Re: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

[Lizhenbin <[e-mail]>]

Hi Acee,
Thanks for your suggestion. It should take more work for the solution proposed by the draft.

Best Regards,
Robin



From: Lsr [mailto:[e-mail]] On Behalf Of Acee Lindem (acee)
Sent: Wednesday, March 10, 2021 10:40 PM
To: Tony Przygienda <[e-mail]>; Les Ginsberg (ginsberg) <ginsberg=[e-mail]>
Cc: Gyan Mishra <[e-mail]>; Robert Raszuk <[e-mail]>; Huzhibo <[e-mail]>; Peter Psenak (ppsenak) <[e-mail]>; [e-mail]; Tony Li <[e-mail]>; lsr <[e-mail]>; Tianran Zhou <[e-mail]>; wangyali <[e-mail]>
Subject: Re: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

The draft specifies a TLV to tag LSPs and immediately jumps to the conclusion that this is sufficient for implementations to provide separation  of routing and non-routing information. Now there are claims that this COULD imply flooding optimizations. Until the draft specifies how this separation is accomplished, it is a non-starter. I’m not encouraging the authors to do this as we already have multiple instances that provides a natural separation and we don’t need complex protocol extensions. If you do persist with this ill-advised draft, I’d ask you actually implement what you are proposing as a POC.

Thanks,
Acee

From: Lsr <[e-mail]<mailto:[e-mail]>> on behalf of Tony Przygienda <[e-mail]<mailto:[e-mail]>>
Date: Wednesday, March 10, 2021 at 8:11 AM
To: "Les Ginsberg (ginsberg)" <ginsberg=[e-mail]<mailto:ginsberg=[e-mail]>>
Cc: Tianran Zhou <[e-mail]<mailto:[e-mail]>>, Gyan Mishra <[e-mail]<mailto:[e-mail]>>, Robert Raszuk <[e-mail]<mailto:[e-mail]>>, Huzhibo <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>, Tony Li <[e-mail]<mailto:[e-mail]>>, lsr <[e-mail]<mailto:[e-mail]>>, wangyali <[e-mail]<mailto:[e-mail]>>, "Peter Psenak (ppsenak)" <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

I just add that SPT'ing the stuff to constrain the "flooding graphs" you can create wonderful chains with choking points for flooding in the graph, especially if you want to slosh lots info around that way. And in particularly devious critical case ALL your topologies may end up flooding through a single link as a choking point. RFC1925 (11) to which I should write an errata with (ii) as "Tony's corollary" along the lines of "it is often that the re-invention of the idea with a new name breaks it or makes it more broken if it was broken before" and (iii) "if the re-invention is broken the only remedy is often to force its implementation and attempt @ deployment" ;-)

-- tony

On Wed, Mar 10, 2021 at 1:58 PM Les Ginsberg (ginsberg) <ginsberg=[e-mail]<mailto:[e-mail]>> wrote:
Aijun –

Inline.

From: [e-mail]<mailto:[e-mail]> <[e-mail]<mailto:[e-mail]>>
Sent: Tuesday, March 09, 2021 10:14 PM
To: Les Ginsberg (ginsberg) <[e-mail]<mailto:[e-mail]>>; 'wangyali' <[e-mail]<mailto:[e-mail]>>; 'Gyan Mishra' <[e-mail]<mailto:[e-mail]>>; Peter Psenak (ppsenak) <[e-mail]<mailto:[e-mail]>>
Cc: 'Robert Raszuk' <[e-mail]<mailto:[e-mail]>>; 'Huzhibo' <[e-mail]<mailto:[e-mail]>>; 'Tony Li' <[e-mail]<mailto:[e-mail]>>; 'lsr' <[e-mail]<mailto:[e-mail]>>; 'Tianran Zhou' <[e-mail]<mailto:[e-mail]>>
Subject: RE: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

Hi, Les:

Is there any deployment of Multi-Instance to transfer the application info as recommend in https://tools.ietf.org/html/rfc6823#section-5?
From my POV, the transfer of the application information needs not even be flooded to all of the device’s participated circuits, as that does for the normal routing LSP.
Such information needs only be reached to the IGP nodes, then they can follows the shortest path calculated by the base routing instance(reach the flooding reduction effect).
Thus, why not considering using the existing routing instance? It can certainly decrease the overhead of neighbor configurations/maintenances for the new routing instance.

