Re: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06

[[e-mail]]

Hi Ketan,

Top posting in effort to also take a step back.

I could understand the following sematic for the anycast flag: (beware) this prefix may be an anycast prefix
In which case, this is an additional indication, it’s not mandated for any existing behavior, existing behaviors are unchanged and routers needs to be equally capable of handling anycast prefix which don’t have this AC-flag (just like today).
Does this align with your objective?

If so, I have the following comments:


“A prefix that is advertised by a single node and without an AC-flag MUST be considered node specific.” (*2)



I disagree with this sentence which change the existing behavior and does not align with the above semantic.

For prefix without the AC-flag, one has no new information compared to today and the behavior should be unchanged.

The semantic is AC-flag set --> anycast prefix (semantic is not: AC-flag unset --> prefix is unicasted)


“Both SR-MPLS prefix-SID and IPv4 prefix may be configured as anycast and as such the same value can be advertised by multiple routers.”

Sorry I’m not familiar with OSPF, but ideally the semantic would be the same for IS-IS. For IS-IS, multiple L1L2 routers (or ASBR) would typically advertise the same prefix when those prefixes are redistributed from another area/domain.  My reading is that the advertisement of the same prefix by multiple ASBR/L1L2 routers does not qualify those prefix as anycast. Is that a correct understanding? Regardless, I would welcome a clear definition of “anycast”  in the context of IGP. (for MPLS, I guess that we could say that a prefix is advertised by multiple LERs but I’m not sure there is an equivalent term for IGP)


Some minor comments:

“The AC-Flag MUST be preserved when re-advertising the prefix across areas. »
Ideally also across (IGP) redistribution. (I guess one could say that this implementation specific but if we need the AC-flag we also need it across domains)

A priori, removing the term “SR-MPLS” does not change the fact that the AC-flag could be set on SR-MPLS SID. So the removal seem mostly cosmetic^W editorial to me 😉

Thanks
--Bruno
From: Ketan Talaulikar <[e-mail]>
Sent: Friday, March 22, 2024 3:30 AM
To: DECRAENE Bruno INNOV/NET <[e-mail]>
Cc: Acee Lindem <[e-mail]>; lsr <[e-mail]>; [e-mail]; Dongjie (Jimmy) <[e-mail]>; Tony Przygienda <[e-mail]>
Subject: Re: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06

Hi Bruno,

Please check inline below with KT3.


On Thu, Mar 21, 2024 at 11:28 PM <[e-mail]<mailto:[e-mail]>> wrote:
Hi Ketan,

Please see inline [Bruno2]

From: Ketan Talaulikar <[e-mail]<mailto:[e-mail]>>
Sent: Thursday, March 21, 2024 4:19 PM
To: DECRAENE Bruno INNOV/NET <[e-mail]<mailto:[e-mail]>>
Cc: Acee Lindem <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>; Dongjie (Jimmy) <[e-mail]<mailto:[e-mail]>>; Tony Przygienda <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06

Hi Bruno,

Please check inline below with KT2 for responses.


On Thu, Mar 21, 2024 at 7:16 PM <[e-mail]<mailto:[e-mail]>> wrote:
Hi Ketan,

Thanks for your quick reply.
Please see inline [Bruno]

From: Ketan Talaulikar <[e-mail]<mailto:[e-mail]>>
Sent: Thursday, March 21, 2024 2:18 PM
To: DECRAENE Bruno INNOV/NET <[e-mail]<mailto:[e-mail]>>
Cc: Acee Lindem <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>; Dongjie (Jimmy) <[e-mail]<mailto:[e-mail]>>; Tony Przygienda <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06

Hi Bruno,

Thanks for your feedback. Please check inline below for responses.


On Thu, Mar 21, 2024 at 4:12 PM <[e-mail]<mailto:[e-mail]>> wrote:
Hi all,

I would also welcome a clear specification of the semantics.
Such that the meaning and implications are clear on both the originator and the receivers/consumers.

e.g., from the originator standpoint:
- The originator MAY advertise the Anycast Flag if CONDITIONS1 are met (which allow for some useful features such as….)
- The originator MUST advertise the Anycast Flag if CONDITIONS1 are met (otherwise this breaks …)

Please specify the CONDITIONS1.

