rpk@mit-eddie.UUCP (Robert Krajewski) (05/08/84)
Today I attended a seminar by David Clark about the network environment at MIT's Laboratory for Computer Science. One of the points made was that although Chaos and TCP/IP may have different camps, it's a relief to know that you can run both protocols on the same piece of Ethernet cable. And then Jerry Saltzer drops the bomb and says that the new IEEE802 standard that started off with Ethernet has replaced the protocol type field in the Ethernet frame with a length field ! My question is: How can anybody be so brain-damaged ???? -- ``Bob'' (Robert P. Krajewski) ARPA: RpK@MC MIT Local: RpK@OZ UUCP: genradbo!miteddie!rpk or genradbo!miteddie!mitvax!rpk
rpw3@fortune.UUCP (05/12/84)
#R:mit-eddi:-177900:fortune:5900020:000:920
fortune!rpw3 May 11 18:30:00 1984
I posted a fairly long article about this 802.3 vs. Ethernet thing a
while back, but the key is that none of the (currently) Xerox-registered
Ethernet frame types (such as XNS or IP/TCP) are legal values for the
length field. So, 802.3-only boards won't receive data from Ethernet-only
software, but they can share the cable (if you don't mind each XNS or IP
packet potentially bumping an "invalid length field" error count in the
802.3 system). But sophisticated Ethernet systems will be able to talk
to 802.3 systems (both ways) by treating each possible value of the 802.3
length field (46 to 1518) as a valid sub-type of the 802.3 "type".
Ugly, but then again sometimes it's good to remember:
"Politics is the art of the possible."
;-}
Rob Warnock
UUCP: {ihnp4,ucbvax!amd70,hpda,harpo,sri-unix,allegra}!fortune!rpw3
DDD: (415)595-8444
USPS: Fortune Systems Corp, 101 Twin Dolphin Drive, Redwood City, CA 94065mark@cbosgd.UUCP (Mark Horton) (05/14/84)
Pretty much everyone I've talked to so far who has a product out is going with Ethernet and ignoring the IEEE 802.3 standard. Does anyone know of vendor products available that support 802.3 instead of Ethernet?