| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2213.1 | Anybody care to answer? | CMOTEC::CHOI | Shaking and Moving | Fri May 12 1995 14:07 | 26 |
| Hi,
Nobody seems to have answered the basenote - perhaps I'll be luckier!
My customer has a similar problem. He has a few separate ethernet LANs
which he wants to join via an FDDI ring. However, the LANs have a
variety of Ethernet traffic on them and he is concerned to know whether
the various frame types will be translated to and from FDDI correctly.
His Ethernet frame types are:
- Ethernet Type II
- 802.2 + 802.3
- 802.2 + snap
- raw 802.3 (IPX)
Can someone give me details as to how each of these frame types are
translated to FDDI, and how those frame types are translated back to
Ethernet?
Bottom-line: Will the translation work correctly
Ethernet -> FDDI -> Ethernet?
Thanks,
Clinton
|
| 2213.2 | | NETCAD::B_CRONIN | | Fri May 12 1995 15:36 | 35 |
|
Bottom line:
As long as there is no mixing of raw IPX with other IPX
datalink types on the same bridge it will work correctly.
Towards the FDDI, the bridges will do the following:
The bridges will snap encapsulate the enet type 2 traffic with an OUI
of 00-00-00. The 802.2+802.3 traffic will go through as 802.2 traffic,
the SNAP traffic will go through SNAP encapsulated with the OUI in the
SNAP. Raw 802.3 will either go through looking like 802.2 with SAPs
of FF, or, if the switch is on, it will be SNAP encapsulated with a
type of 81-37.
Coming off the FDDI,
The bridges will turn the snap encapsulated enet type 2 traffic
into enet frames. The 802.2+802.3 traffic will be turned back into
802.2+802.3 traffic. The SNAP traffic will go through SNAP encapsulated
with the OUI seen in the SNAP. Raw 802.3 seen on the FDDI (i.e. sap =FF)
will go through as RAW 802.3. Raw 802.3 that was SNAP encapsulated on the
FDDI will be turned back into raw 802.3 if the switch is on, or
forwarded as enet type 2 with with a type of 81-37 if the switch is
off.
|
| 2213.3 | One more thing! | CMOTEC::CHOI | Shaking and Moving | Tue May 23 1995 08:15 | 29 |
| Hello again.
Thanks for the full reply.
I've had a chance to think about this and would like one further
clarification.
Assuming the switch to convert 802.3 raw is on, how does the
FDDI->Ethernet bridge decide to convert a snap FDDI frame?
It can obviously convert to one of:
- 802.3 raw or
- 802.2+snap or
- Enet Type II.
Now, I assume if it sees an OUI of 00-00-00, it knows to convert back to
Enet type II, but how does it distinguish between the other two cases?
Specifically, how does it tell between IPX traffic that was originally
raw 802.3 and IPX traffic that was (802.2 + snap)?
I would assume that the type for both these cases will be 81-37. If
so, will it decide based on the OUI? If so, what are the values of the
OUI?
Sorry to be picky, but the customer just wants to be absolutely sure...
Clinton
|
| 2213.4 | | NETCAD::B_CRONIN | | Tue May 23 1995 12:13 | 16 |
|
The bridge makes an either/or decision on what to do when it sees
the SNAP with an 81-37 protocol type. It assumes that all stations
connected to this bridge want to talk in either raw 802.3 or ethernet
format, but not both.
If the CSMA/CD segments are running raw 802.3 format, the switch will
be set for raw 802.3 mode enabled, and all packets on the FDDI will be
converted to raw 802.3. If the switch is set to disabled, all of them
will be converted to Ethernet frame format.
If I understand your question correctly, I'd also add that there is no
need for the receiving station to know what format was used
by the originating station. In fact, this allows a limited amount of
mix and match to occur, as long as the mix and match is per bridge,
not per segment.
|