Re: [BEHAVE] UDP zero checksums and v4 to v6 translators

[Iljitsch van Beijnum <iljitsch at muada.com>]

On 4 aug 2009, at 15:27, Rémi Després wrote:

> > That's pointless, because the IPv6 spec, against which  
> > implementations have been heavily tested, reject such packets.

> You seem to have missed that the proposal includes a relaxation of  
> the constraint that zero-checksum UDP datagrams MAY be accepted by  
> hosts ion the future, just to avoid unnecessary black holes in case  
> of v4 to v6 translations.

Although actual IPv6 packets flowing over the internet aren't as  
numerous as their IPv4 counterparts and at least "legacy" IPv6 stuff  
can't be as old as the oldest IPv4 stuff, we still have a very large  
installed IPv6 base out there that isn't going to adapt to modifying  
standards at the drop of a hat.

If you send UDP packets with 0 checksums to IPv6 hosts, expect many of  
them to reject those packets for a very long time, at the very least 5  
years.

> Is progress from what is implemented today a taboo?
> (I hope not.)

It's not taboo, just very hard. Apple and Microsoft aren't  
particularly quick with building in updates to the network stack, and  
after they've done that, it can take a long time for holdouts to  
upgrade. This is much harder than updating applications from competing  
vendors running in user space.

But nobody is stopping you from coming up with a new UDP-like  
encapsulation that doesn't have a checksum.