet spec suggest that Xerox was trying to
maintain a master list; is it accessible?  Failing that, does anybody
know if there are specific protocol number(s) reserved for site-specific
protocols?  We may have to implement one, and I'd prefer to leave the
door open to TCP/IP etc. on the same Ethernet.
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,linus,decvax}!utzoo!henry

randy@bcsaic.UUCP (randy groves) (06/05/85)

In article <222@thunder.UUCP> gamiddleton@thunder.UUCP (Guy Middleton) writes:
>Sorry if this is a dumb question.
>It it possible to run both (for instance) XNS and TCP/IP on a single
>ethernet?  How about DECNET as well?  Is anybody doing this?

We are running TCP/IP, XNS, Ungerman/Bass, Sun ND (I think this is a TCP
clone), and DECNET all on the same ethernet cable.

All but TCP/IP have been running for months.  So far so good.

--
Randy Groves

striepe@muscat.UUCP (Harald Striepe) (06/06/85)

> Sorry if this is a dumb question.
> It it possible to run both (for instance) XNS and TCP/IP on a single
> ethernet?  How about DECNET as well?  Is anybody doing this?
> -- 
> 	Guy Middleton, Lakehead University, Thunder Bay, Ont.
> 	..{allegra,clyde,decvax,utcsrgv}!watmath!thunder!gamiddleton
> 	..ihnp4!mmm!thunder!gamiddleton

We are running quite a number of DECnet systems (VMS and Ultrix) as
well as TCP/IP. No problems.
It really works...
-- 
					Harald Striepe
					DEC Western Region TBU
					decwrl!muscat!striepe
					decwrl!dec-rhea!dec-wer521!striepe
					WER521::STRIEPE

jmg@cernvax.UUCP (jmg) (06/07/85)

In article <222@thunder.UUCP> gamiddleton@thunder.UUCP (Guy Middleton) writes:
>Sorry if this is a dumb question.
>It it possible to run both (for instance) XNS and TCP/IP on a single
>ethernet?  How about DECNET as well?  Is anybody doing this?
>-- 
Perfectly possible. We have managed all of DECNET, TCP/IP, XNS and private
protocols, on a mixture of transceiver types, controllers, computers etc.

We have even added in ISO 8802/2 style packets (protocol ID field replaced
by length field).

One main worry is to avoid protocols which make widespread use of broadcast
and multicast packets (i.e. send lots at frequent intervals).

chris@umcp-cs.UUCP (Chris Torek) (06/08/85)

>>It it possible to run both (for instance) XNS and TCP/IP on a single
>>ethernet?  How about DECNET as well?  Is anybody doing this?

>We are running TCP/IP, XNS, Ungerman/Bass, Sun ND (I think this is a TCP
>clone), and DECNET all on the same ethernet cable.

Sun's ND is stuck on top of IP, same as TCP and UDP.  It's protocol #
77.  (That's an unoffical number, by the way, and conceivably could get
preempted any time... but that's pretty unlikely.)
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@maryland