Templin, Fred L wrote:
Some of the terms related to Ethernet should be clarified, I believe.
The IEEE 802.3 standard does not recognize the terms 'jumbo' or
'superjumbo'. These are as you point out of the standard. The max
frame
size in the original Ethernet standard was 1518. This was slightly
increased about a decade ago to 1522 when IEEE 802.1 added four bytes
for VLAN ID. More recently the IEEE 802.3as extended the max size of
the
Ethernet frame to 2048.
The 1518 would be (14 header + 1500 ULP + 4 trailing CRC).
Focusing only on the ULP, can we say that 1500 bytes is
the nominal MTU of today's Internet, and therefore all
attached devices SHOULD configure a minMRU of 1500? Then,
can we also say that applications that send packets larger
than 1500 bytes (with DF=1) are RECOMMENDED to use RFC4821?
Finally, could we also say that tunnel decapsulators SHOULD
configure a minMRU of 2048 to account for encapsulations
that might extend a 1500 byte ULP out to the IEEE 802.as
maximum frame size?
Asked another way, can all of this be addressed in a brief
BCP that captures something like the above?
This is the advice from RFC4821 (not in standards language):
Since protocols that do not implement PLPMTUD are still subject to
problems due to ICMP black holes, it may be desirable to limit to
these protocols to "safe" MTUs likely to work on any path (e.g., 1280
bytes).
I think it would be unrealistic to say tunnels SHOULD have a MRU of at
least 2048 given the preponderance of gear that only supports 1500
bytes. However, it's relatively easy for non-4821-capable protocols to
limit sizes to something a little smaller that will provide adequate
headroom for tunnel headers.
-John
_______________________________________________
Int-area mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/int-area