[Les:] Hmmm…I do not know what you mean by “IGP Nodes”. Given we are always talking about nodes which are running IS-IS I would assume all nodes are “IGP Nodes”. Apparently you have some other distinction in mind, but what it is I do not know.

If your concern is that when using MI it is still necessary to calculate an SPT in order to track the set of MI/ITID nodes which are currently reachable, I would remind you that https://tools.ietf.org/html/rfc8202#section-4 recommends minimizing the number of instances/flooding topologies. MI has never advocated  – nor required – that there be a 1-1 mapping between a flooding topology and an application. The idea that if there are thousands of MFIs (to borrow the term from your draft) it would be required to have thousands of IID/ITID pairs when using MI is incorrect. The number of flooding topologies should be minimized as per RFC 8202 Section 4.

I also point out that if it is your intent as part of your draft to have a separate flooding topology/MFI (which is what your text suggests),  you will have to build an MFI specific SPT in order to flood the MFI specific LSDB only to the nodes who ae interested in this MFI specific LSDB.
Your other alternative is to share a single flooding topology for multiple MFI specific LSDBs, in which case you can reduce the number of SPTs which need to be calculated but at the cost of potentially sending MFI specific information to nodes which are not interested in it. In this regard, the options available are identical to what MI offers.

The difference between MFI draft and RFC 8202 is that with the latter we have complete separation of the Update Process(es) used for distributing the application information from the operation of the Standard Instance. The scaling characteristics of the two solutions are identical, something which you seem to have misunderstood.


And, https://datatracker.ietf.org/doc/html/draft-bowers-lsr-isis-gen-info-clarifications-00 describes the necessary for the correlation of app info and other info that advertised in different routing instances.
If we put them into the basic routing instance, and make their flooding does not influence the core routing calculation function, won’t be it more attractive?

[Les:] I am aware of and have read this draft. I do not believe it has any relevance to this discussion. Here we are discussing whether there is a need for yet another definition of how to instantiate Update Process(es) used for sending Application information. Let’s please confine this thread to that discussion.
I believe it is only your confusion regarding the number of IID/ITID pairs that might be required when using RFC 8202 that suggests to you that this draft might be relevant. Hopefully I have corrected your misunderstanding.

    Les

Best Regards

Aijun Wang
China Telecom

From: [e-mail]<mailto:[e-mail]> <[e-mail]<mailto:[e-mail]>>
Sent: Wednesday, March 10, 2021 4:25 AM
To: wangyali <[e-mail]<mailto:[e-mail]>>; Gyan Mishra <[e-mail]<mailto:[e-mail]>>; Peter Psenak (ppsenak) <[e-mail]<mailto:[e-mail]>>
Cc: Robert Raszuk <[e-mail]<mailto:[e-mail]>>; Huzhibo <[e-mail]<mailto:[e-mail]>>; Aijun Wang <[e-mail]<mailto:[e-mail]>>; Tony Li <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; Tianran Zhou <[e-mail]<mailto:[e-mail]>>
Subject: RE: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

Yali –

I find your responses not very helpful.
Please see inline.

From: wangyali <[e-mail]<mailto:[e-mail]>>
Sent: Tuesday, March 09, 2021 6:22 AM
To: Les Ginsberg (ginsberg) <[e-mail]<mailto:[e-mail]>>; Gyan Mishra <[e-mail]<mailto:[e-mail]>>; Peter Psenak (ppsenak) <[e-mail]<mailto:[e-mail]>>
Cc: Robert Raszuk <[e-mail]<mailto:[e-mail]>>; Huzhibo <[e-mail]<mailto:[e-mail]>>; Aijun Wang <[e-mail]<mailto:[e-mail]>>; Tony Li <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; Tianran Zhou <[e-mail]<mailto:[e-mail]>>
Subject: RE: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

Hi Les,
Glad to receive your email. Please see inline [Yali].

