Path: utzoo!utgpu!watserv1!watmath!att!rutgers!aramis.rutgers.edu!athos.rutgers.edu!hedrick From: hedrick@athos.rutgers.edu (Charles Hedrick) Newsgroups: comp.sys.hp Subject: Re: HP TCP/IP router/bridge? Keywords: TCP/IP HP router bridge Message-ID: Date: 17 Oct 90 22:27:33 GMT References: <45306@apple.Apple.COM> <8445@fmeed1.UUCP> Distribution: na Organization: Rutgers Univ., New Brunswick, N.J. Lines: 21 The problem with HP's 802.2 TCP/IP is that it doesn't use the standard 802.2 IP encapsulation. It uses their own homegrown one, which (among other things) uses HP's "probe" protocol instead of ARP. I have no idea why somebody would do a TCP/IP implementation that can't talk to any other TCP/IP implementation, but they did. (Supposedly this was fixed in later releases, though the code we are using on an HP/3000 still seems to be this way.) As far as I know, the only product that will translate between HP's braindamage and normal IP encapsulations is a cisco router. In general there are problems with mapping between IEEE and Ethernet encapsulations. E.g. Appletalk uses IEEE for phase II and Ethernet for phase I. A simple translation would cause Appletalk to get the two phases confused. I believe there are similar problems with other protocols. The net result is that bridges with IEEE/Ethernet translation need interesting options to make everything work. This makes supposedly transparent Ethernet/FDDI bridges a very dicey proposition. (Not that it can't be made to work for many useful cases, but that the obvious mapping causes problems.) But this isn't relevant to the HP case andway, since their IEEE encapsulation isn't standard. So a normal IEEE to Ethernet translation wouldn't help.