Re: [Roll] Stephen Farrell's Discuss on draft-ietf-roll-applicability-home-building-11: (with DISCUSS and COMMENT)

[Stephen Farrell <[e-mail]>]

Hiya,

On 15/07/15 10:23, Robert Cragie wrote:
> On 14 Jul 2015 23:36, "Stephen Farrell" <[e-mail]> wrote:
> 
> Hi Stephen,
> 
> The text before 4.1.8.2 doesn't imply that some nodes are not using
> security, it only implies that symmetric key scope used by nodes may be
> less than the whole network. 

Well, I'm still not getting it sorry;-) Unless you mean that
the n/w might accept/use >1 group symmetric key? If that's it
then just saying it that way would be good.

> I will add clarification regarding this point.

Grand - I'm sure clarification is all that's needed.

Cheers,
S.

> 
> Robert
> 
>>
>> Hiya,
>>
>> On 14/07/15 23:17, Robert Cragie wrote:
>>> Hi Stephen,
>>>
>>> Thanks for your further review. Answers and comments inline, bracketed by
>>> <RCC></RCC>
>>>
>>> Robert
>>>
>>> 1) This could be my ignorance of zigbee, but how can we
>>>> use layer 2 security for only some network nodes?  (In
>>>> other words, I don't see how 4.1.8.2 works.)
>>>>
>>>
>>> <RCC>All network nodes use L2 security once they have joined the network.
>>> Prior to that, they communicate using an authentication protocol which is
>>> unsecured at L2. Enforcement points police unsecured traffic to ensure it
>>> is only related to authentication, thus preventing data or other control
>>> plane traffic unsecured at L2 from being allowed into the network. So, to
>>> your point, the only traffic allowed unsecured at L2 is authentication
>>> traffic and those nodes are not yet participating in the network.</RCC>
>>
>> Right, that's what I thought. So I can't see how the text that
>> implies that only a subset of nodes on the network are using l2
>> security works which seems to be implied by the bullet just
>> before the start of 4.1.8.2.
>>
>> [snip]
> 
> 
>>
>>
>