 Hello, hello Hello, go ahead and paste the agenda Can everyone open that? Yeah, I got it. Yeah We'll give everyone a couple more minutes to join. I know my last meeting was running over as well. So Hey me That's okay The next yeah Is everyone that's just joining able to see the agenda in the chat or should I I can't see the agenda One more minute and then we'll get started at five after go ahead and share my screen can everyone see my screen Yep. Yeah All right. Good morning Report Alex or have you already started it looks like it's recording perfect. It's recording. All right. So good morning This is Wednesday, November 13. This is the CNCF storage. So So on the agenda As Hopefully everyone knows by now. We had pretty sad news Brad Childs who was one of the Cotech leads of the CNCF storage SIG and one of the co-leads of the Kennedy storage SIG his Passed away. So if you would like to record a video memory for Brad that will be given to his family There is a link in the agenda that first item Is the sooner you can do that the better and I believe and so I can correct me if I'm wrong that they're planning on Playing like a montage of these on the stage as well at KubeCon so It'd be appreciated Yep, that's right You guys so yeah, we're looking forward to any videos that you can share so the sooner you can do that better. Thank you And in addition to that, I will also post this later in the agenda There is a Google Doc in the CNCF where you can also just add in some You know memories of Brad if you'd prefer not to do a video and write something else instead or do both So the first piece of business is the use case documents that we've been talking about for a while that Louisa's put together Can you guys see that Yeah, yes If you click on the instead of clicking on that click on the my branch there right above that use case that's yeah, that way you can actually view what it looks like and then Go down to use cases On the on the top. I don't I haven't changed this part yet. Oh, you haven't okay. Yeah, just the directory use cases so What I've done here is kind of taken your feedback from the previous meetings Had it in a way that it's simple to add simple to edit simple to track versions Simple to remove and simple to add new ones and keep it a Keep it concise for those who are new So what I tried to do is I created a small template that's right there and I created a sample use case Using the template so we can click on the template first and you can see how simple it is and I took the top part from Kubernetes Style of doing caps. So this is just an idea every again. This is a straw man. I'm not Saying this is the way it should be but it is just a start and If you guys feel this correct direction, we can continue going this direction and we can give it in another direction So I took the cap title Area like if you click on raw here then you can see that area looks it's just taken from from the caps in Kubernetes and It provides way to track it and know the state of the document and things like that and then the document is just regular markdown The document the way I have Structured is to make it simple for readers and then refer back to the storage landscape document in any any Words or values that we want them to to to learn about and and So it's just pretty much table-driven and then I can show you an example of what it looks like if we go back one So if you go back another because this is a template and Then go down and Yeah, you can go like that. It's fine. Yep, you go on. So this is what it looks like Here's a very small example of how to do an object storage NAS gateway using Minayo This is a small example of how to deploy Minayo and have The storage back in for it in the use case is as it describes here I'm modeled for the requirement of accessing files over a shell file system and an object stored at the same time Okay, so for that use case you can go down and it says what the goals are For the small document, but you can if you go scroll down a little bit like you go so here you can see the Storage landscape summary It tells you that's recommended to use those it tells you if you use one of those This is what you should check for if you use a block store and make sure you have a file system on top And you also need a service to be able to to share that If you go down a little bit more in file systems It tells it says you look if you have a remote file systems It's actually really nice because the storage system already has a model for for clients to access and It has a model for you to go from node to node to connect and disconnect So so pretty much this is like You know and then again, I said look object stores and key values that there they just don't work in this use case So this is the straw man If you can notice also things like durability and file systems and other stores are all links They all refer to that section in the Landscape paper so as a user that I'm not very familiar with storage and I want to click on file systems I can click on that and it puts me right on that page on the storage landscape so I can read that more information about it And again, the goal is to make it concise I didn't want to go into too much detail and we can add more detail, but it is just a use case recommendation What do you guys think I think it's excellent? I Love having the tables to be able to concisely determine where it fits and where it doesn't I