SNIA Experts on Data
Listen to interviews with SNIA experts on data who cover a wide range of topics on both established and emerging technologies. SNIA is an industry organization that develops global standards and delivers vendor-neutral education on technologies related to data.
SNIA Experts on Data
When AWS S3 Keeps Changing, Who Keeps Up?
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Your customers don’t care who wrote the code, they care that “AWS S3 compatible” actually works with the apps they already run. That’s why SNIA developed the Cloud Object Storage Plugfest Program – a collaborative, multi-vendor interoperability testing environment for S3 developers.
In this episode, we sit down with Cloud Object Storage developers, Mark Trinh (Oracle), Jason CWIK (DDN), Veer Kumar (Nutanix) and Michael Hoard (Chair of the SNIA Cloud Storage Technologies Community), to explain why this Plugfest is the fastest path to real S3 interoperability and faster bug fixes.
The Plugfest participants speak candidly about friction points in S3 compatible object storage: clients that behave differently than the documentation suggests, API options that leave too much room for interpretation, and edge cases that only appear when you test across many vendors. They highlight why AI-driven workloads are raising the stakes for S3 compatible storage and discuss why an NDA-based Plugfest model matters for fast troubleshooting, direct engineer-to-engineer collaboration, and practical fixes.
SNIA is an industry organization that develops global standards and delivers vendor-neutral education on technologies related to data. In these interviews, SNIA experts on data cover a wide range of topics on both established and emerging technologies.
About SNIA:
Welcome And Quick Introductions
SPEAKER_04All right, everybody. Welcome to the latest, the greatest. Well, they're all great, but I'm particularly enjoying what we're about to see here because we're going to be plugging some plug fest. Uh, thank you for watching the SNEA Experts on Data podcast. My name is Eric Wright. I'm the lucky co-host uh of all this, along with my SNEA rep here today, Michael Horde. Uh so Michael is gonna lead us through what we're talking about with what's coming up. We've got Plug Fest around object storage. There's a lot of really great stuff, and we got three other amazing humans joining us for the call today. So I'm just gonna do a quick round to get everybody to do a quick introduction. Michael, start with you since I already called you out.
SPEAKER_01Sure. Uh hi, this is Michael Horde, and I'm the chair for the SNEA cloud uh storage technology or CST community, and we have a couple different uh activities going on. One is educational webinars, and please check those out. We've got several on agentic AI, and we uh the topic for today is the plug fest, and that's a community-driven effort to promote uh interoperability, and most of our focus is on S3, which we'll get into that detail. The last uh activity is a um the the uh uh cloud object storage test tools, and that's a technical working group, and we're finishing a white paper right now, so we're gonna publish that soon, and that's uh for the purpose of creating open source test suites that we'll make available through uh GitHub at SNEA, and that's to drive uh interoperability automation. So pretty cool stuff. Thank you.
SPEAKER_04Yes, you had me at interoperability. Uh next one in my round here. I'm gonna go counterclockwise, Mark. If you want to introduce yourself and uh and then uh we'll make our way around the rest of the crew as well.
SPEAKER_00Sure. My name is uh Mark Trin. I'm part of Oracle. I've been with Oracle for about 10 years now. Um I'm pretty excited about this because uh Plugfest, we just recently joined uh uh almost a year ago, and we've been getting a lot of benefits from from being a part of Plugfest. So uh I'm excited to talk about some of the things that we've been working on in uh with Plugfest.
SPEAKER_04My favorite thing to plug is Plugfest. Uh Jason, uh next up. Yeah, I'm Jason Swift.
SPEAKER_03So I'm a senior director of engineering at DDN for DDN Infinia. Uh, this is also great for us because we're a new product and the Plugfest gives us some early access to do some testing. So we're super excited about it.
SPEAKER_04And Vier, uh last but very certainly not least, uh tell us about yourself and uh where we can find you for connecting online.
SPEAKER_02Yeah, so this is Vir. Uh I work with Nutanix. I'm I'm part of Objects team, uh, mostly working in replications, uh replication area. This is the first time we are participating in PlugFest, and we had a lot of learning which we like to share. Excellent.
