Re: [Dime] AD Evaluation of draft-ietf-dime-doic-rate-control-10

Ben,

I've updated the document based on our comments.  See more below.

Steve

On 12/21/18 5:06 PM, Ben Campbell wrote:
> Hi,
>
> This is my AD evaluation for draft-ietf-dime-doic-rate-control-10. I previously reviewed version 8, however since some time has passed I reviewed this version “from scratch”.
>
> In general the draft is in good shape. I think it’s ready for IETF Last Call, which I will request shortly. Please note the last call window will be extended due to the upcoming holidays.
>
> I have a few minor comments that can be resolved along with any last call feedback.
>
> Thanks!
>
> Ben.
>
> -------------------------------------
>
> §4, paragraphs 2 and 3: Am I correct to assume that, as new DOIC algorithms get added, nodes could support both of these and something else? If so, then in paragraph 2 I suggest s/ “ support both the loss and rate based abatement algorithms”/ "support at least the loss and rate based abatement algorithms”
SRD> No, only loss is required to be supported.  The statement is that,
because loss is always required, supporting rate implies supporting loss
and rate.  I don't think a change is required here.
>
> .... and in paragraph 3, I suggest adding something to the effect of “... and MAY indicate support for others.”
SRD> I agree this is a good change.
>
> (nit) §5.5, 2nd paragraph: "It is also possible for the reporting node to send overload
> reports with the rate algorithm indicated when the reporting node
> is not in an overloaded state.”
>
> I suggest s/ “indicated when” / “indicated even when”
SRD> Okay.
>
> (nit) §5.6, first paragraph: The algorithm is detailed in 7.3.

>
> §7.3.1: "To apply abatement treatment to new Diameter requests at the rate
> specified in the OC-Maximum-Rate AVP value sent by the reporting node
> to its reacting nodes, the reacting node MAY use the proposed default
> algorithm for rate-based control or any other equivalent algorithm
> that forward messages in conformance with the upper bound of 1/T
> messages per second.”
>
> This is redundant to similar normative text in §5.6. I suggest keeping just one (probably this one since it’s more precise) and use descriptive language for the other.
SRD> Okay, I changed 5.6 to the following:

   When determining if abatement treatment should be applied to a
   request being sent to a reporting node that has selected the rate
   overload abatement algorithm, the reacting node can choose to
   use the algorithm detailed in Section 7.
>
> §9: Do the authors think that the rate algorithm might be more effective at DoS mitigation than the loss algorithm? If so, that might be worth a mention in the security considerations.
SRD> Good suggestion.  I've added the following paragraph to the
security section:

   In addition, the rate algorithm could be used to handle DoS attacks
more effectively than the loss algorithm.
>
>
>
>
> _______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime