RIFT minutes:

1. RIFT update
Tony Przygienda (Tony P) presenting

Bruno Rijsman: (Jeff Tantsura read) On slide two it is actually safe to reject a packet based on field after the fingerprint before validation the fingerprint.

Tony P: Yes, but there is no need unless we we will do a little suggest that I rejected which is protecting something like rift major because I see absolutely no benefit but then you look for the fingerprint right now there is nothing new for look before fingerprint that makes any sense.

Bruno R: (Jeff T read) For example, it is safe and efficient to reject a pack and because of our differential comes before validating the fingerprint, or however, if you accept the packet, you must validate the interruption before taking any practical action.

Tony P: Okay, so that may still need a small grind because Bruno's absolutely correct You can add shoot look at the nuances before the fingerprint because they could you can rejected without being exposed to computational attack, because the fingerprint of computation is and this is an envelope, where if someone forces you to compute a lot of envelope damages basically as a denial of service attack surface against your computational resources. Bruno is correct.

Pascal Thubert: So I was just thinking about discussion about all the ipv4 or ipv6 address adjacency Yeah. So you we can explore the world that you pronounce which are correct on the document, but the resulting RFC5549 which which explains all that.

Tony P: So at this point optimal link?

Pascal T: Yes. RFC5549 is that you have adjacency and you can solve my address and you can forward.

Tony P: perfect because that was what I was looking for. I was only finding the seven series which was all about tumble link.

Pascal T: ?????

Tony P: was a text is not explained about the router requires we're kind of making it implicit right? But it will be great to have to five series reference where we basically don't even have to explain how you do it okay yeah because because I refuse to put it to respect like I'm not here to describe how to build a forwarding path right.

Jeff T: so this is specifically to avoid definitely that?

Tony P: Yeah.

2. Python implementation update
Tony P presenting.

Bruno R: (Jeff T read) for Linux does support a net cost multi pass and it is he has a shot to 256.

Tony P: So it’s 256 wonderful and we do have to support in lineups which means if you get that as a host version, you will actually see that it will load balance based on bandwidth on the fabric not about. 

Bruno R: (Jeff T read) that is published security report that has other results.

Tony P: Yes, so there is security review. I didn’t even see that. I’m not sure what he’s talking about with the publishing okay.

3. draft-wei-rift-applicability-01
Sandy Zhang presenting

???:(from Juniper) Just one comment on the previous slide. Do you want to also add some statement about scalability a number of adjacencys? Which could be a plus?

Tony P: So Yeah, a number just in syrup. It's important. He's logic interpretation specific, but the scalability of the protocol the based on the fact that he summarizes up right. And you have much less information also the blast radius. That is our official report. That's a little bit further right.

Alvaro Retana (AD): Now that you mentioned working to adoption, I wanted to mention that What the Church says is that one of capability statement that talks about how employees configure different policies. What I think this document does, is it basically summarizes the features, And then it says, oh, because it has those features, I can use them fabrics.

Sandy Z: There is many texts in the use case parts.

Alvaro R: But It looks like there's maybe one or two paragraphs for each of the use cases. I don't know that that's a lot. I don't think that's a lot. And I don't think that one or two paragraphs, you can talk about how to deploy and configure for each one of these cases. So again, with a grain of salt, I haven't read the whole thing. But It seems to me that this draft in the current form is not what we want to meet the charter. This sounds more. And I'll take this the wrong way. But this sounds more like a marketing summary. This is all the great things about Rift, which is great. But It doesn't tell me if I need to apply rift to a fabric to the fabric for example, what considerations Should I take what I configure the plan? Why or whatever considerations are?

Sandy Z: So we will add more details in the use cases. Right? Okay.

Alvaro R: I think there should be.

Jeff T: something good message to quality to work to clarify use cases to increase number of use cases.

Sandy Z: Okay.

Tony P: So you have five use cases of making the work to be stern magic number we have to cross?

Sandy Z: We will add more detail about it.

Tony P: No, it's more like question to up into ADs versus over the what's wrong with five. And the second one, we are under the curse over our own design. Because if you run the risk of ZTP is literally the configuration cover with the one sentence which the whole for the fabric on the top of the fabric and you're done.

Alvaro R: It So there's nothing wrong with five. This wasn't one with seven, This falling under one with one. So what we want, or what I thought we agreed from the beginning, but we did the charter with the statement is to say okay, now here we have the protocol specification. This is how you use that you just said something interesting, which is well, this is so easy. It's just a few right here. And then that's it. So maybe there aren't that many considerations. We specifically mentioned this in the charter. Because you know, I want to make sure that when people are going to use, So when people are going to use RIFT that they know how to deploy that if there are the considerations that people consider them. And so that, you know, I don't know if I need to think about the size, or whatever it is. That is to say that just like security, You could have a section in your in the big documents, which is already 140 pages. To say these are the deployment observations. There's an RFC that I forget which one it is, In the 5000 series, I'm going to say 5143 But whoever's the fetch fact checker, you can check that talks about the deployment and management considerations. And so it talks about different putting a new protocol what kinds of things we need to think about in the network. If I'm doing something that's the type of thing they want to see. It may be that feel approval is so simple and so easy to use that You'll really have considerations and that's fine as long as you explain why not just say great. We want to do something with it.