KT> Whether a prefix is anycast or not is configured by the operator. This spec does not require implementations to detect that a prefix that it is originating is also being originated from another node and hence may be an anycast advertisement. We can clarify the same in the document.

[Bruno] As an operator, why would I configure this? What for? What are the possible drawbacks? (i.e., can this be configured on all prefixes regardless of their anycast status)

KT2> If anycast property is configured on all prefixes, then it is an indication that none of those prefixes resolve to a unique node. That has consequences in terms of usage. E.g., taking the TI-LFA repair path use-case, we won't find the Node SID to be used to form the repair segment-list.

[Bruno2] Given OSPFv2, by SR you mean SR-MPLS I guess. For TI-LFA, if you want a Node SID, why not simply picking a SID having the N flag. That’s its semantic. Also with SR-MPLS we don’t do much aggregation so I’m not sure to see use for prefix. (by prefix, I mean not a /32 address)

KT3> Yes, that is why we had the N flag for that specific use case. I assume there are no concerns with the use of the N flag and its semantics.


I would propose those points be discussed in the operation considerations section of this draft.
In the absence of reason, this is not likely be configured IMHO.

KT2> Sure. Thanks for that feedback. We can certainly do that in the draft. I hope this isn't blocking the adoption in your view though, right?

[Bruno2] I haven’t asked for blocking the adoption. I asked for clearly specified semantic.


e.g., from the receiver standpoint:
What does this mean to have this Anycast Flag set? What properties are being signaled? (a priori this may be already specified by CONDITIONS1 above)

KT> In addition to the previous response, for the receiver this means that the same prefix MAY be advertised from more than one node (that may be happening now or may happen in the future). This can be clarified as well.

[Bruno] OK. If this is happening now, this is a priori already visible in the LSDB.

KT2> This is tricky. If the prefix is originated in a different domain, it gets tricky to determine if the prefix is anycast or dual-homed since the LSDB has a local area/domain view.

[Bruno2] Agreed for prefix. For Node-SID you have the N-flag. Regarding origination in another domain, would the ABR/L1L2 node be able to detect this and set the anycast flag by itself?

KT3> It cannot if the case is of anycast originating from different domains/areas.


Any reason to duplicate the info (I would guess that’s easier for implementation but since this is not guaranteed to be implemented one would need to also check in the LSDB. So doubling the work).

KT2> This extension brings in simplicity for the receivers provided that operators can configure this property.

[Bruno2] aka moving the complexity to the service provider. I guess you would not be surprised if I prefer the other way around (have computer do the job instead of humans, have vendors do the job rather than operator 😉) Configuring states and having to maintain/updates them forever is akin to a technical debt to me.

KT3> Here, I think, we may have a point of disagreement. While it is outside the scope of this document, I hope we agree that there is a lot more involved in the configuration of anycast prefix and the service/use-case behind it. The Anycast property config provides a very small additional "state" to be provisioned as part of a larger anycast service/use-case provisioning. It allows the operator to robustly indicate this property of the prefix (they know it is anycast) via the IGP without requiring routers and applications to algorithmically figure this out (that might not always be correct). I think of it as a useful optional lego block in the set of IGP extensions.


KT2>  Like I mentioned above, this starts to get more complicated in multi-domain scenarios. Perhaps we can think of this as the complexities that we experience in determining this property via an LSDB/topology-db that motivate us to bring forth this easier and more robust way.

Any specific reason requiring the knowledge of the future?

KT2> Perhaps at time T1, there are two nodes originating the prefix. Then at time T2, one of them goes down (or becomes disconnected), do we assume that the prefix is now not anycast? Then what happens if that other node comes back up again. For certain use-cases where anycast prefix is not desired, it may be helpful to completely avoid use of this prefix. The operator knows their design and addressing and perhaps is able to provision this prefix property correctly from the outset.


[Bruno2] I guess there could be such use cases. But a priori in the general case, when that other node come back 1) before IGP convergence nothing change from a routing standpoint, 2) during routing convergence you know about this other node and can do what you want. This includes updating your FRR protection. If this is really a concerned (to assume anycast status while it’s not certain) I find a sentence problematic in the draft “A prefix that is advertised by a single node and without an AC-flag MUST be considered node specific. ». TIn fact, the receiver does not know whether this is a node specific prefix or an anycast prefix advertised by a node not supporting this extension (or an operator not doing the right configuration).