From: Les Ginsberg (ginsberg) [mailto:[e-mail]]
Sent: Tuesday, March 9, 2021 11:15 AM
To: Gyan Mishra <[e-mail]<mailto:[e-mail]>>; Peter Psenak (ppsenak) <[e-mail]<mailto:[e-mail]>>
Cc: Robert Raszuk <[e-mail]<mailto:[e-mail]>>; Huzhibo <[e-mail]<mailto:[e-mail]>>; Aijun Wang <[e-mail]<mailto:[e-mail]>>; Tony Li <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; Tianran Zhou <[e-mail]<mailto:[e-mail]>>; wangyali <[e-mail]<mailto:[e-mail]>>
Subject: RE: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

Sooo, I have been reluctant to comment on the shortcomings of this draft because I feel there was no need for the draft to be written in the first place.
I had hoped that the authors would think about this a bit more and realize the flaws in the proposed solution – or – as Acee suggested during the WG meeting – they would attempt an implementation and discover what had not yet been realized. This then would end the time the WG is spending on this – which IMO is the right outcome.

As that has not yet happened, perhaps some comments can speed this process along.

The goal of the draft is to support per-MFI LSDBs in the standard instance of IS-IS.  Since it is not possible for a legacy node to differentiate LSP.xx-yy w MFI #1 from the same LSP with MFI #2 (or with no MFI at all), it is clear that an MFI LSP cannot be flooded to a legacy node EVER!!
[Yali]: In our proposal, we considered the scenario where some routers that do not support MFI, and recommended that all MFIs share one LSP Number space in the standard instance of IS-IS protocol. And if routers that do not support MFI but receive the LSPs and SNPs carrying MFI-ID TLV, then routers SHOULD ignore the MFI-ID TLV and continues processing other TLVs.

[Les:] Yes – I understand this. But, as I point out below, this case is already provided for by RFC 6823.  We do discourage doing this – as we discourage your proposed use of the standard instance. But it is already defined – so your draft adds no new functionality for this case.

In order to prevent this, a node has to know whether a given neighbor supports MFI or not. But since the draft defines no signaling in hellos, it cannot tell whether the neighbor supports MFI (not to mention which MFIs – which is important for avoiding flooding MFI LSPs to nodes that aren’t interested in that MFI) you are forced to rely on receiving LSPs (or SNPs) – which brings us to the chicken/egg problem. Neither I nor my neighbor can send an MFI LSP out of fear that the receiver does not support MFI. So MFI flooding is blocked.

This problem can be solved by including the MFI TLV in hellos (analogous to what MI(RFRC 8202) does). But this is not the end of your issues. If you have a LAN, you could have a mix of legacy routers and MFI routers – and again you cannot allow legacy routers to receive MFI LSPs as they will look just like legacy LSPs to the legacy nodes. This means you will have to find a way to avoid ever having MFI PDUs received by legacy nodes (RFC 8202 uses different MAC multicast addresses).
[Yali]: Correctly. You’re right. There indeed be a signaling in Hellos. MFI-ID TLV can be included in IS-IS Hellos. Because we want to focus on extensions to Update process, we originally plan to implement this part in the second version.
We also have describe the Interoperability Considerations in the condition where some routers do not support MFI in the draft.

[Les:] Glad you agree. But the point is – when you have finished your draft – you will simply end up with a variation of RFC 8202. So it won’t be simpler – which was one of your key goals.
This is one of many reasons why I see no need for your draft.

Sooo, once you have addressed both of these issues you will have repeated what RFC 8202 already does. No new benefits here.

This then leaves you with one possible use case: support a single LSDB for all MFIs in the standard IS-IS instance. But, this use case is already provided for (though strongly discouraged) by https://tools.ietf.org/html/rfc6823#section-5 .
[Yali]: This proposal is used to solve the issue that how we can isolate the impacts of non-routing information flooding under the condition that there is only the standard instance of IS-IS protocol implemented in the network and no one router support Multi-instance. Sec.5 also mentioned the issue as follows.