SPEAKER_04Well, like I said, one of the things that I love about SNEA as a technology and a human community is the engagement you can have with peers, because as much as we're kind of single-tracked inside our own organizations, doing, you know, we're innovating as as best we can with the tools and teams we've got. What is incredible is that looking at the standards driven and kind of this community broad community-driven approach where we can all learn from each other because effectively we're all consuming each other's resources and we're no longer the single source of anything in in any organization. So the fact that SNEA gave us a place to innovate internally with an external focus so that we can say, hey, this is what we're doing, and I've seen the successes of of this kind of collaboration. So, Michael, do you want to give a quick rundown on on what is the plug fest series, like and and yeah, you know, why should companies get get on in? Because this is a pretty fantastic way to meet some amazing people.
SPEAKER_01Awesome.
What SNIA PlugFest Actually Does
SPEAKER_01Yeah, I'll I'll uh just have a brief uh statement. So um so we really created the Plugfest as a means for developers to get together. Uh previously, there had not been any forum within the industry that uh addresses this. And so uh we we were meeting the need for just getting together and doing tests. There's no other uh organization in the world that does that right now, and um and so some of the uh benefits are the um uh time to uh fix bugs in the in the plug fest, you have the peers working together directly, and this uh and so they're sitting at a table or online, a virtual table, and you get immediate response to not only find bugs, but then the collaboration that uh is required to do troubleshooting, uh getting traces, uh, and then uh different uh you know dumps of uh the intermediate data out of testing, journaling. But then uh the contrast is that uh you would spend weeks, if not months, trying to find your peer in the industry, and often uh the support teams will not allow engineers to uh talk to each other, and so it's not only uh next to impossible, but uh you know the the rate of bug uh uh resolution is is just minimal without this kind of activity. And uh, but we're delighted uh to have the the history we have now. It's um uh our first plug fest was uh um was in September of 24, and that was uh in-person at SDC 24. We had another one April of last year 25, then another uh face-to-face in uh uh September uh for SDC 25. Now what we're doing is uh online plug fests. So we're doing three of those. We had one in February, we have one, uh another one in May, and another in uh July, then and those are virtual, and that's really to speed up the cadence of the activity. Uh, and again, this is all uh member driven, so it's really up to the community to you know sit at the table, figure out what do we want to test, what is the the key area that's missing or uh pain point, and and it's all member driven. And so then we'll uh at the end of this year, we'll also have the in-person SDC 26. So that's kind of a thumbnail of uh what we're up to.
SPEAKER_04Well, the good thing is the you know, you talked about this idea of what we get together virtually and in person, so it's a good combination. You know, we I remember years ago I worked at another organization. This is gosh, it's got to be like almost 10 years ago at this point, and I'd heard about this. Like I got asked, you know, I was on the marketing team, and they said, you know, an engineer came to me and said, Hey, can we host this thing called a plug fest? I was like, sure, what is it? Yeah, and and that was back, you know, around SMIS kind of controls. I was working at an organization and we were doing optimizations for everything across the full stack. And I just said, This is amazing. So we're gonna invite other people to hang out with us. And so I went to my team and I said, I want to run this this event, we're gonna host a bunch of people, we'll pay for pizza. It's basically like a meetup. And they said, Okay, great. You know, what's the advantage to us? And I said that we meet other people who are doing the thing that we're saying we can make better instead of us reading API docs and reading manuals. That's like we we can do a lot with what we see in in a swagger doc, but we really can't take true customer experience and true like field experience. Uh I'll start with you, Mark, because you mentioned, of course, you're this is sort of a fresh thing for your team.
Real Interop Wins From Oracle
SPEAKER_04What's got you kind of excited about you know what you see as the opportunity for being a part of Plugfest? And and specifically on the object store and the S3 side, what's what's coming out of your camp that that you want to bring to that that room and and that discussion?
SPEAKER_00Yeah, it's it's been pretty exciting. I, you know, like I said, we've been at Plugfest almost a year ago, and everything that Michael said was basically experience that we got, right? We got a chance to sit in the room with like-minded people that also has object storage. They have the clients that they use against our object storage S3 implementation. And we get a chance to directly interact with their tests against us, and then if they reach any issues, I can immediately um help figure out okay, what is the issue and help them resolve it, as well as get to learn about their behaviors that are interfacing with Oracle's object storage, right? And so uh because of that, last year we we found you know differences due to testing that with with different clients and came back uh from last year's plug fest with new features added. And then at this recent um online plug fest that we had, I was able to directly test their their same clients and even new clients with this new feature and get immediate feedback from from that. So that these are really great opportunities for for us to kind of get our um S3 implementation correct and working with multiple different clients.