KT3> We have the N and the AC flag. If they are configured properly, then there is no ambiguity. But what if they are not? What if there is a prefix w/o either of the flags set and say for the use-case like TI-LFA we need to use that as a node identifier (because there is nothing else from that node). That is the ambiguity that we are trying to cover. Btw, that same text is there in RFC9352/9513 and therefore also in this document for consistency across the IGPs.





If this is specific to SR,  please say so.

KT> It is not specific to SR, it is a property of an IP prefix.

But even in this sub-case, SR anycast has some conditions, both for SR-MPLS and SRv6.

KT> This document does not discuss either SR-MPLS or SRv6 anycast. It covers an OSPFv2 extension to simply advertise the anycast property of any IP prefix. The discussion of SR anycast belongs to some other (SPRING) document ;-)
[Bruno2] I’m sorry but “SR-MPLS” is the second word in the abstract. So I believe this document covers SR-MPLS. IMO anything specific to SR-MPLS caused by this document should be covered in this document.

KT3> That is a mistake that Les has also pointed out. We will fix that.





SR-MPLS:  https://datatracker.ietf.org/doc/html/rfc8402#section-3.3.1

“determining the second label is impossible unless A1 and A2 allocated the same label value to the same prefix.”

“Using an anycast segment without configuring identical SRGBs on all
   nodes belonging to the same anycast group may lead to misrouting (in
   an MPLS VPN deployment, some traffic may leak between VPNs).”

So for SR-MPLS, where we did not have anycast flag at the time, the burden of respecting the conditions seems to be on the receiver. In which case, Anycast flag didn’t seem to be required.

KT> True. But that was also beyond the anycast property of the prefix - it also involves checking the Prefix SID associated with it (plus other considerations) and that is something quite different.
[Bruno2] That’s about anycast SR-MPLS SID which is the scope of your document.

KT3> Agree



SRv6: https://datatracker.ietf.org/doc/html/rfc9352#name-advertising-anycast-propert
“All the nodes advertising the same anycast locator MUST instantiate the exact same set of SIDs under that anycast locator. Failure to do so may result in traffic being dropped or misrouted.”


So for SRv6 the burden is on the originator, and we felt the need to define an anycast flag.

KT> Note that RFC9352 does not restrict the advertisement of anycast property of the prefix to SRv6. It applies to all IPv4/IPv6 prefixes - irrespective of SRv6, SR-MPLSv4, SR-MPLSv6 or plain old IP. This is the same for RFC9513 - since OSPFv3 supports IPv4/IPv6 prefixes as well as SRv6, SR-MPLSv4, and SR-MPLSv6.
[Bruno] Indeed. And note that  RFC9352 did specify some specific conditions (MUST) before a node may advertise this anycast flag. A priori there is a reason for this. A priori the same reason would apply to SR-MPLS, no? So why this sentence has not also been copied from RFC9352 and adapted for SR-MPLS? (the sentence is “All the nodes advertising the same anycast locator MUST instantiate the exact same set of SIDs under that anycast locator. Failure to do so may result in traffic being dropped or misrouted.”)

KT2> You have a good point. All I can say is that RFC9352/9513 were focussed on SRv6 extensions and therefore covered only those aspects. This document is not an SR extension and I feel it is better that these aspects related to SR anycast (SR-MPLS or SRv6) are covered in a separate document in a more holistic manner.

[Bruno2] On my side, speaking about holistic manner, I would a priori have a preference for the document defining the anycast flag to cover the anycast properties in an holistic manner.

KT3> I understand your point of view. My view is that, the way existing RFCs stand, we cover only the base protocol semantics of anycast in this document and cover the overall SR anycast aspects in a separate (SPRING?) document such that it also covers those aspects for ISIS and OSPFv3.




Interestingly, the conditions seem different…
Authors seems to use RFC9352 and RFC9513 as a justification. I’m not familiar with OSPFv2 but my understanding is that it does not advertise SRv6 SID. So presumably there are some differences

KT> I hope the previous responses clarify.