[Les:] If – as you seem to agree – you can only support a single LSDB, there is no need for any protocol extensions and no need for the new MFI TLV. You could advertise GENINFPO for multiple apps in the standard instance.

So your table below is irrelevant.



‘’ Flooding of information not directly related
   to the use of the IS-IS protocol in support of routing degrades the
   operation of the protocol.  Degradation occurs because the frequency
   of LSP updates is increased and because the processing of non-routing
   information in each router consumes resources whose primary
   responsibility is to efficiently respond to reachability changes in
   the network. ‘’

So in this proposed draft, we give an optional solution by separate multiple Update processes for different kinds of information, such as following example shown in the Table (I presented in the yesterday meeting). Each Level 1/Level 2 LSPs associated with a specific MFI carries flooding information, such as routing topology, belonging to the specific MFI #0. And Level 1/Level 2 PSNPs and Level 1/Level 2 CSNP containing information about LSPs that transmitted in a specific MFI#0 are generated to synchronize the information propagated in the MFI #0 LSDB that is subdivided from a single LSDB within the standard instance of protocol.

[cid:image001.png@01D7168C.02ACF1C0]

As shown in table, different flooding information can be propagated in the unique Update process. In each MFI, priority and update parameters can be customized in dependent on the requirements on the flooding rate of different information. Hence, in our proposal, we also provide an optional solution that can be used in the standard instance to solve the issue also mentioned in section 5 as follows.

‘’ One of the most egregious oversights is a
   failure to appropriately dampen changes in the information to be
   advertised; this can lead to flooding storms. “

I do understand that some folks want to advertise VTN info in IGPs and that the WG will be discussing this. I am not in favor of doing this – but I recognize it is a legitimate topic for discussion. And if the WG were to approve such functionality we have MI available to be used to provide separation.
(Note that MI has been implemented and successfully deployed by multiple vendors.)
[Yali]: Well, then may I post a question (or maybe a Survey) here for folks from operators? The question is that when deploying thousands of slices, are you willing to implement thousands of IS-IS instances to flood thousands of slices topologies and so on? From our side, we can't imagine what a huge project it is.
We’re appreciate if you are willing to give us your answer or opinion.

[Les:] This is not an appropriate response. No one – not you, not me, nor anyone else on this thread – has suggested that thousands of IS-IS instances would be required. For you to suggest this  induces uncalled for hysteria.
It is obvious that if you could send info for multiple applications in the standard instance you can do so in a single non-standard instance – but we would at least have separation from the routing instance. Certainly there is no need for thousands of instances.
Please do not compromise a meaningful dialogue with this misdirection.

If you read https://www.rfc-editor.org/rfc/rfc8202.html#section-4 there is a discussion of related issues. Section 4.3 is particularly relevant to your comment.

   Les

draft-wang-lsr-isis-mfi is an unnecessary proposal, seriously flawed, and not achieving any of the goals stated in its introduction.
I ask that the authors abandon this proposal.

   Les


From: Lsr <[e-mail]<mailto:[e-mail]>> On Behalf Of Gyan Mishra
Sent: Friday, March 05, 2021 8:11 AM
To: Peter Psenak (ppsenak) <[e-mail]<mailto:[e-mail]>>
Cc: Robert Raszuk <[e-mail]<mailto:[e-mail]>>; Huzhibo <[e-mail]<mailto:[e-mail]>>; Aijun Wang <[e-mail]<mailto:[e-mail]>>; Tony Li <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; Tianran Zhou <[e-mail]<mailto:[e-mail]>>; wangyali <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsr] New Version Notification for draft-wang-lsr-isis-mfi-00.txt

Hi Peter

On Fri, Mar 5, 2021 at 10:56 AM Peter Psenak <[e-mail]<mailto:[e-mail]>> wrote:
Gyan,

On 05/03/2021 16:46, Gyan Mishra wrote:
> Yali
>
> I agree with a Peter.
>
> As for resource isolation and provisioning of a VTN I think you really
> need separate LSDB instances provided by MT or MI as better suited for
> network slicing.