SPEAKER_04That's amazing. And Veer, I know that you've you know again, you're you're a full stack kind of kid. So you you're your your platform obviously touches a lot of different areas of the physical and virtual stack. So what's what's your hope that you want to come away with with the upcoming plug fest? And what's been your experience so far?
SPEAKER_02Yeah, so let me start with why we started with this uh this whole initiative. Um we uh we have a certain set of customers uh who provide object stored as a service to the other um company uh customers. So they work with multiple vendors, and the most most of the time what we have seen is certain set of APIs and certain sort of configuration which works with one vendor, doesn't work with another, or they have a little different behavior than expected, uh right? So we were seeing those issues. What we wanted to do is uh we were looking for either we do it ourselves, uh bring multiple solutions and try to test with each other and and set the expectations right for the customer. Uh, that if you use this, then you will uh you will get this kind of behavior. And then we got to know about this plug fest. Uh we immediately jumped on it, and uh this time I think we are like uh uh I tested with around nine to ten vendors, and and it was fantastic learning. Um, we were able to test with them, and they also tested with our product. We figured out the issues, we immediately raised the concern with the uh um specific people, and we are fixing it and uh making it more better and stronger and and more easy to manage uh as it was before.
SPEAKER_04That's uh lessons of uh also I always quote charity majors on this as we all test in production, and that's often where we find the stuff out is you know, we go to the field and we don't have a chance to bring that knowledge back as often as we want, but yet we find all these RD teams independently and like being able to share even among young teams in multiple product areas is its own challenge, let alone being able to go out to the community. So having this as a forum to get together is is super important. Jason, of course, your org has no shortage of BUs and platforms and products that you're dealing with. So, what what's got you excited and what do you see as the biggest impact opportunity that you and DDN can have in this?
SPEAKER_03Yeah, so I mean, for us, obviously, you know, Infinia's the new kid on the block, and we GA'd about a year ago. And so, you know, this is my third object store. So I did uh EMC Atmos, uh, I did uh Dell uh ECS, and now I'm with DDN. And every time, no matter how much testing you do ahead of time, you know, how much you know prep work and trying to follow the documentation you do, inevitably you put it in the customer's hands, right? And they're gonna break something. So, you know, getting getting exposure to as many applications as possible, you know, in the shortest time possible, you know, helps you plug those gaps because inevitably they exist. A great example for us is that, you know, we were um we were uh the the key in an S3 object is 1024 bytes of UTF 8. Well, it turns out there are some invalid sequences in in uh UTF-fate that you need to deny because you can take it in the URL, but when you put it back into XML to put back in the list object responses, also now you have invalid XML. So it's it's imperative that you filter out these invalid UTF-8 sequences. So that is just you know invaluable to us to you know get a get those bugs out of the way before they're found by customers. Um, so and then you know, what we do too then is you know all talk together about and share these experiences to help improve the test suite so that you know everybody here can be operating at the same level because ultimately it's in all of our best interests to have compatibility, right? The storage vendors, we want to have maximum compatibility with other storage vendors just because we want to you know grow the ecosystem of you know S3 compatible applications. And then the you know, the client vendors, you know, they want to test against as many uh you know storage systems as possible because we all do things a tiny bit differently, right? And so we want to make sure that we identify those flight differences and get ahead of those.
SPEAKER_04And there's definitely those, like you said, those are things that we can't in our subset, we only have exposure to our current customer base, and it's really tough to go outside of that. And the last thing you want is that interoperability test to come up negative when you're not there to watch, and yes, and all of a sudden, you know, then it becomes you've got vendor engineering teams who are like, is this our problem? Is it their problem? So being able to do this and not wait for it to be in the wild is kind of huge.
NDA Testing, Fast Debugging, Shared Learning
SPEAKER_04Michael, you talked too about being able to store trace data and test data, because I know that's like the IOTA, you know, project has been fantastic for so many people. So what's the what's the piece that you're seeing as what the S3 project, you know, and this it particular initiative is gonna bring in the very short term, and and what's the benefits you've already seen in the field from from your folks?
