Re: [Bier] BIER: draft-eckert-bier-cgm2-rbs-01 with performance analysis simulation

["Xubing (Robin)" <[e-mail]>]

>On Tue, Feb 15, 2022 at 09:59:40AM +0800, [e-mail] wrote:

>> Hi Toerless,

>> Thank you very much for your response!

>> I have one more question, let we consider CGM2/RBS solution only.

>> Since all the bits have local significance in CGM2/RBS,  what if we want to send one flow to multiple BFERs but we don't want to indicate any path?

>> In this case we can't reach the BFERs because we don't put any path bits in the BitString and there is no global BIFT in the intermediate nodes, right?

>

>Right. This is how we've introduced CGM2/RBS and it makes it very simple and IMHO elegant.

>And seemingly from the first simulations also quite competitive versus the "flat bitstring" BIER/TE approach in terms of replications required with same address size. More simulations would be great.



Let me introduce some detail about the simulation as follow:

We divide the overall topology into 160 domains, each of which includes 180 egress points, providing the total of 28000 egress points.

We randomly select egress points as group members, with the total number ranging from 10 to 28800. And total samples is 60.

In order to compare the encoding mechanism more fairly, we limit the size of the bitString to 256 bits (the typical size of BIER ) in both BIER and CGM2.



[image]


Conclusion:

  1.  BIER reaches its 160 packet replication limit at about 500 users, while CGM2 encoding reaching its limit of 125 replications at about 12000 users.
  2.  The decreasing of replications above12000 users about CGM2 is caused by the use of node-local broadcast as a further optimization.
  3.  CGM2 can completely break the 256-bit encapsulation limit, which cannot be easily broken by BIER, thus allowing the source to send fewer multicast streams.
  4.  CGM2 encoding performs is significantly better than BIER in sparse scenario.





Sincerely,



Bing Xu

Network Laboratory, 2012Labs, Huawei Technologies Co.,Ltd.

Tel : [phone] / Email : [e-mail]



This e-mail and its attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient(s) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!



-----Original Message-----
From: Toerless Eckert <[e-mail]>
Sent: Tuesday, February 15, 2022 9:42 PM
To: [e-mail]
Cc: [e-mail]; Xubing (Robin) <[e-mail]>; zzhang=[e-mail]
Subject: Re: [Bier] BIER: draft-eckert-bier-cgm2-rbs-01 with performance analysis simulation



On Tue, Feb 15, 2022 at 09:59:40AM +0800, [e-mail]<mailto:[e-mail]> wrote:

> Hi Toerless,

> Thank you very much for your response!

> I have one more question, let we consider CGM2/RBS solution only.

> Since all the bits have local significance in CGM2/RBS,  what if we want to send one flow to multiple BFERs but we don't want to indicate any path?

> In this case we can't reach the BFERs because we don't put any path bits in the BitString and there is no global BIFT in the intermediate nodes, right?



Right. This is how we've introduced CGM2/RBS and it makes it very simple and IMHO elegant.

And seemingly from the first simulations also quite competitive versus the "flat bitstring" BIER/TE approach in terms of replications required with same address size. More simulations would be great.



Cheers

    Toerless



> Best regards,

> Sandy

>

>

> ------------------原始邮件------------------

> 发件人：ToerlessEckert

> 收件人：张征00007940;

> 抄送人：[e-mail];[e-mail];zzhang=[e-mail].o<mailto:[e-mail];[e-mail];zzhang=[e-mail].o>

> rg;

> 日 期 ：2022年02月11日 00:51

> 主 题 ：Re: [Bier] BIER: draft-eckert-bier-cgm2-rbs-01 with performance

> analysis simulation On Thu, Feb 10, 2022 at 09:16:28AM +0800, [e-mail]<mailto:[e-mail]> wrote:

> > Hi Toerless,

> > Haven't read the new version of CGM2/RBS, just have the same question with Jeffrey.

> > If I understand BIER-TE and CGM2/RBS solutions right, the forwarding tables of the two methods are only focus on the local links.

> > So the table's scale for the two solutions is almost same.

> Jeffrey wasn't asking about BIER-TE (i think).

> > The commen issue of the two solution is the BSL limitation.

> > BIER-TE solution has an exclusive issue that the BP needs to be assigned globally.

> > But in CGM2/RBS solution, the BP needs not to be assigned globally.

> > So the main difference of the two solutions is: is it necessary to assign the BP globally, right?

> That is one core difference, yes. And it eliminates IMHO all the

> scaling challenges we have with BIER-TE (that we try to compensate for

> to some degree with controller based bit optimizations). But as

> discussed in the answer to Jeffs message, the fact that the encoding

> in CGM2/RBS is a tree means that we also eliminate the BIER/BIER-TE

> issue that a bitstring can carry a lot of unnecessary (zero) bits and

> need to make multiple copies just because some BFER or adjacencies have their bits in different SI (in larger topologies).

