Hi Micheal,
>Probably, just don't be so specific about the values.
>I think that "object-value" can be any amount of CDDL that want to write.
>You could use a map for "service-identification"
>Since it's all encoded in CBOR, you can't have a 3-bit value, only a 1-byte
>value.
A New GRASP Objective Option as an example is defined as follows, according to
draft-ietf-anima-asa-guidelines-01.
- name: s-deployment-value, which is need to be assigned by IANA registration
- objective-flag: F_DISC and F_SYNCH
- objective-value: a data structure as follows
-- s-identification, which is used to distinguish different network services
-- s-deployment-type, which indicates that it is new, or updated, or deleted
-- s-resource-information, which are the assured resources of the network
service.
-- resource-type, which indicate bandwidth, or latency, etc.
-- resource-value, which indicate the value.
-- s-deployment-status, which indicate deployment success or failure, and the
reason for the failure
Do you think it is available?
Best wishes,
Joanna
-----Original Message-----
From: Anima [mailto:[email protected]] On Behalf Of Dangjuanna
Sent: 2021年7月12日 22:54
To: Michael Richardson <[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: [Anima] New Version Notification for
draft-dang-anima-network-service-auto-deployment-00.txt
Hi Michael,
Please see the below.
Best wishes,
Joanna
-----Original Message-----
From: Michael Richardson [mailto:[email protected]]
Sent: Friday, July 9, 2021 12:36 AM
To: Dangjuanna <[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: [Anima] New Version Notification for
draft-dang-anima-network-service-auto-deployment-00.txt
Dangjuanna <[email protected]> wrote:
mcr> 1) do the ANS message flow over the ACP? I assume so.
JD> Correct. ACP of the autonomic node send or receive the ANS message. I
JD> was a little hesitant to use the SPE and UPE names when writing, so
JD> the text was a little confusing. I suggest that the SPE and UPE be
JD> unified as an SPE in the future because it is a single device on the
JD> network. Do you think it is OK?
I don't know if you should combine SPE/UPE/PE. I don't know enough about this
area of traffic engineering.
mcr> I think you want to do an IANA Action here to define your own?
mcr> EX10 is just... an example.
JD> Yes, the new objective-name for network service deployment need to be
JD> defined and added. EX10 is only one example :)
I suggest you pick a name, and put it into the IANA considerations.
[JD] I fully agree with you.
> path-identification-value, resource-type, same comment.
JD> Could you help me modify them according to your understanding?
Probably, just don't be so specific about the values.
I think that "object-value" can be any amount of CDDL that want to write.
You could use a map for "service-identification"
Since it's all encoded in CBOR, you can't have a 3-bit value, only a 1-byte
value.
[JD] I need to learn this definition. Hopefully, I'll understand in a few days.
> You probably mean to say that it uses the security provided by the ACP.
JD> Yes, do you think whether it is enough?
Yes, I think it is, but you need to say so, and explain what threads there are,
and why the ACP provides an acceptable solution.
[JD] I thought it was relatively simple. This scenario requires the GRASP
protocol. The security risks are mainly in the channels between autonomic
nodes.
> Then the ANS communication would occur with that discovered node.
> It would, as you say, negotiate the path.
JD> Correct, this process is necessary in the entire program. But in the
JD> network service auto-deployment scenario, this discovery mechanism
JD> has not been changed, so I did not emphasize it in the text. Where do
JD> you think it is more appropriate to add it?
I think that you need to explain more of it.
[JD] I try to write more about it.
> I don't know [I-D.ietf-spring-segment-routing] or [I-D.ietf-mpls] that
> well, but I didn't think that they included on-path (RSVP-like)
> discovery of route.
> Probably, I've missed something.
JD> The routing protocol can advertise routes and calculate the optimal
JD> route, which I understand should be regarded as a route discovery
JD> process. I also would like to hear more of your thoughts on this
JD> point.
I don't have the background in traffic engineering to know.
--
Michael Richardson <[email protected]> . o O ( IPv6 IøT consulting )
Sandelman Software Works Inc, Ottawa and Worldwide
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