SPEAKER_01Yeah, excellent. So uh several things here. So, first of all, uh you know, the main value and benefit that I think Jason and others uh mention is finding bugs before the customer finds them. And it's not just the test results. Um, obviously that's uh the main uh you know entree for the plug fest. However, I think there's also just the socialization and meeting your peers in the network in the in the industry, you know, sitting across the table. So you immediately get your like-minded people and you're working uh you know uh uh together on the same uh projects. What we do is we talk about themes. Okay, what's the theme for what we're gonna focus on? And and I think with uh Mark, you know, a year ago and uh Jason uh just recently and Veer uh was uh just the ability to to compare notes. And I remember the very first uh plug fest we had, uh, we were going around the room and people were sharing uh, well, we're using out of five options on this one command, uh, we picked number three. And people go, Why did you do that? You know, and uh, well, don't you just pick one? You know, and so so it's just like a vacuum of uh in of industry knowledge, and there's nowhere any of this stuff is documented as far as a common command set with options and the and just the rapid pace of innovation that AWS is doing, you know, those all combined, we have got to be talking together, and so it's comparing notes and uh just sharing best practice. What we're and you speak of the the trace files, so we are not um uh you know collecting those traces to be able to create a repository right now. We're not doing that, and the reason for that is we have an NDA, and so all the testing is done under NDA, and that's uh based on a very long uh uh policy for SNEA and and uh 30 years ago I was working on uh fiber channel uh plug fest, and I was on the board for the FCIA at that time, and so this is a very well um orchestrated uh initiative. You have NDAs in place, and that's just to secure uh when you find a bug, it's not going to be publicized. Uh, we just uh find bugs, fix them, and move on. And um, and then the final point uh is uh we are intending to up level some of the uh lessons learned into best practice, and that's what we're going to document. And uh I'm I'm right now finishing a white paper with our test tools team. So we're we intend to have uh white papers, things of that sort. Um, and over time we will uh create the test tools. And I do want to speak about uh some of the developments on AWS side, uh, but but that's kind of a summary of uh you know what we're doing right now. But but go ahead.
SPEAKER_04Yeah, well, I think the this highlights the I'll call it the closed open, the optimal mix, because this is where as you when you come in as a participating vendor into the community, we want to have obviously want to share as much as we can, but also know we've got the protection. And you know, when you mentioned about IOTA being fully open, I know for some folks that became a problem because that means that ultimately you've got to be very careful about what data you put in there, and that makes it, you know, it makes it tough just because as much as you want to, there's a lot of extra, you know, loopholes you got to work around and and sort of steps you got to run through. So it's nice that we can collectively share, you know, data, stories, and and actual field information, but yet still maintain the proprietary side of the world safely. You know, it's it's so wild that people think that we're all like you know, like angrily, you know, punching each other in the ear when we see each other in a Starbucks line because we were for competitive companies. It's like that's not the case at all, you know. And uh, you know, Mark, you know, given that you've come from a spot where kind of a quiet part of the world, there's an amazing amount of of storage that's being doled out, you know, from your your camp. And what's been the the sort of favorite happy accident you've had as a result of being in these communities where you say, Hey, you know, we we may have found something that is it just us? Like, what are the the anecdotal side of the world stories that have made this beneficial for you as you know an individual as well as a team?
SPEAKER_00Yeah, so for us, like I said, just talking to different vendors and almost like complaining, right? Like, oh yeah, we ran into this issue, and then seeing the same thing, like, oh, you know, we ran into that same issue, and then they might have a solution that we never thought of, and we collaborate right there. Uh it it's been it's been uh like amazing just sharing the same kind of war stories that that we go through.
SPEAKER_04You can't just do the like what it worked on my computer. And uh and VR, same thing again, deals thinking that we don't necessarily look at like storage being the only or most important part, but when it all comes down to it, all the compute in the world isn't gonna do you bugger all good unless we can actually store it more.
