 It's like you're right next to a mall or something I'm sitting in a messy study fighting the sins. I have my fake stadium behind me. That's very cool too. Is that Boston? Yeah this is the old stadium for the New England Patriots which was worse than a high school stadium. It was terrible. It was cement seats and everything. This is way before the Patriots were very good. This is when they started to get good and then they created a new stadium. It was terrible. You can see openings here and here. There's no more opening. The whole thing is like a big oval. It was terrible. High schools were better than that. So morning Ray and T-Pool. I guess you are new to the SIG? Yeah I am new to the SIG. This is my first time calling. Excellent. Awesome. Welcome. Yeah I attended every once in a while. People. Very cool. Welcome. Nice. Wait a few more minutes. People are joining late today. Yeah when you don't have to drive anywhere and work from home some people get the time changes. Yeah I'll put the agenda in the chat. Morning SIGU. Morning Kristin. Morning. I was just going to suggest that we wait a minute or two because we've got fairly low attendance today. We probably have enough people now. Let's start off. I've posted the agenda in the chat window. So there's a few things I wanted to go through. First off I've incorporated all of the final changes on the database sections and management interface updates etc and that Jing, SIGU and everybody else had sort of provided feedback on and contributed with. So the final document for the version two of the landscape is now ready. It would be good if we could get sort of a final review. And I'm proposing that we close off any comments for the next SIG call. And then I'll raise services requests so that we can publish this officially. Does that sound okay with everybody? That's good. Sounds great. Just in summary Alex, the difference between this and the last one we published is it just that we added the database section is that right? There is the database section and the database comparison table and then there are quite a number of changes in the management interfaces section to kind of all of that was sort of refreshed to take into consideration some of the new things that happened with CSI and object interfaces and that kind of thing. Sounds great. And just to be clear that optional text is something that will be used as reference for the other document, right? That's correct. Yes, yeah. So in fact, yes, that optional text that we had written, we said we were going to include it in the use case template and I've put that into the, I've put the headers into the use case template. So the next doc I just wanted to talk about was the performance and benchmarking doc. We made a bit of progress in February and there is quite a bit of content there, but we've now stalled. There are actions on myself to tidy up the intro section of the documents which I'm happy to do over the next week or so. But we do need to flesh out the benchmark example sections which Nick Connolly and a couple of the guys from Percona were going to help out with, but we haven't gotten that yet. So I'll chase up, but in the meantime if anybody else has any updates or has any ideas for this, I would really appreciate any help that we can get in. I'd really like to clean this up and have a version that's ready for review for the next meeting because it's been outstanding for it too long at this stage. And then again, similar sort of thing with the use case templates. We had a bunch of discussions. It was, we had some fairly contentious discussions as to what we're including in the use case template. We finally agreed to have a use case based on groups of categories and we decided to move the markdown into Google Docs and put in a couple of other sections to cover off some of the other comments that we had previously discussed. But we've kind of stalled at this stage and I think it's probably a bit of coordination and a bit of work where we can maybe spend, we need two or three people I guess to sit down together for an hour or so and flesh out one example and then release this for comments and further work. So Luis, I guess you'll put your hand up for this. But anybody else who would like to help out with this is again super appreciated. Awesome. So how about we set up a call maybe in the next week or so and just use that to flesh out this example. In the meantime, if you want to give some thoughts to what would go into it, that would be useful too. Sounds great. Just realized I wasn't meeting the entire time. Sounds great. And then we had a little bit of surprise. So Jing pings me and reminded me about the survey that we had prepared last year and it turned out that I think following the webinar that we had done, we got 30 or 40 new people responding and we just hadn't checked up on it unfortunately. So this kind of fell through the cracks, which is a little embarrassing. But the results are now in because we have at least 54 people had responded. So there's some meaningful data there. I guess the question is now, how do we process this? It would be very useful obviously to maybe summarize a bit of this