“The prefix may be configured as anycast”
Putting the burden on the network operator is not helping clarifying the semantic. We need the receivers/consumers and the network operators to have the same understanding of the semantic. (not to mention all implementations on the receiver side)

KT> I hope again the previous responses have clarified.
[Bruno] Not yet. Cf my first point about an operation considerations section.

KT2> Ack for introducing operational considerations.




So please specify the semantic.
This may eventually lead to further discussion (e.g., on SR-MPLS)

KT> That discussion is important and we've had offline conversations about that. However, IMHO, that is beyond the scope of this document and this thread.
[Bruno] Too early to tell on my side.

KT2> How about now? :-)

[Bruno2] I’d say this discussion in this is in scope of this document. Another thread works for me. I picked that thread as I don’t usually read OSPF documents but have been convinced by Tony P.’s argument.

In summary, I understand a bit more the point of view of this document. But I’m still concerned that different implementations could have a different reaction to this flag. For a link state protocol this seem possibly problematic.

KT3> OK. Let me take a step back. The Anycast property of the prefix has been defined for 2 of the 3 IGPs - this document is covering that 3rd IGP. As authors, we have already shared the various updates that we have agreed to make to the document to clarify the semantics of the anycast property of a prefix in OSPFv2. We will continue to incorporate WG inputs should the document be adopted. However, as co-author, I do not agree that it is in the scope of this document to delve into the use-case (they are informative examples and so will be very brief about them in this document) and the document should also not delve into the broader SR anycast aspects. That later discussion belongs in SPRING. I will leave the adoption of the document with that proposed scoping to the WG decision.

Thanks,
Ketan


Thanks
--Bruno

Thanks,
Ketan


Thanks,
--Bruno

Thanks,
Ketan


Thank you
--Bruno

From: Lsr <[e-mail]<mailto:[e-mail]>> On Behalf Of Tony Przygienda
Sent: Wednesday, March 20, 2024 5:44 PM
To: Acee Lindem <[e-mail]<mailto:[e-mail]>>
Cc: Ketan Talaulikar <[e-mail]<mailto:[e-mail]>>; Dongjie (Jimmy) <[e-mail]<mailto:[e-mail]>>; lsr <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>
Subject: Re: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06

I think the draft is somewhat superfluous and worse, can generate completely unclear semantics

1) First, seeing the justification I doubt we need that flag. if the only need is for the SR controller to know it's anycast since it computes some paths this can be done by configuring the prefix on the controller itself. It's all centralized anyway.
2) OSPF today due to SPF limitations has a "baked-in weird anycast" since if prefixes are ECMP (from pont of view of a source) they become anycast, otherwise they ain't. I think the anycast SID suffers from same limi8ation and is hence not a "real anycast" (if _real anycast_ means something that independent of metrics balances on the prefix). Hence this draft saying "it's anycast" has completely unclear semantics to me, worse, possibly broken ones. What do I do as a router when this flag is not around but two instances of the prefix are ECMP to me? What do I do on another router when those two instances have anycast but they are not ECMP? What will happen if the ECMP is lost due to ABR re-advertising where the "flag must be preserved" .
3) There is one good use case from my experience and this is to differentiate between a prefix moving between routers (mobility) and real anycast. That needs however far more stuff in terms of timestamping the prefix. pascal wrote and added that very carefully to rift if there is desire here to add proper anycast semantics support to the protocol.

So I'm not in favor in adopting this unless the semantic is clearly written out for this flag and the according procedures specified (mobility? behavior on lack/presence of flag of normal routers etc). Saying "

It

   is useful for other routers to know that the advertisement is for an

   anycast identifier.

" is not a use case or justification for adding this.

if this is "anycast in case of SR computed paths that are not ECMP" then the draft needs to say so and call it "SR anycast" or some such stuff. If it is something else I'd like to understand the semantics of this flag before this is adopted.

-- tony




On Wed, Mar 20, 2024 at 5:10 PM Acee Lindem <[e-mail]<mailto:[e-mail]>> wrote:
Hi Ketan,

On Mar 20, 2024, at 12:07, Ketan Talaulikar <[e-mail]<mailto:[e-mail]>> wrote:

Sure, Acee. We can take that on :-)

I hope it is ok that this is done post adoption?