MT does not provide LSDB separation, only MI does.

thanks,
Peter

   I thought that each MT topology was a separate RIB meaning separate LSDB.  The RFC is confusing.😄

https://tools.ietf.org/html/rfc5120

6<https://tools.ietf.org/html/rfc5120#section-6>.  MT SPF Computation





   Each MT MUST run its own instance of the decision process.  The

   pseudo-node LSPs are used by all topologies during computation.  Each

   non-default topology MAY have its attached bit and overload bit set

   in the MT TLV.  A reverse-connectivity check within SPF MUST follow

   the according MT to assure the bi-directional reachability within the

   same MT.



   The results of each computation SHOULD be stored in a separate

   Routing Information Base (RIB), in normal cases, otherwise

   overlapping addresses in different topologies could lead to

   undesirable routing behavior, such as forwarding loops.  The

   forwarding logic and configuration need to ensure the same MT is

   traversed from the source to the destination for packets.  The

   nexthops derived from the MT SPF MUST belong to the adjacencies



conforming to the same MT for correct forwarding.  It is recommended

   for the administrators to ensure consistent configuration of all

   routers in the domain to prevent undesirable forwarding behavior.



   No attempt is made in this document to allow one topology to

   calculate routes using the routing information from another topology

   inside SPF.  Even though it is possible to redistribute and leak

   routes from another IS-IS topology or from external sources, the

   exact mechanism is beyond the scope of this document.