and present it in maybe the next QCron whenever that is. But also, should we actually formally review this in terms of helping to define any priorities going forward? So I'll open it to the floor for comments. Does the CNCF have a blog website where we could just write us a blog and say these are the results and just kind of comment on them. Just keep it as simple as possible and just let everybody come in and kind of consume that information. Just an idea. Yeah. That's a good one. I think it'd be good to have a presentation to this group at least and perhaps a discussion of the outcomes. First of all, just to verify that we think we've got sensible responses. I don't know what makes sense. We've kind of sanitized the results or made sure that they make any amount of sense. So let's do that first and then we can perhaps say I think a blog or something like that in addition to perhaps QCron present would be good. Excellent. Sounds great. Yeah. And you know what? We could probably also ask Kim if we could do a webinar if we prepare any presentations. That might be useful too. Okay. So then moving on, we have a number of project reviews in progress. The first one is the TIKV graduation. So as background, this project is a distributed, sharded, replicated key value store, which is used as the backing store for TIDB. And it's currently an incubation project. A due diligence stock was written as part of the PR. And I've had a brief review of that due diligence stock. But I guess one of us needs to go through that in a bit more detail and perhaps try out TIKV. And we also need to schedule a presentation of the project as an update because the last presentation that they did was prior to incubation. Erin, I saw you had volunteered to look at the DD stock, but did you actually get to do that? No, but I have time today. So you can wait till the end of the day. I can look at it. Awesome. All right. If you need any help, I'm happy to help out with that as well. And then the next step was going to be the next project was going to be Harbor. So Harbor has been through a number of the SIGs. I believe it's with SIG Runtime who's driving it. We had some concerns after and Saad had prepared a review of Harbor. And during the talk meeting, we had some concerns around the use of the fact that there's a single point of failure with some postgres dependencies and things like that, which we wanted to highlight. So I guess there is a final step there to move our concerns from the issue I think that Saad had created into the due diligence doc. Unfortunately, Saad wasn't available to join today. So maybe I'll ping him offline. But if not, I don't know, Quintin or myself maybe can move the information in. Because I remember, Quintin, you had quite strong feelings on this. Yes, I can. If Saad wrote to Delta, Saad did most of the due diligence, I just sort of chimed in at the end as is my penchant. But yes, I think if he's written the stuff down, I'm happy to summarize it in the consolidated document or move it there or whatever and sort of make sure it's represented correctly. It would be great if he could, but if he's not able to, I can do that. All right. So we'll ping Saad after this. So next one then is the Rook project. So similarly, this is a project moving from incubation to graduation. Jared and team had presented to the SIG a couple of weeks back and we're going to work in our next progress, the move to graduation. They had completed things like the security review and things like that. However, I believe there was an issue where the due diligence doc hadn't been completed at the point of incubation, which basically means it must have slipped through the crack somehow, which is not great. So the challenge now is to implement the full due diligence doc for the graduation request. I don't know if any, if Quinton or Aaron has any more background on this, that's worth sharing. I don't have any more background on it. Alex? I just saw a discussion with Jared. So just to be clear, it's not about a document. It's about an actual due diligence. It's about somebody actually going through the project and basically checking out that we think it's up for grad, that it's of sufficient quality for graduation. And I thought someone had done that. Did we miss that? I guess that's what I... So I didn't do that and I was not around when they... So maybe let me just clarify. So they went into incubation at some point in the past and I was not around at that time. I was either not on the TOC or I was on long vacation or something, but certainly I don't recall voting for it or having anything to do with it. And then, and I hadn't looked into the exactly when that happened, but it was a few years ago, I think, in 2008 or something. And Jared pointed out on some other thread or email that they had not actually had a proper due diligence at that time. So I don't know, I haven't gone through the archives to find out like how the vote happened or why the vote happened without a due diligence, but it sounds like that's what happened. And all of what I'm saying should be verified because it's kind of based on emails I think I've read. And so my suggestion