Subject: Re: real ARP, anyone?
To: None <matt@lkg.dec.com>
From: Ignatios Souvatzis <is@beverly.rhein.de>
List: tech-net
Date: 10/24/1995 09:00:06
Hi Matt,

> 
> In  <199510231755.SAA00931@theory.cs.uni-bonn.de> , you wrote:
> 
> > Hello,
> > 
> > does anyone work on a real ARP implementation? I mean, one which can
> > handle multiple protocols for link level and user protocol?
> 
> ARP is a protocol defined in Internet RFCs for various media.
> It is not, in my view, a general purpose algorithm.  

No, it is _defined_ and designed to be generic in rfc826 by Davit C Plummer.
The rfcs for specific media just define the protocol id numbers and address 
lengths, which should be inserted. It was a quite ingenious piece of work, 
foretelling any possible use.

Btw., the author was an old DECineer: he used symbolic names like
ARP$CODE for the fields of the packet and the variables of his
pseudo-code. :-)

> > No, FDDI does not count: it seems to masquerade as Ethernet for ARP
> > purposes. 
> 
> Of course it does.  That's what the RFC for IP over FDDI tells you to do.

Ah, I didn't read that, but so I guessed from the code.

> Even ARP for Token Ring isn't that different from ARP over Ethernet.

Hm. :-(

> No luck there.  IPv6 doesn't use ARP.  (it has its way of doing link
> layer address resolution).

Hm. :-(

So I'm out of luck.

Regards,
	Ignatios Souvatzis