Re: Running code (Was: I-D Action: draft-ietf-6man-ipv6only-flag-03.txt)

[Mark Andrews <[e-mail]>]

> On 25 Oct 2018, at 2:04 pm, Manfredi (US), Albert E <[e-mail]> wrote:
> 
> -----Original Message-----
> From: ipv6 <[e-mail]> On Behalf Of Mark Andrews
> 
>> Yes.  The assumption when the flag is set is that all the boxes are IPv6
> capable.
> 
> Okay, you are speaking as the network admin.
> 
>> If I was implementing it the interface would become un-numbered for IPv4,
> the DHCP client would be disabled, mDNS would not be attempted over IPv4.
> 
> And, as a network admin, wanting to stop all IPv4 traffic for whatever network admin reason, it is feasible tout de suite, today, to simply filter Ethertypes 0x0806 and 0x0800, in your switches and router Ethernet interfaces. No need to wait for the IETF.
> 
> And similarly, from an equipment vendor's perspective, say your goal as smartphone maker is to extend battery life, make it your own responsibility to do what's best for your user. Don’t expect the network admin to care. Probe the network, see if IPv6 is sufficient for your user's needs, and if yes, then no op the IPv4 stack. The network admin couldn't care less, potentially. Why count on him/her?
> 
> So in short, why do we actually need any flag in the IP header?

It isn’t a flag in the IP header.

Not all equipment can filter on Ethertypes.

Heuristics are not enough.  I know of no machine that can say “There are no IPv4 only devices are expected to be connected to this network in the future” based on heuristics.  Just trying to work out if you can connect to every device on the network over both IPv4 and IPv6 for every service they offer is a impossible task which is where you would need to be to decide to automatically shutdown IPv4 link-local.

Mark

> Bert
> 

-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: [phone]              INTERNET: [e-mail]