Re: [v6ops] IPv6 fragmentation experience

On 03/21/2018 01:10 PM, Joe Touch wrote:
>
>
>> On Mar 21, 2018, at 2:41 AM, Ole Troan <otroan@employees.org 
>> <mailto:otroan@employees.org>> wrote:
>>
>> For IPv4 (and only IPv4) I do think we have a strong case against 
>> fragmentation.
>> Routing on the transport layer ports is now a fundamental part of the 
>> IPv4 Internet architecture.
>> If the layer 4 information isn't there, then the packet cannot be routed.
>
> DPI into the packet body is also a fundamental part too.

Not generally at scale, and not with many of the new and upcoming 
protocols.
>
> However, if the lack of layer 4 info creates a situation where a 
> packet can’t be routed, then the device isn’t a router.

I don't understand this. I understand you mean the subsequent fragments, 
except that routers as I understand them perform "network layer 
forwarding" [RFC1812, Requirements for IP Version 4 Routers]
Layer 4 is not the network layer; are devices that need L4 info 
considered routers?
>
> Yes, the Internet has many many broken parts. That includes RFCs that 
> are inconsistent with our core architecture requirements.
>
> If anyone in the IAB is listening, you’d earn your middle name better 
> by dealing with those than trying to seek new “fun” stuff on which to 
> pontificate.

As I take off several hats, I'm not sure what to ask of the IAB here.

Is it to be the "compliance arm" you mentioned elsewhere in the thread?

Lee

>
> Joe
>
>
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops