All,

Chiming in here from my travel to the IETF.

At the upcoming Friday meeting of the RTG WG, there will also be a presentation 
of the Routing on Service Addresses (ROSA) draft 
(https://datatracker.ietf.org/doc/html/draft-trossen-rtgwg-rosa). Given the 
emphasis on the routing aspect, this has been positioned towards the RTG WG, 
not the INTAREA WG, unlike the SAN work.

While I understand that there is an overlap with CATS (being myself involved in 
a number of CATS drafts), there are also differences in that ‘service routing’ 
does not require compute awareness (positioned as an anycast problem, you may 
see that an awareness of compute resource status is just one of possible inputs 
into the decision making, if at all). I would also argue that compute awareness 
does not require ‘service routing’ at L3/L3.5 either. However, CATS may 
investigate such solutions for use cases with requirements for high dynamicity 
of the (compute-aware) decision making, so I can see this fall under the 
‘groundwork’ aspect of the CATS charter.

However, given the above, I positioned the discussion with the RTG WG (for the 
ROSA draft) so as to discuss the argumentation for service routing (at the 
proposed L3.5) without necessitating a clear way forward for CATS in terms of 
the groundwork being done. The use cases in the ROSA draft are thus positioned 
wider, not focused on compute awareness alone.

One of the discussion topics on my slides for Friday is to receive guidance 
from the RTG WG chairs and the wider community as to where to pursue this work 
going forward but the focus here is on a solution for service routing, not one 
specifically that may be useful for compute-aware scenarios. It may ultimately 
feed into CATS, both as part of the ‘groundwork’ or as part of a possible 
solution if deemed necessary due to shortcomings of existing ones (although I 
am not trying to preempt the groundwork since that conclusion will need to 
drawn clearly). Further, some of its use cases and requirements may well be 
applicable to CATS, where there is compute awareness used as input into the 
routing decision.

I am happy to receive any comments, both on the draft specifically as well as 
on the above argumentation on its positioning.

Best,

Dirk

From: Cats <cats-boun...@ietf.org> On Behalf Of Adrian Farrel
Sent: 26 March 2023 07:18
To: c...@ietf.org; draft-zhou-intarea-aware-route-san-datab...@ietf.org
Cc: int-area@ietf.org; 'John Scudder' <j...@juniper.net>; 'Erik Kline' 
<ek.i...@gmail.com>; 'Alvaro Retana' <aretana.i...@gmail.com>; 'Andrew Alston' 
<andrew-i...@liquid.tech>; 'James Guichard' <james.n.guich...@futurewei.com>; 
'Eric Vyncke (evyncke)' <evyncke=40cisco....@dmarc.ietf.org>
Subject: [Cats] More on "Service Aware Network"

Further to my previous, I’ve now had a chance to skim 
https://datatracker.ietf.org/doc/draft-zhou-intarea-aware-route-san-database/

It seems to me that this draft is “routing related”, and highly related to 
CATS. In particular, the overlay of “service routing” with edges and relay 
nodes is leveraged through knowledge of “compute resource status” injected into 
the routing overlay network nodes.

Many of the issues addressed here seem to be directly overlapping with CATS in 
terms of what problem is being solved, what metrics need to be advertised, and 
how the overlay is built. While (obviously) this draft intends to leverage the 
solution set developed as “SAN”, it would be worthwhile synchronising the use 
case and problem statement with that in CATS.

One of the jobs of CATS, in the medium term, is to evaluate existing tools and 
mechanisms to see what can address the CATS problems and where the gaps are. So 
we would certainly welcome learning more about SAN and the use cases that the 
authors hope to address.

