Tunnels might not support 1500B source MTUs.
So using 1500B IP MTU works only where it works and silently fails elsewhere,
e.g., if the receive-side reassembly MTU can’t fit both the packet being sent
and the tunnel overhead for a single tunnel.
I.e., if you send 1500B and go over a tunnel, even an IPv6 one, the tunneled
packet will have a size larger than 1500B. Even IPv6 does not *require*
receivers to reassemble packets larger than 1500B, so it might not work.
Had the system stuck with 1280, reassembly would be guaranteed.
I DO agree that we should make MTUs a bit bigger, but there are two hazards:
1. We really can’t go over 1500B until Ethernet does everywhere
2. Whatever we pick, there has to be room for tunnel headers somewhere
Just “pegging the meter’ is asking for failure where tunnels are concerned.
Joe
—
Dr. Joe Touch, temporal epistemologist
www.strayalpha.com
> On Aug 7, 2025, at 3:06 PM, Paul Vixie <[email protected]> wrote:
>
> Many large scale DNS operators assume 1500 without regard to v4 vs V6. Works
> fine. The law is wrong and always was wrong and should be updated.
> Paul Vixie
> Aug 7, 2025 23:01:16 [email protected]:
>
> Additionally:
>
>> On Aug 7, 2025, at 1:15 PM, Templin (US), Fred L
>> <[email protected]> wrote:
>>
>> Many tunneling protocols live by “grace” and assume 1500 everywhere. But,
>> robust
>>
>> tunneling protocols need to live by the “law”, and the law says 1280.
>
> And its subtle - for IPv4, they assume 1500 everywhere but it’s really 576
> after reassembly at the receiver.
>
> For IPv6, it’s 1280 over each hop BUT up to 1500 after reassembly at each
> receiver.
>
> These and other aspects are why this isn’t just an op-ed.
>
> Joe
>
_______________________________________________
Int-area mailing list -- [email protected]
To unsubscribe send an email to [email protected]