AI Workloads Reshaping Object Storage
SPEAKER_04And as we see more and more AI workloads and the importance of how object storage, and in fact, with the very recent changes that happened in with S3 itself, now being able to mount it as a file system. So that like being able to expose those primitives now in a different abstraction, that opens the door now for inevitably all of us to kind of do the same thing. And then if those adoption patterns work in one vendor, then now we can share that because that's the other thing is. We don't know if that's actually a pattern that people will go for. And so you you talked about kind of your customers and and other customers. You know, what are the stories that you've captured because you were able to talk to somebody else's engineering teams about this stuff?
SPEAKER_02All right. So uh so the the the major thing which uh which uh like I I remember a story vaguely. Uh in 2020, AWS changed from signature v2 to signature v4. And and they updated the SDKs, and all the clients which were using those SDKs suddenly start working with third uh S3 compatible uh storages. Now, if we had a certain set of tools, which is common across all the vendors, which we which uh Michael is talking about, if we have that kind of tool, uh we could catch those issues quickly. And and fact of the matter, at the time we just had the release and we had to release a dot release again to fix that issue immediately. Yeah, so that is one kind of thing. Another uh thing which we are uh looking at given the uh else uh AWS release, the file compatibility, uh like our solution already had the file compatibility from last three, four years. Um, but if we had shared this information with other vendors and like if we have collaborated more, uh, if we had that platform to do that, uh other vendors would have come up and we would have come up uh with that solution way sooner than what we are seeing today. So uh similarly, we it's going for S3 vectors and any of the new features which is coming in. So it's not necessarily that AWS also can drive things um um uh first. We all can come together and and and bring things and make it more standardized, and maybe AWS can adapt from us.
SPEAKER_04And it's so funny coming back for like I'm like an old school OpenStack kid. So I remembered even in the early days with that community and and Randy Bice in particular, who really wanted to fight for like let's have broad cross-compatibility. And there was a big project around getting S3 as kind of a standard bearer inside the OpenStack in the Swift ecosystem. And it was a real tough battle because we were sort of really adamant that it had to be entirely open source, like from bottom to top. And that created a bit of a blocker that we couldn't adapt to what would obviously become a fairly broadly adopted standard in by customers just by volume. Uh, and now, of course, you know, the workloads that are consuming this stuff are fundamentally different than they were even two years ago, because now object is the new hotness. We thought that everything was gonna be the traditional, you know, journaled file system, and it's gonna be that's where everything is gonna live. But now seeing how the AI workloads are fundamentally changing the consumption pattern, and Jason, obviously, that you're working on a particular platform in the org that's really targeting that exact use case. So, what are you seeing in the that the AI is now the consumer, not the deliverer of the service?
SPEAKER_03Yeah, I mean, when when AI is consuming data, there's a lot of different changes we're seeing. One is just simply the volume, right? We're not talking about, you know, where it used to be a petabyte of data was huge. I mean, we're these these AI training workloads literally have hundreds of petabytes of data in them. And so that really presents a challenge for a file system to store that many, that, that many objects and that volume of data uh successfully and make sense out of it at the end of the day. Um, so and so that's kind of the training side, but then also then on now we're seeing the inferencing side pick up. I mean, over the past, you know, six months here, we've really seen that inflection point where you know the training is now starting to or the the inferencing side using the AI now is starting to surpass the training side of things. And that's a fundamentally different workload now, because now you're talking about fine-grained key value pairs that are gonna be cached inside the layers uh of the AI. And so it requires an even slightly different access pattern than to be to handle those small, fine-grained objects that are gonna have, you know, a random identifier of a GUID. And that's something that's definitely extremely hard for a um for a file system to take care of, you know, talking about billions of these small, fine-grained objects.
SPEAKER_04So, and also I think this opens up the next level of innovation because now as we set this as the sort of a fixed abstraction and we agree that these are the primitives and these are what we're gonna expose, and everybody has figured it out and we've actively done this. Now, when people want to start tackling key value store cash acceleration and working with metadata optimization and acceleration, now we know we're we're built on a foundation that we can consistently rely is going to be there. And this is where that sort of standards-driven approach is is pretty amazing. And the funny thing is we've almost accidentally arrived at it. Like obviously, we're we're all having the same customers or common customer patterns. So it's not like we're just like making up you know object storage as we go, but uh, as you talked about, like we've now converged where you know we call it the S3 object standard because that's kind of become the de facto phrasing of it, you know. Same reason, like these are I call these Kleenexes, but they're they're maybe facial tissues if you were really get deep on the on the naming convention. Uh so Michael, what do you see for what do you want to see more of in this community?
