Re: [6lo] Generation of IPv6 IIDs

[Fernando Gont <[e-mail]>]

On 07/24/2014 01:17 PM, Kerry Lynn wrote:
>     > I think we should assume that those documents will be updated in
>     the future to provide the option to work with other IIDs even if
>     that comes with lower efficiency.
> 
>     Note that RFC 6282 already provides the option:
>     - If the IID can be derived from MAC address, great.
>     - If not, things will still work just with reduced efficiency.
> 
> To recap, re-using the L2 addresses for IIDs contributes to the
> efficiency of
> IPHC [RFC 6282].  Using locally-assigned short MACs provides the greatest
> efficiency.

I'd personally say that sticking to MAC-derived IIDs is not a good idea.
It's okay if that's still employed before something else is engineered,
though.



> Requiring randomly assigned IIDs can have a significant impact on 6lo
> networks
> with short frames, which already have reduced payloads due to security
> and other
> headers.  In some cases, random IIDs might necessitate an additional
> frame per
> message.
>  
> 
>     Section 10 of RFC 4944 also provides the same option, but has since
>     been replaced by the encoding specified in RFC 6282.
> 
>     Also note that 802.15.4 interfaces have both a stable MAC address
>     (EUI-64) and a locally-significant, dynamic MAC address (Short
>     Address).  Many of the arguments against MAC-address-based IIDs may
>     not apply when using Short Addresses.
> 
> I'm not sure that any of the arguments do apply when using locally assigned
> short MAC addresses.  Please note that draft-ietf-6lo-6lobac *only* uses
> locally
> assigned short MAC addresses.

If you generate your IIDs from a constant IID, they do apply. Please see
e.g.
<http://tools.ietf.org/id/draft-ietf-6man-ipv6-address-generation-privacy-01.txt>
for a longer version of the story.

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492