>
> To me it seems a common LSDB shared among network slices VTN underlay
> could be problematic with network overlap issues.
>
> Kind Regards
>
> Gyan
>
> On Fri, Mar 5, 2021 at 10:28 AM Peter Psenak <[e-mail]<mailto:[e-mail]>
> <mailto:[e-mail]<mailto:[e-mail]>>> wrote:
>
>     Hi Yali,
>
>     On 05/03/2021 15:31, wangyali wrote:
>      > Hi Peter,
>      >
>      > Thanks for your question. Please see inline [yali3].
>      >
>      > -----Original Message-----
>      > From: Peter Psenak [mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>]
>      > Sent: Thursday, March 4, 2021 11:20 PM
>      > To: wangyali <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Gyan Mishra <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Robert Raszuk <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>
>      > Cc: Huzhibo <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Aijun Wang <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Tony Li <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; lsr <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Tianran Zhou <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      > Subject: Re: [Lsr] New Version Notification for
>     draft-wang-lsr-isis-mfi-00.txt
>      >
>      > Hi Yali,
>      >
>      > On 04/03/2021 14:45, wangyali wrote:
>      >> Hi Peter,
>      >>
>      >> Please see inline [Yali2]. Thanks a lot.
>      >>
>      >> -----Original Message-----
>      >> From: Peter Psenak [mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>]
>      >> Sent: Thursday, March 4, 2021 6:50 PM
>      >> To: wangyali <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Gyan Mishra
>      >> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Robert
>     Raszuk <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >> Cc: Huzhibo <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Aijun Wang
>      >> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Tony
>     Li <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; lsr
>      >> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Tianran Zhou
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >> Subject: Re: [Lsr] New Version Notification for
>      >> draft-wang-lsr-isis-mfi-00.txt
>      >>
>      >> Hi Yali,
>      >>
>      >> On 04/03/2021 11:42, wangyali wrote:
>      >>> Hi Peter,
>      >>>
>      >>> Please review follows tagged by [Yali].
>      >>>
>      >>>
>      >>> -----Original Message-----
>      >>> From: Peter Psenak [mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>]
>      >>> Sent: Wednesday, March 3, 2021 5:37 PM
>      >>> To: wangyali <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Gyan Mishra
>      >>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Robert
>     Raszuk <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>> Cc: Huzhibo <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Aijun Wang
>      >>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Tony Li <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; lsr
>      >>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Tianran Zhou
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>> Subject: Re: [Lsr] New Version Notification for
>      >>> draft-wang-lsr-isis-mfi-00.txt
>      >>>
>      >>> Yali,
>      >>>
>      >>> On 03/03/2021 06:02, wangyali wrote:
>      >>>> Hi Peter,
>      >>>>
>      >>>> Thanks for your comments. Yes. I am improving this sentence.
>     Please review the following update.
>      >>>>
>      >>>> OLD: " And Level 1/Level 2 PSNP and Level 1/Level 2 CSNP
>     containing information about LSPs that transmitted in a specific MFI
>     are generated to synchronize the LSDB corresponding to the specific
>     MFI."
>      >>>>
>      >>>> NEW: "And Level 1/Level 2 PSNP and Level 1/Level 2 CSNP
>     containing information about LSPs that transmitted in a specific MFI
>     are generated to synchronize the MFI-specific sub-LSDB. Each
>     MFI-specific sub-LSDB is subdivided from a single LSDB."
>      >>>
>      >>> please specify sub-LSDB.
>      >>> [Yali] Thanks for your comment. But to avoid introducing a new
>     term, I change to use "MFI-specific LSDB" instead of " MFI-specific
>     sub-LSDB ".  And we give the explanation that "Each MFI-specific
>     LSDB is subdivided from a single LSDB."
>      >>
>      >> I wonder what is the difference between "MFI-specific LSDB
>     subdivided from a single LSDB" versus the "MFI-specific LSDB".
>      >> [Yali2]: Actually I am trying to optimize and accurately
>     describe the key point that multiple Update processes associated
>     with each MFI operate on a common LSDB within the zero IS-IS
>     instance, and each Update process is isolated from each other and
>     does not affect each other.
>      >> So we say "MFI-specific LSDB subdivided from a single LSDB",
>     which may explicitly indicate each MFI-specific LSDB shares a common
>     LSDB but each Update process associated with a MFI is isolated.
>     However, from your previous question and suggestions,  "MFI-specific
>     LSDB" looks like unclear and misleading. Any good idea on improving
