Hi Ian,

Many thanks for the prompt review. Much appreciated!

Please see inline.

Cheers,
Med

De : Ian Farrer [mailto:[email protected]]
Envoyé : mardi 8 novembre 2016 12:24
À : BOUCADAIR Mohamed IMT/OLN
Cc : softwires; [email protected]
Objet : Re: [Softwires] Volunteers to review draft-ietf-softwire-dslite-yang?

Hi Med,

Thanks for the reminder. I’ve been meaning to review this for a while.

Cheers,
Ian


General Comments:
1, How does this draft relate to the model in draft-sivakumar-yang-nat? There 
seems to be a lot of redundancy between them, which is understandable given the 
respective functions of a CGN and an AFTR. Wouldn’t it be better to reference 
the ietf-nat model for the nat specific functionality?

[Med] This is fair point, but given the current progress of 
draft-sivakumar-yang-nat I prefer to avoid that dependency and build the 
DS-Lite as self-contained modules.

Section 1
2, It would be useful to include a YANG model for B4 configuration as well to 
be inline with what is covered by the softwire-yang draft. Operators that are 
planning on moving to configuring CPEs with Netconf would also find this useful 
and given that there is a  fairly small amount of confiig necessary for the B4, 
it would be neater to include this here rather than another draft.

[Med] The B4 module will be quite simple. I don’t have objection to include 
that module, but this will require to change the scope of the draft.


Section 2
3, "A device can enable multiple AFTR instances”
I’m only familiar with one vendor’s DSLite AFTR implementation and I believe 
this only supports a single instance. Suggest something like - "The model 
supports enabling one or more instances of the AFTR function on a device”
[Med] OK with this change. FWIW, there are vendors who support the activation 
of multiple instances on the same AFTR.


4, The tree model needs a clean up for the data type column positioning
[Med] This will be fixed.


Model comments:
5, aftr-parameters/ipv4-address - You could use 192.0.0.1 as a default value 
here.
[Med] Agree.

6, aftr-parameters/leaf-mss-clamping-enable - Does this need an option to 
configure the MSS size to be used for clamping?
[Med] thank you for catching this. When it clamping is enabled, a value should 
be provided.

7, aftr-parameters/max-softwire-per-subscriber - the description could 
reference the subscriber-mask leaf to clear up how a ‘subscriber’ is defined.
[Med] Makes sense.

8, Should it be possible to configure an IPv6 range(s) for clients that the 
AFTR instance will service? If so, then the subscriber mask should be a child 
of this object.
[Med] The base AFTR behavior does not require explicit configuration of the 
IPv6 prefix to service. I’m hesitating to add that.

9, container logging-info - Can this be extended to contain timestamping 
information for the start/endtime of an active translation? I’m aware of a 
large commercial CGN deployment which requires these timestamps for regulatory 
compliance.
[Med] I agree that timestamping is needed, but this model assumes that the 
structure of the logging entries are fixed. Because there is no recommendation 
about such entries (except a discussion in RFC6888 and RFC6269), we avoided to 
zoom more on that part.

9, Also, I’ve seen several different logging/ protocol options that are 
implemented by vendors (local, syslog, netflow/ipfix, proprietary). Should a 
way of configuring these be available in this model?
[Med] May make sense to be added.

11, mapping-entry/lifetime - what is the unit for this leaf?
[Med] Seconds ; the module will be updated.

12, statistics/traffic-statistics - are there any useful sent ICMP set error 
messages that could be included here for e.g. to identify MTU problems?
[Med] Will be happy to align with lw4o6/MAP-E.

13, The notifications section seems quite light given then number of 
configurable threshold values that are available. e.g. 
max-softwire-per-subscriber (would be useful to know how often this is 
exceeded), no. of exceeded port-quota or port-set-size.
[Med] This section focuses only on critical events.

What about a counter for drops of packets received with unsupported transport 
protocols?
[Med] We can but this may be just logged locally.


Grammatical comments:
Section 2:
s/each responsible to service a group of B4s/each responsible for serving a 
group of B4s/
s/with a dedicated configuration data/with dedicated configuration data/

leaf hold-down-timeout
s/not reassigned till this timer/not reassigned until this timer/

[Med] Fixed. Thanks

On 08 Nov 2016, at 10:04, 
<[email protected]<mailto:[email protected]>> 
<[email protected]<mailto:[email protected]>> wrote:

Dear WG,

We are seeking for volunteers to review draft-ietf-softwire-dslite-yang.

If you are willing to review, please send a note to the list.

Thank you.

Cheers,
Med
_______________________________________________
Softwires mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/softwires

_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to