AWS Tooling And The Road To Alignment
SPEAKER_01Okay, great. Thank you. Um, yeah, so we are eager to work with AWS. We see that this whole uh activity is uh compatible, but also um uh complementary to what AWS is doing. Uh just over the last year, AWS has announced uh quite significantly new energy around interoperability. And at the last reInvent, uh there was there were several presentations on that that touched on it. But even more significant than that was they released a set of tools. Uh the first one is a Smithy tool, and that's uh IDL, a uh interface definition language, and that covers not just S3, but it's the entire uh AWS suite of services. And the second tool is the um the API modeling tool, and those two combined with some of their nested other tools for auto uh code generation, uh, you're able to actually uh develop your own SDKs. And the beauty of that is the uh that AWS themselves have been using those tools for quite some time. I I don't know if it's 10 years, but uh but it's very uh a significant amount of years. So what we're eager to do is engage with uh AWS and uh and uh of course all of their partners and be a part of of this activity. So I think the key the key point here is that SNEA has been doing this for over 20 years as an example with Microsoft. So Microsoft has their SMB protocol and it's proprietary, they own every aspect of code generation and roadmap. Uh however, they've been working with SNEA with Plugfest for over 20 years. I think it's um on the order of 28 years, uh it's in that ballpark. So SNEA has a track track record. And as far as facilitating, hosting, and guiding this kind of plug fest activity, that's really what we want to accomplish. And again, it's to work directly with AWS, uh, be part of um, you know, the ecosystem that will find the bugs before customers find it. Um, and we work with uh under NDA that we described before. So this is not trying to um, you know, control things or anything like that. It's much more to just allow the developers and the community to do their work, but then have timely communications uh with the source of truth, which in this case is S3. Um, I have to congratulate and thank AWS for releasing those tools. And so we have a lot of plans, more time uh to talk. Uh so Eric, we may have another podcast on the test tool side, uh, but that's a a glimpse of where we're headed. Thank you.
SPEAKER_04Yeah, and this this is the the fun of being able to like now we're the pace of innovation in the world is kind of moving really fast, and it's hard to tell which like which is the dog and which is the tail. We've got stuff that's moving around compute and memory that you know, everybody's like, well, this is like the current bottleneck. And like, well, underneath, I'm not sure if you're familiar where the memory ends up eventually, it goes down a disk, you know, and we're gonna solve problems that are currently hardware problems with software accessing storage. Like, I think my prediction is that we'll see much, much more kind of that tier two inference that we can move more of it down into the storage layer, take advantage of you know, high-speed storage and treat it as memory. We're seeing more stuff around with mempalace and these exciting new open source projects that are coming up, and they're all gonna leverage this common underlay, team underlay. All right, like it all ends on metal eventually, and that's the the cool thing. And and we all have different metal, but we all access that metal in a common way so that we know we can we can innovate on it. So, Veer, what's your what's your thing that you're you're you want to see and and and what do you want to bring from your team and and what do you hope to bring to that plug fest?
Automating Tests As APIs Shift
SPEAKER_02All right, so we are uh we are thinking to collaborate on these testing tools. Um uh so given that um nowadays we have agents who upgrade the uh software in in clients very quickly. So as soon as AWS releases something, it gets immediately get updated into the application and and it starts breaking our stuff very quickly, right? Because these release cycles are getting become smaller every now and then because of the code generation and whatnot. So, what we would like to see is if if we can get the collaborations uh going more uh more frequently, then what we can do is we can find those issues very quickly. Uh with these um uh with these test tools, we can um I mean navigate those uh unknown. Uh sometimes what happens is it's not just the bug, it's just the API also changes. So the behavior of the API changes, right? And sometimes it's not just the API, it's just the restrictions inside the API which has changed. Like um Jason described about the the characters used in the object name. Uh similarly, we have metadata where a size in Airbus is could be only two kbs, and for some customers, we might have allowed bigger than that, bigger than two kb, and now customer is moving from one vendor to another, and they they see that their metadata is getting lost without them even knowing, right? So if we can get uh ahead of all these issues, then that's that's the best thing we are looking for. So if if we have this common test which validates what works and what doesn't work with each vendor, that also helps the customer uh to navigate and navigate their migrations and um uh interoperability uh expectations.