Tony P: So Yes, I remember that are still in debt. So thanks for this reference. Yes. required to rethink where we had deployed consideration was basically Miss cabling right that's where the stuff pops up on the security this stuff I saw with Bruno was more implementation considerations when the way the world works he had to lose a few stuff otherwise, keys are keys well you can put something in like you use OSPF and ISIS is key management same thing I mean, if that contributes anything And we have to configuration which is an embarrassingly you know little stuff to say really about and I saw one thing where I self employment considerations really as review really want to place right so I had something about the loop better word culture the loop use parts of the fabric.

Alvaro R: I'll just for the record or six minutes for the RFC706.

Dmitry. Afanasiev: As one of the co-authors with draft Yeah, at least some parts could be more detailed sort of a me because analysis was done later. So it's not in this job but by the particular DC properties if you want to do though for the cause and keep them out of out of state small Okay there is a an RFC how how to drought in in DC using PGP but becomes really really really difficult to was minimized state and to the Holland for example, because it essentially requires doing conditional desegregation based on non local information of the very difficult to do without a lot of additional logic.

Jeffrey Zhang: Okay, so Exactly We need to have I guess, one more revision.

4. 
Tony P: Alright, so here’s a question to AD, so do we do we have to last call everything in a package. What's the deal also with our era. Protocol specification is very thoroughly cookie, We're doing donations. And I have people asked me, when will they have RFC for the art piece? And they're like, for Easter policy here all week. Can we go to the last call the protocols back one older stuff, of course is absolutely necessary committed to.

Alvaro R: So You said What's up for those? Yeah that's a different working group to have last call everything. I think when we did the milestones, We said that we would process together the protocol and security, which has now been documented. Yeah. So Now that means that you notice, we have the last couple of things. So we can progress protocol. Now again, if for example, in the case with this capability statement, which we put later the the milestones, If this is going to be folded into the actual protocol spec, because it's so simple, then, Yeah. But otherwise, we should get it. Don't make promises on the time frame for gamers.

Tony P: Oh, I didn't reply to your he'll deliver a product on time. It's called it just allowed to try to get the product out. Okay. Thanks.

John Scudder: For the suggestion, mostly to the chairs. Jeffrey, I think what I heard models and reading my email, which of course wasn't doing during the discussion of the applicability draft was for some dependability, draft. And then there were, you know, some criticisms of it, whatever. And then you say, well, maybe it's not working for it out yet. And, gosh, it sounds to me, like the working group is discussing a draft that the working group is trying to produce. Which sounds to me exactly like what working group is supposed to adopt a draft to do? Like, normally you don't, you know, wait until the draft is perfect. Before adopting it you adopted? When you're like, yeah, this is kind of what we're supposed to be working on. And then you insisted to be perfect at last call time, but it's your group, I mean, run it the way you want to.

Jeff T: I think the intention was just to address I don't recommend this being another religion and not to delay for another six months. Thank you.

5. WG status & plan
Jeffrey Zhang & Jeff Tantsura

Pascal T: We have discussions we have two big slight very mobile devices you don't want to go state and intermediate layers which is wanting to the status of us either turn off from Spine to the leaf very mobile carriers So so we have things like that which is very special cases. We don't know if it's useful. So Yes, the ages of floating here and if we are ready to realize that even if you don't need it, right, so just unclear that you have a need for it mobile or highly scalable networks were maintained extend state and intuitive stable would be inefficient. discussed it we go to do it, which was developed.

Tony P: yeah, So all these things are kind of the back burner, because from my side I see people being entertained by and I asked the way that or are willing to talk different scenarios. I just don't see use cases I mean my case like SR, right, there's loop back. So talk extensive loops. And I call in deployment considerations, one or two things there, like, Watch for this in very anomalous scenarios. But for example, SR, was leave to leave and then deactivated your tunnel anyway. So why would you even need that young. But it's a good that's why he's anxious each step along and care because the simple enough people pushing for having deployment needs for discovery for the month of helping of servers is has a lot of interest driven by hardware coming and people like realizing that if you can monitor all the servers, lots of problems you're solving today in weird ways to simply disappear, right.

Jeffrey Z: Yeah, I understand the way there are things that we talked about before starting to work, some work before we have not really focused on them means that they are not possible. So just some things on the list, we can decide what to do.

Jeff T: So ideally, we should keep the documents alive. If later on the meter arrive. There's some fundamental workout. My second drug import can use some other things.

End.