>     the expression are welcome.
>      >
>      >
>      > it's not the name that is as important. It's the concept that
>     looks questionable - how well can you isolate the update processing
>     if the data are part of the same LSDB and whether such update
>     process separation would prove to be useful at all. I don't know, so
>     far I have not seen any evidence.
>      > [yali3] This draft defines a new TLV, i.e. MFI-ID TLV,  which may
>     be included in each Level 1/Level 2 IS-IS LSPs and SNPs. Hence, each
>     Level 1/Level 2 LSPs and SNPs associated with each Update Process
>     can be uniquely identified by MFI-ID.
>      > In this draft, each flooding instance has its own separated
>     Update process, which isolates the impact of application information
>     flooding on the IS-IS protocol operation. So each Level 1/Level 2
>     LSP associated with a specific MFI carries flooding information
>     belonging to the specific MFI. And Level 1/Level 2 PSNP and Level
>     1/Level 2 CSNP containing information about LSPs that propagated in
>     the specific MFI are generated to synchronize the MFI-specific LSDB.
>
>     - by using the same LSDB to store the MFI specific information only a
>     limited separation can be achieved. Multi-instance gives you better
>     separation.
>
>     - you carved the MFI specific LSP space from the common LSP space. This
>     may result in the non routing apps consuming the space that would
>     otherwise be required for regular routing information, compromising the
>     basic functionality of the protocol. Multi-instance does not have that
>     problem.
>
>     my 2c,
>     Peter
>
>
>      >
>      > thanks,
>      > Peter
>      >
>      >
>      >
>      >
>      >>
>      >> thanks,
>      >> Peter
>      >>
>      >>>
>      >>> thanks,
>      >>> Peter
>      >>>
>      >>>
>      >>>>
>      >>>> Best,
>      >>>> Yali
>      >>>>
>      >>>> -----Original Message-----
>      >>>> From: Peter Psenak [mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>]
>      >>>> Sent: Tuesday, March 2, 2021 5:12 PM
>      >>>> To: wangyali <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Gyan Mishra
>      >>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Robert
>     Raszuk <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>> Cc: Huzhibo <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Aijun Wang
>      >>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Tony Li <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; lsr
>      >>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Tianran Zhou
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>> Subject: Re: [Lsr] New Version Notification for
>      >>>> draft-wang-lsr-isis-mfi-00.txt
>      >>>>
>      >>>> Yali,
>      >>>>
>      >>>> On 01/03/2021 10:49, wangyali wrote:
>      >>>>> Hi Peter,
>      >>>>>
>      >>>>> Many thanks for your feedback. First of all, I'm sorry for
>     the confusion I had caused you from my previous misunderstanding.
>      >>>>>
>      >>>>> And I want to clarify that a single and common LSDB is shared
>     by all MFIs.
>      >>>>
>      >>>> well, the draft says:
>      >>>>
>      >>>> "information about LSPs that transmitted in a
>      >>>>       specific MFI are generated to synchronize the LSDB
>     corresponding to
>      >>>>       the specific MFI."
>      >>>>
>      >>>> If the above has changed, then please update the draft
>     accordingly.
>      >>>>
>      >>>> thanks,
>      >>>> Peter
>      >>>>
>      >>>>
>      >>>>
>      >>>>>
>      >>>>> Best,
>      >>>>> Yali
>      >>>>>
>      >>>>> -----Original Message-----
>      >>>>> From: Peter Psenak [mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>]
>      >>>>> Sent: Sunday, February 28, 2021 8:23 PM
>      >>>>> To: Gyan Mishra <[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>; Robert Raszuk
>      >>>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>>> Cc: Huzhibo <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Aijun Wang
>      >>>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>;
>     Tony Li <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; lsr
>      >>>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; Tianran Zhou
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>; wangyali
>      >>>>> <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>>> Subject: Re: [Lsr] New Version Notification for
>      >>>>> draft-wang-lsr-isis-mfi-00.txt
>      >>>>>
>      >>>>> Gyan,
>      >>>>>
>      >>>>> On 26/02/2021 17:19, Gyan Mishra wrote:
>      >>>>>>
>      >>>>>> MFI seems more like flex algo with multiple sub topologies
>     sharing
>      >>>>>> a common links in a  topology where RFC 8202 MI is separated at
>      >>>>>> the process level separate LSDB.  So completely different and of
>      >>>>>> course different goals and use cases for MI versus MFI.
>      >>>>>
>      >>>>> I would not use the fle-algo analogy - all flex-algos operate
>     on top of a single LSDB, contrary to what is being proposed in MFI
>     draft.
>      >>>>>