SPEAKER_04Uh didn't Bill Gates tell us that 640k was going to be enough? I've been told this. Apparently, we uh we we blasted past that one. Uh Mark, what do you see as your internal goals and your external goals for for this plug fest?
SPEAKER_00Yeah, so I'm excited to see more companies join Plugfest and more of a selfish thing, right? I I want to get more clients testing against ours. Our big thing is we have so many different customers that that use S3, and they're they are on so many different clients, and testing it in-house is just not feasible. So having this Plugfest where we can actually leverage that and test against it is is really useful. The other thing is that we're we're looking possibly from the test tools, Twig uh, some kind of reference implementation of S3 that we can use against and say, hey, um this this S3 reference implementation works with Oracle and it gives some kind of high confidence for us that most um clients will be able to work with uh our S3 implementation.
SPEAKER_04I love it. I love it. And Jason, I'll say somebody who clearly is the fourth time around here, you've probably forgotten more about object storage than I'm ever gonna know. Beyond the storage, though, what does the collaboration across the other twigs help you? Like thinking about the security integration, all the other, like because we we're focusing on this as like a purely, you know, what is the right S3 abstraction? What are the right, you know, ways in which we can we can manage those resources? But then that ultimately goes to Swordfish and Redfish, and it goes out to the security twigs, and because they're we have different teams that don't necessarily sit beside each other, and we often find out right before something's going live, like, oh, make sure you run that past the security team and and make sure that it's cool runnings. So, you know, you probably have had some interesting experience over the years of it's not just the storage that's gonna matter when we bring this to production.
SPEAKER_03Yeah, absolutely. So, I mean, there's a lot of things. If you look at AWS, they you know, a lot of their services are very tightly coupled, especially when you think about IAM and STS and uh and how that moves off into EC2 and things. And so, really, like IAM is just this really complicated beast, right? And so, and you know, obviously as storage vendors, it doesn't make a lot of sense for us to implement the entire surface area of IAM, but you know, having that insight into what parts are really important uh so that you know we can you know ensure what we're doing is secure. And and and again, doing it in an interoperable way between us to keep, you know, to keep the load off of the you know software uh vendors to to make sure that they're doing things in the same way that's gonna work, you know, evenly across a lot of our applications is really important.
SPEAKER_04Yeah. And uh, you know, even despite we being told that we're it's it's another AGI has been discovered and we're the countdown is on. We've got, I think it's 12 months or maybe it's 18 months until we're out of a job. And but it's just that kind of keeps resetting every week. So I feel I feel okay. So when all that is going on, the plug fests and and that technical community is where we're gonna see, I think, the most good work that happens for all of us, and plus we get a chance to learn at how other people are doing things and bring those concepts back into our own internal teams. And that's what I really think is the best opportunity because instead of us all accidentally arriving at the same outcome, you know, within a some variation, it's so much easier to be collaborating in advance and then kind of running to the finish line together, arm in arm, instead of you know, hey, look, we're all trying to get to the finish line as a business, but we also are doing it in a way that supports each other because no one's going to be alone in a customer environment. So we have to build and and engineer
How To Join The Next PlugFests
SPEAKER_04accordingly. So, Michael, let's drop some details then for folks on how do they get involved and what is the the next best thing that we can point them towards on the SNEA website to make sure they can get the information they need.