think that helps I Think that helps if we're able to like layer on as well be able to take these type of things Side-by-side that's been a big discussion on the TOC like We have so many projects. I don't think that's true storage, but you know, where do we see the landscape going and where do we have gaps where we To foster projects that fill the gaps And again, I tried to make sure I used the same words that's a landscape document So the last gate document is like store stacks section block store section I just referred to those so I kind of applied that document to this use case in other words so So I think this is this is This is really good. It's kind of really simple. I like the fact that ties into the landscape document What do you mean in the storage stacks section? So Things like centralized distributed charted. I think those were topologies right in the last case. Yeah, what do you mean by recommended? Yeah, it's I'm not I'm trying to not say that it's not recommended. I Guess that's the word This like is a binary thing. Maybe we cannot make it binary make it three meaning like it doesn't matter Um, right But right now it's binary, but it really doesn't matter. It's How the story stack does it because it's it's more of an a protocol front, but you're right I mean, maybe maybe instead of just recommend it and not recommend it. We could have three states Like it it and a like I'm not applicable something like that Yeah, yeah, all right Is that is that good like this? Yeah, yeah, no, I I really like that I kind of like the idea you're right of saying You know, you can use distributed or pipe coverage for example But the optimal is might be something right? Yes, exactly. Yeah All right, and to me Do we want to go? Do we want to provide An example in the use case So like Would it be worth saying for example, you know, you know if you want to start many or with a Storage class this is how you do it for example Yeah, I put it as a non goal that oh this doesn't show a step-by-step, but at the bottom of this page. There's a reference Okay, I saw I saw the reference to the to the Minio documentation, right? Yeah, and it pushes you and tells you exactly how to do it. So Yeah, right there or the references the bottom Yeah, I have a bit of a concern here, which is that This could be interpreted as and this could become a list of kind of product or project promotional pages You know we in this particular example, we kind of seem to start with the assumption that Minio is the solution and then essentially justify it and it Sort of doesn't feel to me That's the flavor that we you know understood the problem the requirements and then You know Spelled out a bunch of different ways to do it With pros and cons of each way Which would seem like the sort of balanced approach to Something like this or So the good question and I felt the same way as I was creating it So I I completely agree I feel and this is How I balanced it and I want to get your opinion because I can go either way One of the things I was thinking of is as a customer as a user I want to deploy I'm looking for you know exactly what I want to deploy. I want to deploy out this use case and and We could in at this is just a very simple web page here I knew the minios are used it as an example, but we could you could do the same thing with Seth or Rook And you could create other ones other pages just like it So so it doesn't in nowhere on this page For example, I put that this is the minio is the recommended model. I instead say Here's an example using minio and what it requires, right? And so I try to be clear about that and that gives the opportunity for others To add there to their own pages to github, which their recommendations So that that's how I balanced it What do you think yeah I don't I don't know that a catalog of of you know how to use is is necessarily as useful as the Thinking process by which you take a specified well specified use case and you then Analyze what the requirements are and different ways that you can do it Sure, but we the key thing is that the customers are looking for a specific thing, right? Before they could decide whether they were interested in running minios and as gateway They have a whole bunch of other homework to do to figure out whether that's in fact useful to them Okay, so let's step back a little bit and think This is just an example. I just I'm not need to push this one, but I'm trying to think the goal was to say I Want to deploy Kafka for example? Right now I'm going after a specific project Do we want I mean trying to go back to the root of this product this task, which is do we want to be like that? Do we want to say? You are deploying Kafka as a use case. Here's the recommended storage for that In other words, here's our recommended app And here's the storage for that and I chose the app to be mana minio Now do we not want to do that? Do we want to be really really generic? It's like if you deploy a Like a message bus app Do you want to use that's one? That's my concern So so so my two cents here. I mean many of us possibly a bit something because it kind of It sort of sits in the storage space as well as being a use case rights that can consume storage, but I Think what we will be really want to capture out of these use cases is Different applications that consume storage and Link that to what we have in the landscape looking to say this application