Re: [Roll] Mirja Kühlewind's Discuss on draft-ietf-roll-efficient-npdao-12: (with DISCUSS and COMMENT)

Rahul Arvind Jadhav <rahul.jadhav@huawei.com> Thu, 27 June 2019 01:00 UTC

From: Rahul Arvind Jadhav <rahul.jadhav@huawei.com>
To: Mirja Kühlewind <ietf@kuehlewind.net>, The IESG <iesg@ietf.org>
CC: "draft-ietf-roll-efficient-npdao@ietf.org" <draft-ietf-roll-efficient-npdao@ietf.org>, Peter Van der Stok <consultancy@vanderstok.org>, "aretana.ietf@gmail.com" <aretana.ietf@gmail.com>, "roll-chairs@ietf.org" <roll-chairs@ietf.org>, "roll@ietf.org" <roll@ietf.org>
Thread-Topic: Mirja Kühlewind's Discuss on draft-ietf-roll-efficient-npdao-12: (with DISCUSS and COMMENT)
Thread-Index: AQHVK3mEroguwqI5VE6sJxU/AMt1W6at9xOw
Date: Thu, 27 Jun 2019 01:00:04 +0000
Message-ID: <982B626E107E334DBE601D979F31785C5DF0BF90@BLREML503-MBX.china.huawei.com>
References: <156148288899.31109.1545890009202623647.idtracker@ietfa.amsl.com>
In-Reply-To: <156148288899.31109.1545890009202623647.idtracker@ietfa.amsl.com>
Accept-Language: en-IN, zh-CN, en-US
Content-Language: en-US
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
MIME-Version: 1.0
Archived-At: <https://mailarchive.ietf.org/arch/msg/roll/4OFOa4cGouOGmV-WNigWjivfwpE>
Subject: Re: [Roll] Mirja Kühlewind's Discuss on draft-ietf-roll-efficient-npdao-12: (with DISCUSS and COMMENT)
Precedence: list

Thanks Mirja for the review and feedback.
For the DISCUSS point, I have to reach out to WG to check if anyone has specific recommendations towards this. There is also a similar messaging in parent document RFC 6550 which also does not explicitly puts how this retry should happen.
Please find my responses inline.

Thanks,
Rahul

----------------------------------------------------------------------
DISCUSS:
----------------------------------------------------------------------

I have a small discuss that should be easy to address:

Sec 4.3: " The number of retries are implementation and deployment
dependent."
(and also sec 4.4 point 6)

Please specify a maximum number of retries and also a minimum retry interval (of e.g. 3 sec best with exponential back-off)!

[RJ] Section 4.3 says that it is implementation and deployment dependent. We have avoided to put an exact number here because of the differences in physical layers and wireless dense/sparse deployments this could be used on. I am not sure if putting an exact value like specifying 3 retries or 4 retries is ok. I don't know what to base this value/recommendation on. The DCO/DCO-ACK retry mechanism emulates DAO/DAO-ACK retry mechanism from RFC 6550 which also has not put an explicit details for this.

----------------------------------------------------------------------
COMMENT:
----------------------------------------------------------------------

One question on section 4.6.2: You present use of NPDAO and DCO as two options, however, the problem with the I flag is that the sender does not know if the ancestor understand the signal. Wouldn't it also make sense to use both in some cases, e.g. send DAO with I flag first and if you don't receive a DCO after some limited time, you also send the NPDAO?

[RJ] Yes, it is possible for a node to use both DCO and NPDAO. Section 4.6.2 specifies this in detail.

Nits:
sec 1.2: s/so that the node changing it routing adjacencies/so that the node changing its routing adjacencies/ -> "it" instead of "its"

[RJ] Will fix this.

[Roll] Mirja Kühlewind's Discuss on draft-ietf-ro… Mirja Kühlewind via Datatracker
Re: [Roll] Mirja Kühlewind's Discuss on draft-iet… Rahul Arvind Jadhav