and I think the suggestion of others is that we need to do a due diligence before they graduate and that hasn't been done yet. Right. So remember this was at KubeCon, Austin, and it was just silliness going on. So I'm not surprised that it is kind of messed up. So I'm happy to also take this one on to get a proper document created. It's almost like, I remember when this was happening in Austin and I remember the meetings. And it was the first project that we were trying to get to SCNCF if I remember correctly. So I don't think we had to process down yet. No. And it was the first storage. It was not coming in a sandbox. We didn't have sandbox actually at the time. No, it was in section. Yes, exactly. This was not the first SCNCF project. We had tons of them before that. No, I mean storage. So there was a bunch of discussion. The PR for it is insanely long about people who aren't even associated with this project weighing in. I don't know if it should be grandfathered. I said, look, it was just done before. I think we don't want to have graduated SCNCF projects that have not had a due diligence done. So I would suggest, and it's going to delay the process obviously, and it's going to disappoint Jared, and they've been trying very hard to kind of get their thing graduated. And it looks to me superficially as if it's probably like totally fine to graduate it. But we do need to actually do the due diligence that have been done in the past. That's my feeling at least. That's fine. And I want to emphasize it's not just a matter of a doc being missing. It's a matter of like, I'm not sure anybody has from the TOC or SIG storage have actually gone and looked into the details of that project. It's one thing to see a presentation. It's another thing to actually go and verify that the project is everything it says it is. And I had no reservations about the team or Jared or anybody else, but we still need to do that. I can take this on, Alex. Paul. So Quinton, does this entail sort of like actually speaking to end users and that kind of due diligence, or is that something that the project provides? We can decide that. Years and years and years ago, I put together a guideline for doing due diligence, and it's got a whole bunch of questions to ask. Make sure you have sufficient information to verify this and that in the next thing. And I didn't, you know, I was not explicit about exactly how you should get that information by speaking to users. If you think there's credible information that says that XYZ is using this at so and so scale, then it's up to you to decide whether you think that's credible information or not. And if you want to speak to, I mean, let's face it, even speaking to customers does not necessarily provide credible information. So use whatever methods you think are necessary to ascertain credible information. Right. Okay. So that book went through a bunch of reviews on the other six though. So I guess that's why I'm confused because I know that they went through security review and another review. I mean, there has to be some, I think there has to be some assets that we're missing that were created out of the graduation process. So I don't feel like we have nothing to go on that they haven't gone through some of the process. I just don't think we have all the information consolidated into one place. No, just to ask, I'm curious, are all the projects are graduated? Do they all have due diligence? Yeah, they should have. Yeah, I mean, that got into incubation went through due diligence. That's the problem. Not the graduation stuff is a very small incremental addition to what's required for incubation. So incubation is the major due diligence that happens in the CNCF. Sandbox is, you know, very lightweight because we want immature and, you know, twinkle in the eye kind of projects to be able to be hosted in Sandbox, but due to, but incubation is the big one. And yes, I believe they have all had it except this one. And this is another oversight, but I'm not aware of any others. Just wondering from old, the one, CNCF first started like projects like, I don't remember, like Fluent D or Linker D, one of those two, one of the first few graduated projects. Do they all went through it also? They went through it when they went into incubation, allegedly. I wasn't on the TOC at the time, but I did speak to Alexis Richardson, who was there from the beginning, and he helped me put the due diligence guidelines together. And he said, yes, that that's what they've done for. Okay, cool. Just making sure. Are those due diligence reports or recorded somewhere on GitHub? That's what I'm less sure of. I'm not sure how much of it was written down, but allegedly, the TOC did actually go through the exercise. We can speak to Alexis if you like, and he can give us the details. I believe in the TOC repo, there's a directory with all the previous due diligencees, honestly. I'm sure I saw that the other day. We also have the recordings rate of every TOC call. Yeah, but it goes through recording. So I think this one might be slightly, because so just to summarize what Quinton was saying, all of the main due diligence happens at the