These these type of topologies this type of storage is is optimal or recommended for them And we may want to I mean may want to sort of maybe augment it and have maybe a notes category or something in this in this example template, but That's we can talk about Because in the landscape, but can we have the attributes Like failover and scalability and performance and that kind of thing And make them consistency durability whatever So so maybe we can say, you know, we have a section with notes to cover some of those attributes to be added to this So so it would be a use case like Kafka perhaps or or a database or a message queue or you know, whatever that's they full Use cases that that is consuming storage And we say, okay We have to we have to you know, it is it isn't should be fire system should be local should be remote Should it be a distributed topology or centralized whatever and we and we list all of those things and then we also say for example Say for Kafka for the sake of the example, you know, we can say Kafka is Dependent on lots of sequential IO performance for example You know, and if we do another one for rapid mq we could say, you know, it's it's It's dependent on on lots of random IO and you need very strong consistency or whatever, you know, those those kind of things So that's we can we can the idea is specifically to allow users to better understand how to deploy their stateful applications Exactly, right. I mean, that's that's the model I'm going after and I know that We had It's a kind of a double-edged sword because we're talking about projects as applications that consume storage But they're also promoting promoting those projects. So we have to be careful Again, I this is a straw man, I want to really get your opinion on it. So I completely understand Quinton, what you mean and I'm trying to I don't know how to Satisfy the requirement of of helping customers and users at what the same time being ambiguous I Don't know. I'm thinking about it. So I'm not disagreeing with you is what I'm saying I think we touched on that a little bit last time like if we're so vague This has value, right? Like if we're just Just mentioned the general data structures and put it into a document. It doesn't really help people get started or understand Things there is an issue of these becoming stale like there there could be Something a different offering than many. Oh, I mean, I think we just We like all open source things we would open it up to different use cases that also address the same thing I mean, that's all we can do. That's the CNCF is, you know, landscapes meant to be rich and diverse and have over that And we can also reference each other in the documents when they when they have this thing use case different technologies we can put them all in the same group Or linked to each other so our users can see the other models with other projects So I agree that we want to make this as useful as possible And and you know having it very generic is it makes it difficult to be useful But at the same time I think I agree with Quinton we have to Be aware that, you know, the CNCF can be used as a platform for promotion and as soon as a product is mentioned anywhere It incentivizes anybody who is a competitor in that space I completely agree, yeah Okay, name, you know everyone is gonna want their product name there Right, but but just to clarify right we're not We're not Specifying different storage products here, and we're not saying if you want to run Kafka for the sake of the argument use this product what we're saying is These are the attributes you're looking for out of the storage system, and these are the things that's that's that she needs to That you need to consider for that use case right now The idea of doing anything get up was also a case of right if you can you can now look at Kafka and you can say it benefits from a local file system with strong performance or whatever And then that allows them to look at the storage landscape and make their own decisions as to what storage system They're gonna use or what cloud service or a project or whatever else they're gonna use But we're not we're not actually recommending a project per se we We're trying to We're trying to give people advice on what how to move their stateful applications or how to use stateful applications In these environments and I completely agree with a I agree with the intention of this and I understand that the intention is not to pitch a product it is to Explain a use case and then use a product as necessary as an example But I think even with that You're going to have folks who are gonna say hey, why is my competitors product named here? I want to update this example to include my product. That's not fair Let me in and so I want to make sure that we kind of have a plan in the story for if that happens How do you deal with that? I think there's two more two answers to that One is that it's on github so anybody can edit And we can go through the process they can create they can create their own also so we're not it's not like When when we go through the process of accepting a project into the CNCF we have to go through certain Things but creating a small page on github. That's very very easy and very low Very low ramp. The second one is that It could be that instead and here's another way of doing it instead of us creating this Pages and hosting it here, which could