Re: [MBONED] [Bier] MSR6 BOF 3rd Issue Category: More details are requested about the large scale use cases, including issue 8-11

["Jeffrey (Zhaohui) Zhang" <[e-mail]>]

Hi,

A few of us put together an informational document on this scaling topic: https://datatracker.ietf.org/doc/draft-zzhang-pim-multicast-scaling-considerations/

It tries to show how existing and in-development technologies/solutions can be put together to address scenarios where all three characteristics exist.
It is not data plane specific at all.
Jeffrey



Juniper Business Use Only
From: BIER <[e-mail]> On Behalf Of Yisong Liu
Sent: Tuesday, October 18, 2022 10:22 PM
To: [e-mail]
Cc: [e-mail]; 'BIER WG' <[e-mail]>; [e-mail]; [e-mail]; 'Stig Venaas' <[e-mail]>; [e-mail]
Subject: [Bier] MSR6 BOF 3rd Issue Category: More details are requested about the large scale use cases, including issue 8-11

[External Email. Be cautious of content]

Hi all,

Here are the responses for the 3rd Issue Category: More details are requested about the large scale use cases, including issue 8-11.

Too large scale may request overlay multicast (To be clarified)<https://urldefense.com/v3/__https:/github.com/MSR6-community/MSR6-Issue-List/issues/8__;!!NEt6yMaO-gk!HPjdq8-__y5-vUsrYfBJFu-RqNbXNLDebw1U6v1aKCKoy-jnUCa81F5BD9nuIZ4-9MVJzoqxt1U2pnJiKedGhJA$> #8
 Surely you wouldn't put 10000 receivers in the source route. So you will need overlays.<https://urldefense.com/v3/__https:/github.com/MSR6-community/MSR6-Issue-List/issues/10__;!!NEt6yMaO-gk!HPjdq8-__y5-vUsrYfBJFu-RqNbXNLDebw1U6v1aKCKoy-jnUCa81F5BD9nuIZ4-9MVJzoqxt1U2pnJizUBRFV4$> #10
[Response]Based on the presentation in IETF 114, “Large scale”, “Host-initiated” and “Overlay multicast” are three separate cases. MSR6 is not intended to address scenarios in which all three characteristics exist simultaneously; “Large scale” is for the case of underlay network, for example 5G transport network or the data center network, not an overlay network. “Large scale” also does not describe the number of receivers (e.g, 10k receivers), but rather the size of the network, i.e., the number of network devices.
 What the size of tree or number of flows？How many hosts in a flow? Sparse or dense tree?<https://urldefense.com/v3/__https:/github.com/MSR6-community/MSR6-Issue-List/issues/9__;!!NEt6yMaO-gk!HPjdq8-__y5-vUsrYfBJFu-RqNbXNLDebw1U6v1aKCKoy-jnUCa81F5BD9nuIZ4-9MVJzoqxt1U2pnJiPoU1ido$> #9
[Response]In our consideration, there are three dimensions of multicast scalability: number of services, multicast tree size, network size; all three exist trade-offs. For example, if MSR6 TE encodes the multicast tree into the packet, it could be decoupled from the number of services and network size, but the header overhead is affected by the size of the multicast tree. So it is suitable for large network & small tree & multiple stream scenarios (sparse mode). Large scale doesn’t refer the number of hosts as these are independent cases as mentioned.

Should not count hosts. Tree starts at cell-side router. This problem has been solved for centuries. Major providers have multicast for these emergency message problems.<https://urldefense.com/v3/__https:/github.com/MSR6-community/MSR6-Issue-List/issues/11__;!!NEt6yMaO-gk!HPjdq8-__y5-vUsrYfBJFu-RqNbXNLDebw1U6v1aKCKoy-jnUCa81F5BD9nuIZ4-9MVJzoqxt1U2pnJiD2t4Q6c$> #11
[Response]
1."Should not count hosts. tree starts at cell-side router." It is right as it is mentioned above: Large scale doesn’t refer the number of hosts .
2.With IPv6 and SRv6 already deployed, we think MSR6 is the appropriate transport network solution, compared to existing solutions
3."Major providers have multicast for these emergency message problems" is not very clear. Is there a mistake in the record or can the issue be further clarified?

If you have further comments, please let us know.

Best Regards
Yisong Liu