Re: [secdir] [Roll] Secdir early review of draft-ietf-roll-rnfd-02

[Konrad Iwanicki <[e-mail]>]

Dear Chris,

Thank you a lot for your feedback! Please, find my replies inline. 
Whenever I talk about changes or corrections made in the draft, I am 
referring to its newly submitted version: 03.

On 02/12/2023 22:05, Chris Lonvick via Datatracker wrote:
> Section 1.1 discusses what happens when an LBR crashes. In some places, the
> section uses the term "dead LBR". However, in other places, it just uses "LBR"
> and I can't always tell if it's referring to a live or dead LBR. It would help
> to clarify if in each case, it would be designated that the LBR is live, or
> dead.

Good suggestion. Done.

> Section 5.5 uses the phrase "sufficiently many messages". As I said, I'm not
> familiar with RPL so I don't know how many would be sufficient. Can this be
> better quantified?

It is hard to quantify this. In theory, a single message would be 
sufficient. However, the message may be lost. Different implementations 
and deployments may have different loss characteristics, so the draft 
deliberately leaves this issue open. In any case, even if an 
insufficient number of messages is sent and some node remains 
uninformed, solutions mentioned in the following sentences will address 
this issue. Finally, even if none of these solutions is implemented, 
nothing catastrophic happens: the uninformed node will simply transmit 
data as if RNFD were active. Last but not least, the DODAG root may 
always initiate a new DODAG Version, which by default deactivates the 
protocol and will eventually reach the uninformed nodes as long as they 
are connected. All in all, I would leave the sentence as is.

> Also, s/acros/across

Corrected.

Once again, thanks for the feedback.

Best,
-- 
- Konrad Iwanicki.