cost the issue is that we create the template and we let projects host them on their page on their web pages and they manage that we just kind of promote the the the template and let the projects then posted on their pages, right and They think of refer back to the landscape document just like this github documents doing Then that is not up to us. We don't we don't host anymore It's another way Yeah, I don't necessarily think hosting is is the problem in my opinion in my opinion the function of this group is to provide unbiased balanced and useful information to primarily to the user community as one of our functions to have other functions to and I think that Doing that focused on One yeah, maybe maybe I'm out of line here. I mean I'm prepared to It it just doesn't feel like we're fulfilling the need here The need is, you know, I want to build my own Object store on top of my file systems What what options do I have and and what are the pros and cons of the different options? But that to me is what we were trying to do And correct me if I'm wrong And then I can come to a page like this and says oh, okay. I've got you know at least these options and and most people use this option and But one of the problems is X and and if X is a problem for you you also have option Y and Z They have these other downsides So so this is you know narrows your choice from the whole world is my options to these are actually the Commonly used ones with a pros and cons I can I mean a much better position to make a choice and then maybe I can click on a link That gives me a recipe for doing this Yeah, I think that What I meant by another repo means like we don't Like we don't put it up here, but like we don't advertise it But I think you're here. So an idea with I think following what you said is that in when you write technical papers you on Towards the end you have to write like When you write patterns and things like that you have to write a section of what other things are out there that are similar We could require that when you create one of these You know, there's a section you must have a section button that describes like what other projects are similar to this one, right? So that way every page has is not towards one project, but it has links to others if they provide similar solutions Can I can just clarify what the sort of what you're considering or what the what the objection is? Primarily, right? Is it is it's the listing of the options, which is the problem or the listing of the use cases which is the concern? I think my concern is mostly around having any sort of product names Mentioned in anything that resembles a recommendation or by the CNCS and the reason for this is kind of the history that we had if you all recall the Fiasco with when we try to define what CNCS or cloud native storage meant Turned into a giant brawl. I remember it's different vendors. So I'm very sensitive anytime We have anything that looks like it would potentially cause conflict amongst vendors I just want to avoid avoid that or have a very conscious plan when that starts to happen Maybe a good way would be to put the use case instead of object store NAS gateway using men IO would be Object store gateways and then So it would explain use cases for why we would want to use an object store gateway the topology architecture advantages of that and then under references, yes, it would be up to the author to provide and Examples know we could put reference architectures or example references or Or do we just expect the user to go out and Google those? I thought about that. I think just Sorry, I just I thought about that Erin and the problem was was that in each Implementation of the gateway was different. For example, if I compare mean IO to and they get I view this as an application Consumed storage, right? If I mean IO to open sex with to rock They have different ways of requiring storage that and then they all have the same They all satisfy the requirement of accessing the file and an object. So If I would suggest is that I would write all three, right? Using men IO using rock and and using open sex with for example, because they have different requirements Does that make sense because it opens a swift you don't want I think I get where you're going in order for this to be useful You kind of have to be application specific but at the same time I do think that Products are responsible for providing their own documentation for how to deploy in Kubernetes etc etc and What we want to capture here is not necessarily. Hey, here is how you integrate with a specific product but holistically here is a Product category and here's generally architecturally how it works and then if you're interested about a specific product you go look at that products documentation. I may be okay with what Erin is suggesting with like a references link at the bottom Where you could have specific examples linked out maybe two specific product pages But even that seems a little a little it has a smell to it So can I can I can I just quickly ask a question here, right? Are we are we getting ourselves tied up in knots because the example? That we happen to choose first was happened to be storage related Would we feel this way if we were talking about I don't know how to deploy a SQL database, you know of your choice my SQL or Postgres or whatever on and what type of storage they would require Would that change the