> Cheers

> Toerless

> > Best regards,

> > Sandy

> >

> >

> > ------------------原始邮件------------------

> > 发件人：Jeffrey(Zhaohui)Zhang

> > 收件人：[e-mail];[e-mail]<mailto:[e-mail];[e-mail]>;

> > 抄送人：[e-mail]<mailto:[e-mail]>;

> > 日 期 ：2022年02月10日 04:50

> > 主 题 ：Re: [Bier] BIER: draft-eckert-bier-cgm2-rbs-01 with performance

> > analysis simulation Hi Toerless, Not sure if my understanding is

> > correct, but it seems that RBS does not reduce the number of bits that are needed to encode the tree. Rather, it increases the number of bits (to encode the recursive structure).

> > I agree that it reduces the size of BIFTs, but even current BIER-TE can reduce the number of copies if you use a longer bitstring?

> > Jeffrey

> > Juniper Business Use Only

> > -----Original Message-----

> > From: BIER <[e-mail]<mailto:[e-mail]>> On Behalf Of [e-mail]<mailto:[e-mail]>

> > Sent: Wednesday, February 9, 2022 2:19 PM

> > To: [e-mail]<mailto:[e-mail]>

> > Cc: [e-mail]<mailto:[e-mail]>

> > Subject: [Bier] BIER: draft-eckert-bier-cgm2-rbs-01 with performance

> > analysis simulation [External Email. Be cautious of content] Dear

> > BIER-TE WG:

> > Robin did add a section (6.3) describing an initial performance gain analysis of CGM2/RBS to the github source (https://urldefense.com/v3/__https://github.com/toerless/bier-cgm2-rbs__;!!NEt6yMaO-gk!WS49OT72vlonSWP3yLtLcW_RQARYP00KEiAWpH592AuDXmrOOJH_bgVzQZyfK9En$<https://urldefense.com/v3/__https:/github.com/toerless/bier-cgm2-rbs__;!!NEt6yMaO-gk!WS49OT72vlonSWP3yLtLcW_RQARYP00KEiAWpH592AuDXmrOOJH_bgVzQZyfK9En$> ), and i just did a bit of editorial fixup and posted it as -01 of the draft.

> > This actually is the first time i actually like the HTML'ized version of a draft, because the topology picture is so large it doesn't fit a single page:

> > https://urldefense.com/v3/__https://www.ietf.org/archive/id/draft-ec<https://urldefense.com/v3/__https:/www.ietf.org/archive/id/draft-ec>

> > kert-bier-cgm2-rbs-01.html__;!!NEt6yMaO-gk!WS49OT72vlonSWP3yLtLcW_RQ

> > ARYP00KEiAWpH592AuDXmrOOJH_bgVzQaBIgBPC$

> > The interesting piece about the comparison is that it is actually comparing CGM2/RBS to BIER, and not BIER-TE. Because BIER itself should be requiring less copies than BIER-TE, so the gain of CGM2/RBS over BIER-TE should be even higher, but the fact alone that you get away with fewer packet copies to large receiver sets even though the bitstring also needs to encode the path/tree towards the receivers is really cool.

> > Robin, two Q:

> > 1. The new text mentions "in our graphs", but the text does not include any such graphs (yet).

> > I guess such a graph would be even worse to convert to ASCII than the topology.

> > Maybe post whatever format you have those results in to github (PDF, png...) and then we actually may want to see if/how a PDF version of the draft could include better than just ASCII art. Certainly a good reason to finally try it out.

> > And short term we can just add references to such visuals to the draft.

> > 2; Is it correct to assume that the hops through the topology that you simulated are "just" shortest-path, maybe with some ECMP choice - aka: the same paths that also BIER would choose given some "default" IGP routing setup ?

> > Cheers

> > Toerless

> > _______________________________________________

> > BIER mailing list

> > [e-mail]<mailto:[e-mail]>

> > https://urldefense.com/v3/__https://www.ietf.org/mailman/listinfo/bi<https://urldefense.com/v3/__https:/www.ietf.org/mailman/listinfo/bi>

> > er__;!!NEt6yMaO-gk!WS49OT72vlonSWP3yLtLcW_RQARYP00KEiAWpH592AuDXmrOO

> > JH_bgVzQZXGUd9P$ _______________________________________________

> > BIER mailing list

> > [e-mail]<mailto:[e-mail]>

> > https://www.ietf.org/mailman/listinfo/bier

> --

> ---

> [e-mail]<mailto:[e-mail]>



--

---

[e-mail]<mailto:[e-mail]>