Re: Next steps on advancing core IPv6 specifications to full Internet standard

[Hemant Singh <[e-mail]>]

Got it - thanks, Tom.

Did you see my other email related to both the implementations for SR and
Inband OAM with mixed nodes?  How do nodes know where to add the OAM data
or parse the SR encoded data if one vendor's node uses add EH while another
vendor's node uses encapsulation?    If both implementations continue to
live, we need means to signal to the network what mode is being used for
certain protocols.   Some means I see are a new DHCPv6 option, a new
option/bit in the IPv6 ND RA, a new next_hdr in the IPv6 header, or use
some bits from the Traffic Class and Flow Label in the IPv6 header.

Thanks,

Hemant


On Fri, Nov 18, 2016 at 3:45 AM, Tom Herbert <[e-mail]> wrote:

>
>
>
>> The implementation doesn't distinguish between being a host and being a
> router in this case. If the forwarding instructions are to insert a header
> then that is what's done. Encapsulation is just an alternate method that
> should already be supported and similarly would work for a host or router.
> Basically, we are leaving it up to the user rather inserting EH in the
> network is reasonable, however a warning about inserting EH based on the
> proposed text will be added to the documentation.
>
> Thanks,
> Tom
>