thinking at all because I kind of do feel that it's We There are lots of things that consume storage obviously And we went to great lengths to define all the different storage types and how they work and You know pros and cons and things like that but what we always said we were going to do as the next step was Define the use cases to kind of an examples of How those how real life Consumers of that storage Whether they're databases, whether it's a message queue whether it's a stateful application of whatever type is Consuming the storage now in this particular case We kind of chose an ambiguous example because many always obviously both a storage consumer and a storage provider So so I can I can understand the why here But if this was say, I don't know a database or an example of how to configure a message queue on On storage platform with that have the same concern. I Think you've hit the nail on the head Alex I think if we treated the use case here as building an object store On-premise it looks like is is kind of the assumption Then and and men I always was listed as one of the options that was available if you wanted to build an on-premise object store I think that would be totally fine and and the same goes for the database stuff I think I'm probably less adverse to to mentioning products than Saad is and and I understand sensitivities But I think to be useful. We have to be comfortable saying this is the most commonly Used, you know, I think we should be able to say my sequel is a commonly used relational database And so is Postgres and there are also these other ones that are less commonly used for example but I think There are a whole bunch of things that are common whether you are Deploying Postgres or my sequel or any other, you know of those style of databases Regarding what discs you store your data on how you set up your caching All of those kinds of things and I think they can be dealt with Generically and and I think we can say the pros of putting your, you know log file Your database logs on on ssd or the following and the cons of the following And this is the common way of doing it for performance and if your ability or cost is there more important This is the common way of doing it. I think I'd be totally fine I think the fact that Minio is in the title here is is actually the problem as opposed to focusing on the use case Which is building an object store on-premise Okay, so should we just remove the Minio from the title? Yeah, that's what I was saying. I think it should just be Okay, but here's my concern and again, I'm just being playing devil's advocate It's just that even the sequel model sequel is just a language. I mean we could talk about You know Yeah Yeah, it's just we have to be careful like we could use like Vitesse and now we'll have a very different storage model Than the my sequel and a single note We just have to what I'm trying to say is that sometimes we're gonna have to bite the bullet sometime and and choose All of these also all these databases are persistence so they're all storage products themselves even in our landscape so I I don't know if we're just worried about just Minio specifically or Because if we start going towards databases, that's still persistence and then we're gonna start picking even message cues of persistence It I was thinking actually of using at CD for example as an example as use case But what about console, you know what I mean, so it's like it at CD I think it's a project for CNCF. I think so anyway No, I We have to find a good balance like I mean We would obviously want to highlight the open source ones that are part of CNCF, right? I mean if they're there, we should be able to put them here as a viable option But it would be nice if if the use cases could be put in such a way like why do I need object storage? Why would I want to use a gateway like it explains? I would want to do that Which I think you you've done and then the more specific pieces we would just leave to these references We wouldn't put those specific changes is if I did this in Nuba or I did it in Minio What would the differences be that's up to the user to go experiment what works best for them, right? And so that so then we would leave products specific under references We would open that up. We would let anyone add their reference there You know, we need a disclaimer probably like we do not, you know endorse any of these products These are provided merely as examples And then we're just providing guidelines of different ways that you might use storage based on your workload. I don't know but We can't be So one model I'm thinking Sorry Did I interrupt? No, no, I'm done. Go ahead So the one model I was thinking we can follow is you start with the cloud native CNC of the storage landscape document and See if we can extend that same structure here Where like in the storage document we have sections for object store key value stores Categorizations, so if we actually built the same structure here Then I think that what we can think of is High-level framework that describes how the trade-offs and then we can have Specific examples about Projects The one the one confusion I have is is Minio a product or a project or is Both all three So actually, I think that's where the I think it's the confusion is more specific in the fact that it is this one is all free Like so if you just took an open source project like safe But it