Yup. I realize this is a simple draft to fill an IGP gap but I did ask the question below. Hopefully, we can get to WG last call quickly.

Thanks,
Acee




Thanks,
Ketan


On Wed, Mar 20, 2024 at 9:35 PM Acee Lindem <[e-mail]<mailto:[e-mail]>> wrote:


> On Mar 20, 2024, at 11:17 AM, Ketan Talaulikar <[e-mail]<mailto:[e-mail]>> wrote:
>
> Hi Acee/Jie,
>
> The most common users of the anycast property of a prefix are external controllers/PCE that perform path computation exercises. As an example, knowing the anycast prefix of a pair of redundant ABRs allows that anycast prefix SID to be in a SRTE path across the ABRs with protection against one of those ABR nodes going down or getting disconnected. There are other use cases. An example of local use on the router by IGPs is to avoid picking anycast SIDs in the repair segment-list prepared for TI-LFA protection - this is because it could cause an undesirable path that may not be aligned during the FRR window and/or post-convergence.
>
> That said, since ISIS (RFC9352) and OSPFv3 (RFC9513) didn't have the burden of this justification of an use-case, I hope the same burden would not fall on this OSPFv2 document simply because it only has this one specific extension.

But they also weren't added in a draft specifically devoted to the Anycast flag. It would be good to list the examples above as  potential use cases.


Thanks,
Acee



>
> Thanks,
> Ketan
>
>
> On Wed, Mar 20, 2024 at 8:16 PM Acee Lindem <[e-mail]<mailto:[e-mail]>> wrote:
> Hi Jie,
>
> I asked this when the flag was added to IS-IS and then to OSPFv3. I agree it would be good to know why knowing a prefix is an Anycast address is "useful" when the whole point is that you use the closest one (or some other criteria).
>
> Thanks,
> Acee
>
> > On Mar 20, 2024, at 9:09 AM, Dongjie (Jimmy) <[e-mail]<mailto:[e-mail]>> wrote:
> >
> > Hi authors,
> >
> > I just read this document. Maybe I didn't follow the previous discussion, but it seems in the current version it does not describe how this newly defined flag would be used by the receiving IGP nodes?
> >
> > Best regards,
> > Jie
> >
> > -----Original Message-----
> > From: Lsr <[e-mail]<mailto:[e-mail]>> On Behalf Of Acee Lindem
> > Sent: Wednesday, March 20, 2024 4:43 AM
> > To: lsr <[e-mail]<mailto:[e-mail]>>
> > Cc: [e-mail]<mailto:[e-mail]>
> > Subject: [Lsr] Working Group Adoption Poll for "Updates to Anycast Property advertisement for OSPFv2" - draft-chen-lsr-anycast-flag-06
> >
> >
> > This starts the Working Group adoption call for draft-chen-lsr-anycast-flag. This is a simple OSPFv2 maintenance draft adding an Anycast flag for IPv4 prefixes to align with IS-IS and OSPFv3.
> >
> > Please send your support or objection to this list before April 6th, 2024.
> >
> > Thanks,
> > Acee
> >
> >
> > _______________________________________________
> > Lsr mailing list
> > [e-mail]<mailto:[e-mail]>
> > https://www.ietf.org/mailman/listinfo/lsr
>

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

____________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc

pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler

a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,

Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.



This message and its attachments may contain confidential or privileged information that may be protected by law;

they should not be distributed, used or copied without authorisation.

If you have received this email in error, please notify the sender and delete this message and its attachments.

As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.

Thank you.

____________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc

pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler

a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,

Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.



This message and its attachments may contain confidential or privileged information that may be protected by law;

they should not be distributed, used or copied without authorisation.

If you have received this email in error, please notify the sender and delete this message and its attachments.

As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.

Thank you.

____________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc

pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler

a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,

Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.



This message and its attachments may contain confidential or privileged information that may be protected by law;

they should not be distributed, used or copied without authorisation.

If you have received this email in error, please notify the sender and delete this message and its attachments.

As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.

Thank you.
____________________________________________________________________________________________________________
Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci.

This message and its attachments may contain confidential or privileged information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been modified, changed or falsified.
Thank you.