incubation stage. The graduation stage is kind of like a couple more steps. So for example, at graduation, you just have to show additional users more than one user using it. And secondly, you have to do a security review, which is actually paid for by the CNCF. And I think what we're saying here is that we missed the DD at the incubation stage or at the inception stages, as it was known then, due to some of the previous drama, which we will not regurgitate again at that infamous meeting. Okay. So finally, the Previka project that presented at our last meeting, I'm going to follow up briefly with the project team. I think the Previka project looks very interesting. It's an interesting set of components to which sort of lead it to be able to be used as a message queue and another event processing. In many ways, it would be comparable to perhaps the NATS project, which was already in the CNCF or maybe even something like Kafka. So I think it's worth following up with them. But they're not 100% decided to donate this to the CNCF at this stage. I think they were trying to debate whether to do it as part of the CNCF or as part of the, I believe it's the AI foundation or something like that. That's also part of the Linux foundation. So I'll reshack to them just to see what's happening there. And then we can either close that off or move forward, depending on what they come back with. I think that covers the agenda for today. So I just want to quickly summarize the storage landscape leads to feedback please before next meeting. The benchmarking doc, I really need help with the benchmark examples. Tom, I see you joined the call. I don't know if you can help move that along at all. I wanted to look at it a little bit. Who's been contributing to it so far? We had, so there's myself and Paul Sobri from StorageOS. There's Nick Connolly from the Terra. There's Vadim Chachchenko, which is I believe from Perkona. And there's Saif Al-Harthi, which is that Planet Scale? Yeah, Planet Scale. Yeah. And Tyler Desana, who I think it's Perkona as well. Yeah, Tyler and Vadim are Perkona. So my comments in there were really around the right amplification and different technologies at different layers in the storage stack, how to handle those. I've seen a lot of people benchmark SSDs very poorly and don't get to a steady state because it can be difficult, especially with large SSD pools. So I can try to put in enumerate some of those pitfalls, but it's a big step when this is kind of a primer and getting into that kind of right amplification is going a little deeper. So I didn't know what kind of audience you wanted to really tailor to. Is it just to get awareness or did you really want it to be kind of a checklist, walkthrough, guidance kind of document? So the idea was primarily to kind of inform users and especially in the pitfalls and sort of common considerations. It's kind of like common issues that people might come across. So certainly, you know, if there's a paragraph or two that you can talk about to write amplification and SSDs, I believe we actually had it to do on the list to put something like that in there. I think that would be fantastic. So, you know, if you can just stand it in as a comment or ping me an email or I can add you to the doc as well. Yeah, I put in comments in a descriptive way, but they weren't ready for the, you know, the document itself. I could review that and actually add it in if you want. Yeah, that would be fantastic. Let's see. Do I, yeah, I have, I guess I'm looking at it. I think the comments did get added in as a suggestion. And if you look on the page, they're under commented falls. Oh, yeah, I can see that page three. So all right, I can add that. Yeah. And just ping me if you want to, if you think we need to go more or less detail, we can work through kind of your where you want to go. I can work some more. Cool. Okay. And then just moving through the agenda very quickly. So the next action was on the survey. So I was wondering, Jing, do you want to sit down maybe for half an hour? And we can extract a couple of summaries from that so that we can present it to this sake next time? Sure. All right. And I think we have action items for TIKV on Aaron, Harbour on Sandham and Clinton, Rook on Aaron and Alpha with Pravega. We've gotten through quite a bit in half an hour, but does anybody have any other items that you'd like to raise? I just wanted to actually thank you, Alex, for looking through all this stuff that we have actually achieved is pretty impressive. It's a small team and most of the work has actually been done by you personally. I think we should all take the opportunity to pat you and the rest of the people who have been involved on the backs and things because I think this is one of the, yeah, one of the more effective six, personally, as far as I've seen. And there's a lot of really valuable stuff that's been happening here. So thank you, everyone. Well, thank you. And indeed, thanks to all the team. We've done some magical stuff together. Cool. All right. So in that case, unless there's anything else we can call it today. Thank you. Thank you, guys. Thanks, everyone. Bye-bye. Bye.