is a product. I mean it is open source a party. It's licensed Correct. Yeah. Yeah, so I'm talking about like if you for example, Rook is Rook is also a product and so is safe. Oh Okay So So just to clarify though, right? We are we are not building storage use cases We are we are not going to build a catalog of all the storage projects and say This is the best way of deploying these stateful The these these storage systems. It's not about that if the use cases are End user use cases so in some cases there might be some ambiguity because somebody might be deploying a key value store like at CDP and You might want to have a use case which is how do you deploy at CDP? What is the important thing that you need that the storage? Well, you need the file system. You need strong durability. You need strong consistency, for example I think it's okay to say that and I think it's okay to say that's what that CD means And if and if somebody wants to write a use case for console, then fine, let them do it That's the whole point of having a community with GitHub, right? And I think the same thing applies to to other things whether it's, you know, message queues, databases, or any other type of stateful application I think somebody might say this is how you deploy, I don't know, WordPress on on the stateful application You want the shared file system or Or if you're installing WordPress, you want the database with this type of consistency or this particular application needs a key value store with these attributes As the back end. I think those are all perfect things to have here The what we're trying to remove is the ambiguity from the end users that they kind of look at this and they go, okay There are I look at the CNCF landscape and there are You know, 400 different options. How am I even going to where do I even start from right? And this is pointing them to something to say you have these attributes It needs to be a file system or it needs to be a key value store. It needs to be whatever And this is how you can do it. This is this is what we're trying to build here so I'm just trying to keep everybody on track because I Feel like we were having a circular discussion Discussed this five times or anything now put into the points where we're Where we're putting a template together and we're rehashing the same old discussion I So let's stop No, yeah, exactly what I was gonna say because to me when I six storage And again, I complete here. What's that and Quentin? I'm not I'm trying to balance it and Aaron. So I'm not At the nothing no to that. I'm just trying to come up with as a user What can I can can six storage kind of compile down to me for me? Something like a you know, like a card that I can just look at with bullets according to the the storage landscape, I want to do this I want to do the following and Hopefully by being in github Encourages community involvement I'm hearing it if this is what I'm hearing some I'd agree with it that a Document like this that describes storage use cases by class Object store is a class. It is not commingled with sequel based relational databases That would be a different class, but have a document per class Maybe inside that document you could link Examine at the end with links to supplemental documents maintain outside the CNC out that would Go into a deep dot a deeper dive on that particular variant of a solution So something like Minio would be referenced in the class document of Object store, but the actual Minio specific document is outside The CNC app storage takes realm So let me go back. I think we're getting stuck on Minio, but I view it as an app that consumes storage, right? Yeah, I do too But the reason I brought Minio up was just to use it as an example of an object store Okay The document should stick to the class of object store written in a generic enough way that what is said applies to all Examples of object store. So I don't think you'd argue that Minio is the only yeah Yeah, I completely agree and that's what I was saying before that you could add a Rook one you could add a Sep one as optic open sex Swift one They have their own models and down recommendations on the back in store So why don't we can I just be the Let's move on because we spent 46 minutes on this Okay, everyone be fine with removing the use case to mention a product and instead Requiring authors of these use cases to provide at least two different references in the bottom and it Links to somewhere else out. We provide a disclaimer saying we don't the CNCF does not endorse. These are provided purely for reference and Move on is it would everyone feel comfortable with that? Yeah, I think that's a reasonable compromise Perfect Sorry, just just one parting comment and not about Minio. So I think one of our functions here We're supposed to be cloud native storage experts and people are looking for us looking to us for advice and so So I think it's actually fairly reasonable for us to endorse Ways of doing things saying we think this is a good way of Addressing this use case and we think this is a less good way of addressing this use case and sometimes that will involve mentioning open source projects for example And I don't think we should be petrified of doing that. That's our job What we do have to be sure of is making sure we're being objective about this and transparent And for example, if a person writing a document Works for a company