>      >>>>>>
>      >>>>>>        MFI also seems to be a flood reduction mechanism by
>     creating
>      >>>>>> multiple sub topology instances within a common LSDB.  There
>     are a
>      >>>>>> number of flood reduction drafts and this seems to be another
>      >>>>>> method of achieving the same.
>      >>>>>
>      >>>>> MFI draft proposes to keep the separate LSDB per MFI, so the
>     above analogy is not correct either.
>      >>>>>
>      >>>>> thanks,
>      >>>>> Peter
>      >>>>>
>      >>>>>
>      >>>>>>
>      >>>>>> Gyan
>      >>>>>>
>      >>>>>> On Fri, Feb 26, 2021 at 7:10 AM Robert Raszuk
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>
>      >>>>>> <mailto:[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>>> wrote:
>      >>>>>>
>      >>>>>>          Aijun,
>      >>>>>>
>      >>>>>>          How multi instance is implemented is at the
>     discretion of a vendor.
>      >>>>>>          It can be one process N threads or N processes. It
>     can be both and
>      >>>>>>          operator may choose.
>      >>>>>>
>      >>>>>>          MFI is just one process - by the spec - so it is
>     inferior.
>      >>>>>>
>      >>>>>>          Cheers,
>      >>>>>>          R.
>      >>>>>>
>      >>>>>>
>      >>>>>>          On Fri, Feb 26, 2021 at 12:44 PM Aijun Wang
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>
>      >>>>>>          <mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>> wrote:
>      >>>>>>
>      >>>>>>              Hi, Robert:
>      >>>>>>
>      >>>>>>              Separate into different protocol instances can
>     accomplish the
>      >>>>>>              similar task, but it has<https://www.google.com/maps/search/ar+task,+but+it+has?entry=gmail&source=g> some deployment overhead.
>      >>>>>>              MFIs within one instance can avoid such
>     cumbersome work, and
>      >>>>>>              doesn’t affect the basic routing calculation
>     process.
>      >>>>>>
>      >>>>>>              Aijun Wang
>      >>>>>>              China Telecom
>      >>>>>>
>      >>>>>>>              On Feb 26, 2021, at 19:00, Robert Raszuk
>     <[e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>>
>      >>>>>>>              <mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>> wrote:
>      >>>>>>>
>      >>>>>>>              Hi Yali,
>      >>>>>>>
>      >>>>>>>                  If this was precise, then the existing
>     multi-instance
>      >>>>>>>                  mechanism would be sufficient.
>      >>>>>>>                  [Yali]: MFI is a different solution we
>     recommend
>     <https://www.google.com/maps/search/lution+we+recommend?entry=gmail&source=g>
>     to solve
>      >>>>>>>                  this same and valuable issue.
>      >>>>>>>
>      >>>>>>>
>      >>>>>>>              Well the way I understand this proposal MFI is
>     much weaker
>      >>>>>>>              solution in terms of required separation.
>      >>>>>>>
>      >>>>>>>              In contrast RFC8202 allows to separate ISIS
>     instances at the
>      >>>>>>>              process level, but here MFIs as defined must
>     be handled by the
>      >>>>>>>              same ISIS process
>      >>>>>>>
>      >>>>>>>                  This document defines an extension to
>      >>>>>>>                  IS-IS to allow*one standard instance*  of
>      >>>>>>>                  the protocol to support multiple update
>      >>>>>>>                  process operations.
>      >>>>>>>
>      >>>>>>>              Thx,
>      >>>>>>>              R.
>      >>>>>>>
>      >>>>>>>              _______________________________________________
>      >>>>>>>              Lsr mailing list
>      >>>>>>> [e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>> <mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>>>>> https://www.ietf.org/mailman/listinfo/lsr
>      >>>>>>
>      >>>>>>          _______________________________________________
>      >>>>>>          Lsr mailing list
>      >>>>>> [e-mail]<mailto:[e-mail]> <mailto:[e-mail]<mailto:[e-mail]>> <mailto:[e-mail]<mailto:[e-mail]>
>     <mailto:[e-mail]<mailto:[e-mail]>>>
>      >>>>>> https://www.ietf.org/mailman/listinfo/lsr
>      >>>>>>
>      >>>>>> --
>      >>>>>>
>      >>>>>> <http://www.verizon.com/>
>      >>>>>>
>      >>>>>> *Gyan Mishra*
>      >>>>>>
>      >>>>>> /Network Solutions A//rchitect /
>      >>>>>>
>      >>>>>> /M 301 502-1347
>      >>>>>> 13101 Columbia Pike
>      >>>>>> /Silver Spring, MD
>      >>>>>>
>      >>>>>>
>      >>>>>
>      >>>>>
>      >>>>>
>      >>>>
>      >>>>
>      >>>>
>      >>>
>      >>>
>      >>>
>      >>
>      >>
>      >>
>      >
>      >
>      >
>
> --
>
> <http://www.verizon.com/>
>
> *Gyan Mishra*
>
> /Network Solutions A//rchitect /
>
> /M 301 502-1347
> 13101 Columbia Pike
> /Silver Spring, MD
>
>
--

[Image removed by sender.]<http://www.verizon.com/>

Gyan Mishra

Network Solutions Architect

M 301 502-1347
13101 Columbia Pike
Silver Spring, MD

_______________________________________________
Lsr mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/lsr