Thanks,
Adrian

From: Cats <cats-boun...@ietf.org<mailto:cats-boun...@ietf.org>> On Behalf Of 
Adrian Farrel
Sent: 26 March 2023 05:57
To: 'Eric Vyncke (evyncke)' 
<evyncke=40cisco....@dmarc.ietf.org<mailto:evyncke=40cisco....@dmarc.ietf.org>>;
 c...@ietf.org<mailto:c...@ietf.org>
Cc: int-area@ietf.org<mailto:int-area@ietf.org>; 'John Scudder' 
<j...@juniper.net<mailto:j...@juniper.net>>; 'Erik Kline' 
<ek.i...@gmail.com<mailto:ek.i...@gmail.com>>; 'Alvaro Retana' 
<aretana.i...@gmail.com<mailto:aretana.i...@gmail.com>>; 'Andrew Alston' 
<andrew-i...@liquid.tech<mailto:andrew-i...@liquid.tech>>; 'James Guichard' 
<james.n.guich...@futurewei.com<mailto:james.n.guich...@futurewei.com>>
Subject: Re: [Cats] FW: [Int-area] Intarea WG agenda

Thanks Eric, that’s helpful.

The agenda slot is:

Tuesday, March 28, 2023

17:00-18:00 Tuesday Afternoon Session IV

The agenda item and related drafts are as shown below.

4. Service aware network (SAN) framework and data plane solutions  - Daniel 
Huang
draft-huang-intarea-san-framework-00
draft-ma-intarea-identification-header-of-san-00
draft-ma-intarea-encapsulation-of-san-header-00
draft-zhou-intarea-aware-route-san-database-00
draft-zhou-intarea-computing-segment-san-01
    20 mins

The work certainly seems related, but a quick reading suggests that the 
approach chosen is to mark the traffic according to the service and then route 
according to that marker. This seems (to me) to be closer to APNET proposals 
(https://datatracker.ietf.org/wg/apn/about/) where a charter is currently being 
developed and discussed with the IESG.

I would certainly encourage the authors with their work, but also I would 
invite them to try to contribute to existing work efforts and not create new 
ones where there is overlap.

Cheers,
Adrian


From: Cats <cats-boun...@ietf.org<mailto:cats-boun...@ietf.org>> On Behalf Of 
Eric Vyncke (evyncke)
Sent: 26 March 2023 05:35
To: c...@ietf.org<mailto:c...@ietf.org>
Cc: int-area@ietf.org<mailto:int-area@ietf.org>; John Scudder 
<j...@juniper.net<mailto:j...@juniper.net>>; Erik Kline 
<ek.i...@gmail.com<mailto:ek.i...@gmail.com>>; Alvaro Retana 
<aretana.i...@gmail.com<mailto:aretana.i...@gmail.com>>; Andrew Alston 
<andrew-i...@liquid.tech<mailto:andrew-i...@liquid.tech>>; James Guichard 
<james.n.guich...@futurewei.com<mailto:james.n.guich...@futurewei.com>>
Subject: [Cats] FW: [Int-area] Intarea WG agenda

For information to the CATS WG as the INTAREA WG agenda includes individual 
drafts about "service aware network" (SAN). At first sight, those SAN drafts 
look pretty close to the CATS charter.

So, feel free to join us in INTAREA WG meeting (or on the mailing list).

Regards

-éric

From: Int-area <int-area-boun...@ietf.org<mailto:int-area-boun...@ietf.org>> on 
behalf of Wassim Haddad 
<wassim.haddad=40ericsson....@dmarc.ietf.org<mailto:wassim.haddad=40ericsson....@dmarc.ietf.org>>
Date: Sunday, 26 March 2023 at 13:07
To: Internet Area <int-area@ietf.org<mailto:int-area@ietf.org>>
Cc: "intarea-cha...@ietf.org<mailto:intarea-cha...@ietf.org>" 
<intarea-cha...@ietf.org<mailto:intarea-cha...@ietf.org>>
Subject: [Int-area] Intarea WG agenda

Hi,

The Intarea WG meeting agenda is available at: 
https://datatracker.ietf.org/meeting/116/materials/agenda-116-intarea-01

Presenters, please upload your slides or send a copy to the chairs.


Thanks,
Wassim H.
_______________________________________________
Int-area mailing list
Int-area@ietf.org
https://www.ietf.org/mailman/listinfo/int-area

Reply via email to