that sells product X. They should declare that upfront and And whenever we make recommendations, we should explain why we're making them and what the alternatives on what the pros and cons of the different alternatives are Awesome. Yeah, I appreciate that I think the problem with that approach is that The folks who have the most incentive to go and make the changes are the ones who are getting paid to make those changes Right, so the companies who have products that they're trying to sell are going to be the ones that are very actively trying to manipulate these pages To make their products look good. Yeah, great seems inevitable to me If we become very very product oriented or we have sections that are recommendations for products So I really like Steve's recommendation, which is Let's be generic about the classes or the Kind of categories of storage and the architectures and then leave out the specific product details lead those to the product vendors to To provide because they will inevitably provide their own documentation And then I like the compromise that Aaron said which is hey We could link out to product pages if needed at the bottom put a disclaimer there to say this is not, you know official CNCS indoor stuff. I Think that makes sense. Yeah, I like Aaron suggestion to and maybe add on Just a procedural point that when the PR gets approved maybe Representatives from the people linked below should be on that approval list. So You know in other words if somebody puts Their own product as one of the references and some other product as the second reference Let's have a rule that Somebody associated with that second reference actually approves the text above. Does that make sense? Yeah, actually, that's a good point when I worked in open sex with we made sure that whoever approved the The changes was not the one from the same company, right, right? So I think we should be very clear and actually that's one of the things that may Stop or have a good gate of approved value Documents that go into this repo So, yeah, I'm all for that not I mean sometimes you can't find a person from another company But at least us we can part of the reference here I you know, I see the minial one on the screen if you put a second one there and Put a link to something about that second thing what that was not their preferred entry portal Sure, for example, if I wanted to be nefarious I could link to some five-year-old document that would be bad and we don't want that to happen So I think if somebody's Name product or even open source project is included here that somebody affiliated with that particular Target of the link should at least review the document rather than getting blindsided by some Public document suddenly appearing that they had never seen before okay, so let's Okay, so we're gonna do four things we're going to have the use case be Genericized we're going to provide at least two if not more references to Implementations of that use case with an approval out to those Open-source projects to approve that they want that use case included and it's the appropriate link and We will also add a disclaimer to the references to say these are not endorsed by the CNCF storage SIG merely provided as a reference to implement this particular pattern Is everyone good with that? Can I Are we saying the references are so we're saying the use case becomes a use case Category and then the references are examples of projects in that category That's my understanding Okay, and we think that that's going to be less contentious than actually Creating use cases for specific projects because honestly creating a category and then grouping Products by category seems even more contentious. I'm just I'm slightly more worried about that Well, I would say that the use case for a specific solution or project is linked So it's not like we're saying it doesn't exist. It just There's a top-level document that attempts to cover the category and then provide links to deeper dives into best practices and Considerations for using particular implementations now if those implementations are open source under the CNCF Maybe it belongs under the auspices of this group, but in some cases, maybe these links are going to point to other things You know using the landscape as an example the landscape even allows commercial products Let me it's only eight more minutes. So let me I I I grabbed some of the information that we just talked about today Can I next week is QCon? Maybe I meet with some of you there We'll try to redo the template and so on and then we'll meet after QCon again and see what it's like Let's it would also be good if we could collaborate with the other six Maybe on that Wednesday luncheon to see how they're handling something similar Throughout all the six in the CNCF and then revisit this at the next Yeah, I just want to say thank you for for driving this and my concerns here are not intended to try and derail this I want to keep the momentum going and I just want to make sure that as a sake we don't get embroiled in any more controversy Yeah, I completely agree. I'm with you, man No, it's it's appropriate. So I'm glad we're having the discussion We want to get it right first time So I have one last thought I Think we should one of our objectives is to also make these documents Interesting and useful enough to attract an audience right so if we exclude everything and make it