SPEAKER_01Excellent. So, Eric, we're um we're going to share those links in the in the when you publish the podcast, it'll be in the description. Uh, but there's several things. One is the uh cloud storage technologies community, that's the CST, that's where we do all the planning, but it's for uh the educational, plug fest, and the test tools. So we plan all of that. Uh uh, we also do another standard uh called CDMI, and I think you've had a podcast on that. So that's a that's a fourth item. I I should have mentioned that. Uh, but I think the links uh will provide uh access to the CST in addition to the PlugFest. That's where you can uh register. And we have one uh so two weeks from right now, today, we will be meeting virtually in um uh May 4, 4th, 5th, 6th. And uh so that one's coming up soon. And uh we have our uh third online for this year in July. So you can uh sign up for those. And um, and then uh the in-person will be at SDC as soon as the SDC registration comes up, then we'll provide details on how to register there. So we've really upped our cadence and uh work. I will also put a link to the test tools twig, and that's technical work group. And uh so that's uh that's a way to join that activity. That uh thank you, Veer, and others that uh you mentioned that. So I think those are uh plenty of ways to get involved, and there's a huge amount happening. Uh just uh yesterday we had a call to come up with the agenda uh for May, and uh it's uh more work than we can do in in those three days. So this is something that's ongoing, it's gonna take us um you know a while to uh uh to uh develop those tools. And a lot of the tools are synergistic with uh plug fest. So what we intend is the plug fest is our um is where we do experimentation. So there's uh new tools uh that we're gonna be bringing in. There's a new tool from uh Brazil. We're hoping to contact them. Uh, and then of course the Smithy tools. Uh so those are all the things, and it'll be somewhat experimental, and uh and then over time that will generate the test tools that Veer talked about.
SPEAKER_04Yeah, had me an experimental. It's no fun if like the Chuck Yeager wouldn't be as exciting if somebody got in the plane before him, right? That's the whole fun of this thing is any storage you walk away from is a good storage, I guess would be the good equivalent
Where To Find The Speakers
SPEAKER_04on that one. So for folks that want to get connected to this amazing team, you know, outside uh outside of SNEA or in SNEA, Mark, what's the best way that folks can catch up with you uh IRL or or online somewhere?
SPEAKER_00Yeah, definitely, you know, my LinkedIn profile, you can contact me or uh any you can contact me through email, just my name, Mark Dutchran at Oracle, uh as well. You can get access to me anytime.
SPEAKER_04Fantastic. And Veer, where uh where do we find you in in your world?
SPEAKER_02Yeah, same. LinkedIn, you can search my name, you'll find it. And uh email address v.comar at Nutanix.com. Uh happy to help if someone wants to know how they can get benefited with it and they want to join.
SPEAKER_04And of course, there'll be lots of SDC goodness, both you know, the virtuals and and and in-person opportunities. So that's gonna be great. Looking forward, hopefully, to being at SDC uh and uh and seeing seeing all this stuff in in real life. It's so much nicer. And Jason, uh, for folks that want to catch up with you uh beyond the episode, where do we go?
SPEAKER_03Yeah, I mean, uh LinkedIn's definitely the best place to find me. So uh I'm only there's only a couple Jason Swicks around. So at Jason Swick on uh I think Jason Dash Swick at LinkedIn.
SPEAKER_04So you all have great names. I got the most generic name under the world's you know description. Uh that's what that's the reason I have to put disco posse everywhere because unfortunately ECE has my name. So once like I haven't got a uh a hope of if you search for Eric Wright, I'm like you aren't gonna find me, but you look at disco posse, and guaranteed I'm the only one of those ones. And and Michael, of course, uh being inside SNEA is the best place to find you. But uh, how do we make sure we do that?
SPEAKER_01Yeah, LinkedIn is great, and then uh you uh also can go to the pages that I referred to, and there's contact information at the on the pages.
SPEAKER_04That's right. We got lots of cool profiles, so yeah, I can encourage people to go check out the site. Uh, we've got a lot of great content. There's other fantastic podcasts we've had with some of these amazing humans and many more, and lots of other good learning resources. Absolutely. My thing I recommend to anybody is just look at the the learning that's there. And what you often find is that as soon as it's like LinkedIn in real life, like when we get there and we realize we're up here to each other and we've all got history together. We we really can be so much more connected in a way that we don't think of when we're just in the office. And this is so different than just like I love going to events, it's great to have the hallway track. But what's even better is when you can do like active collaboration and the hallway track, that's like it is just doubly awesome.
Closing Thoughts And Community Invite
SPEAKER_04So uh, so I'll definitely refer people. So head on over to Snea.org and uh for all the folks at AWS, if you're watching this, this is for you, Andy, and all the crew. Get on in. We need uh we need some more reps from the big shop. Um, but anybody else too that wants to get involved with membership, obviously just head to SNEA.org and uh there's nothing but goodness there. So thank you all for that, and we'll see you at the next Plug Fest.
SPEAKER_01Great, thank you, Eric. Thanks, Eric.