totally uninteresting then No, he's going to lead it and it's not very useful Yeah, yeah, it has to be meaningful people have to Figure out be educated. That's what we're supposed to be doing And if it's just general information, we're not looked at as the subject matter experts Then we're all just wasting our time. I agree with that All right We only have six minutes left So let's move on to the performance and benchmarking doc. Is there an update on this? Yeah, so I'll give a very good update so we put an outline of the doc together and have started populating it We've had a couple of meetings with a group of a Group of people section too many things now with users from Percona and data core and Planet scale and storage us and we've split the We split the the document into sections and we're we're kind of Delegating The ordering of the of each of the sections to different people kind of in a similar fashion to how we did the Storage landscape documents So it's it's it's coming along. There's still quite a bit of work to do But we're hoping that we should be having a draft within a few weeks maybe two weeks Keep kind of roll obviously slow things down some looks Okay Obviously the documents open for for viewing and commenting so would love to to have any Feedback if you think we left something out or if you think other things should be out of it or changed or whatever Um is this What was the tool we were talking about in the last meeting was that the volume benchmark that was free and open that we could use Is that do we capture that? Here from the left. Yes, okay, so so for the volume benchmarks. We were we and the database benchmarks both of them were were free. We also The the Percona and Tasking guys are looking to talk a bit about the The suspense TPCC like benchmarks which Which are open as well which they've been developing so that's we'll go into that into that section there Does anyone have any questions or comments? No, just this I think this is super valuable Alex. Thanks for driving this. I'm actually busy trying to figure out Performance of various different alternatives at the moment and it's a nightmare Trying to find exactly this. Yeah. Thank you Cool great. I think this would make for a great cube con talk in the spring too Because I don't think it's real widely known that these tools exist and how to use them So it would be great to better advertise that out either as part of the SIG or is an individual Yeah, that's a really great idea All right. Thank you for that update and thank you for putting that together the next piece. We have two minutes left Dragonfly we briefly discussed this in The beginning of October and I don't think we Circled back around and so I think we had wanted maybe to do more of a due diligence and had I Don't know sod if we had talked to you or Brad about doing that due diligence For that, but I think that was kind of the next steps Was there did you guys do a coat? Did you guys do a review of it? I have not Heard about this. So maybe Brad has or did I don't know. Okay But I'm happy to follow up. Just let me know we can follow up on slack or email Okay, I don't remember seeing this either It was maybe there weren't maybe it wasn't a well attended meeting We kind of all have a very foggy remembrance of it to be honest, and I think maybe Alex and I talked about it separately. We maybe needed to Have the co-leads also better involved. So it could have been a discussion that Do you mind putting the recording there so I can check it out. Yes, I will do that Thank you. Yeah So so just as a quick background dragon fly was was looking to do a sandbox to incubator Promotion and the talk Proposed that the storage sink have a look the project team has presented to us. There should be a recording and a link to the to the Presentation further down in the meeting minutes and I recall we build some notes But it's kind of unfortunately slipped through the cracks, and I don't think we followed up on it So we we do have an outstanding to do I can do to follow up There were Quitter and I had some concerns About the project, but we need to write it up down some notes and share them so if if one of the techniques wants to Re-review the presentation recording. I'm happy to sync up with them and share my notes as well Should we do like a shared document between all of us to? Just have one place where all the comments are or do you want to do separate docs? That would be a great idea. Let's do that Okay They want to share out. Sorry go ahead. I like some I remember Dragon dragon, but a guy says that Charlie is willing to do a due diligence So I don't know if that's gonna happen. Well, it's up to us to do it. It is up to us to do it okay, and the process is still pretty squirrely from digs doing it to the TOC because the other thing that happened is I believe they presented to us and then the following week They also presented to the TOC so we have to Better iron out that process being so thank you for bringing that up and I'll include both Recordings in here as well and then be specific. I'll think up with Saad separately and we'll figure out the due diligence Thank you. All right guys. We're over time I've notated in the agenda where what items we didn't get to so we can talk about them on the next meeting I'm looking forward to seeing all of you at QCon and Have a great rest of your week. Thank you. Take care. Thanks. Thank you all