Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
KubeCon EU Paris News Recap
1.2k Plays7 months ago
Join Bhavin Shah and Ryan Wallner for a recap of announcments and news from KubeCon Paris 2024.
Kubernetes Community Days (KCD) in New York City on May 22nd, use the promo code “KUBERNETESBYTES” to get a 10% discount on your registration fees!
Nethopper
- Learn more about KAOPS: @nethopper.io
- For a supported-demo: [email protected]
- Try the free version of KAOPS now! https://mynethopper.com/auth
News
- https://about.gitlab.com/blog/2024/03/20/oxeye-joins-gitlab-to-advance-application-security-capabilities/
- https://www.redhat.com/en/blog/unveiling-red-hat-openshift-415
- https://developer.nvidia.com/blog/nvidia-nim-offers-optimized-inference-microservices-for-deploying-ai-models-at-scale/
- https://www.acorn.io/resources/blog/our-new-focus-developing-an-llm-app-platform-based-on-gpt-script-technology?fromOther=true
- https://loft.sh/blog/deliver-secure-kubernetes-multi-tenancy-with-new-vcluster-in-rancher-integration/
- https://www.observeinc.com/blog/stepping-on-the-gas/
- https://thenewstack.io/kubecost-2-2-covers-carbon-cost-monitoring-and-more/
- https://thenewstack.io/ovhcloud-unveils-roadmap-to-take-on-hyperscalers-from-europe/
- https://www.suse.com/c/meet-rancher-prime-3-0/
- https://www.suse.com/c/suse-releases-edge-3-0-highly-validated-edge-optimized-stack/
- https://www.fermyon.com/blog/introducing-spinkube-fermyon-platform-for-k8s
- https://www.cncf.io/blog/2024/03/19/announcing-the-ai-working-groups-new-cloud-native-artificial-intelligence-whitepaper/
- https://github.com/Azure/kaito
- https://azure.microsoft.com/en-us/updates/public-preview-kubernetes-ai-toolchain-operator-kaito-addon-for-aks/
- https://cloudnativenow.com/features/solo-io-delivers-on-cilium-support-promise-for-gloo-networks/
- https://docs.solo.io/gloo-network/latest/about/overview/
- https://github.com/kosmos-io/kosmos
- https://gateway.envoyproxy.io/blog/2024/03/14/announcing-envoy-gateways-1.0-release/
- https://newrelic.com/press-release/20240319
- https://siliconangle.com/2024/03/29/aviatrix-revolutionizes-networking-security-distributed-cloud-firewall-kubernetes-kubeconeu/
Recommended
Transcript
Introduction to Kubernetes Bites
00:00:03
Speaker
You are listening to Kubernetes Bites, a podcast bringing you the latest from the world of cloud native data management. My name is Ryan Walner and I'm joined by Bob and Shaw coming to you from Boston, Massachusetts. We'll be sharing our thoughts on recent cloud native news and talking to industry experts about their experiences and challenges managing the wealth of data in today's cloud native ecosystem.
00:00:30
Speaker
Good morning, good afternoon, and good evening wherever you
Bob's New Role at NetApp
00:00:33
Speaker
are. We're coming to you from Boston, Massachusetts. Today is April 12th, 2024. Hope everyone is doing well and staying safe. Bavin, I feel like it's been a while. What's going on, man? Dude, I have some news to share. I switched shirts recently.
00:00:52
Speaker
That's funny. No, I spent jobs recently. So this is, this is me doing the podcast from my new gig at NetApp.
Socks and Mountain Biking Fun
00:00:58
Speaker
Uh, I joined like the Astra product management team. So I'm excited. Very, very cool stuff. They're, uh, lucky to have you. Everybody give a little clap where we are. Congratulations.
00:01:11
Speaker
Thank you, Ryan. I'm excited. They have a good product and it's just about staying in the Kubernetes ecosystem and helping build products for the ecosystem that we do the podcast for. Yeah, we're really glad you're still here in the Kubernetes ecosystem. It would really be a big bummer if you went to a sock manufacturer or something like that. I don't know what I would do there, honestly. You'd run their Kubernetes Edge IoT. Oh, yeah.
00:01:37
Speaker
Something with AI as well, like monitoring through cameras how the socks are coming out, how efficient they are. Nice. Sure. I mean, it makes sense to me. It sounds like a pretty sweet job actually, because I do enjoy a good sock. I know. What's your favorite brand? I don't really have a favorite brand, to be honest. I've been into the, I think it's like Go Vermont socks. They're kind of like a hiking sock. They're super comfortable. Okay.
00:02:05
Speaker
Yeah. Like I don't have like I've heard good things about Bombas. I guess that's
Insights from KubeCon
00:02:14
Speaker
how you say it. Bombas are great. Our previous boss gave us a while ago. A while ago. Did you get any of those? I got them for Christmas. Okay. Maybe I didn't join port work soon enough.
00:02:25
Speaker
Not to get the socks. Like I missed out on them getting acquired, but the bigger bummer is like I missed out on socks. You didn't get bomb buss. This is what happened. Shout out to Sir Besh there if you're listening. Thanks for the bomb buss.
00:02:40
Speaker
How are you doing that? And like, how was your mountain biking trip? Yeah, you know, Arizona was beautiful. Um, you know, from coming to New Mexico in October the previous year and now to Arizona, you know, they're so close, but so different, right? Um, it was like seven days, 800 miles. We had a blast. We only had a couple of days where we got actually snowed on. Okay. Which was, you know, I mean, when you're north of Phoenix in the, like the Teton areas,
00:03:10
Speaker
Uh, it's kind of expected to be cold, but we did not expect to be snowed on. And it did. Um, and it turns out that the fine dust and sand turns to clay and it becomes very slippery and also sticks to all the bikes, like cement. And by the end of that day, uh, we had one guy down with a broken finger, like we had six bikes that were leaking oil from someplace. Oh, why?
00:03:34
Speaker
We just destroyed a hard day of riding, but it was really, really fun. I got.
00:03:41
Speaker
I think I came home with like two terabytes of video data. Nice. For filming that thing. So in the background, apart from this thing that we do here and the day job, I've also been trying to build that video for the company. I don't know if you want to give a shout out to your Instagram account. Like that's, that's a pretty fun way to follow you on. People call me Wally. This is kind of a funny thing. So my Instagram is Wally walls.
00:04:06
Speaker
Kind of a play on my last name, but you can see a whole bunch of stuff I'm working on there the the company that sort of I do this with is Modo Vermont Really fun stuff if you're into that kind of thing. I highly suggest go to talk to Eric and Spencer They're awesome. And yeah, you'll see some of the stuff that we talk about here
00:04:24
Speaker
Well, I know Bhavan and I, we didn't get to go to KubeCon, which is the total bummer for us. But at the same time, we are going to leave you high and dry and not do a KubeCon news episode. Obviously, this is coming from research. We didn't actually spend time there. I know I myself talked to a number of different people that did go. So we do have some kind of first-hand, second-hand accounts from the feeling and what
AI Innovations and Trends
00:04:51
Speaker
it was like. But we're going to kind of focus on some
00:04:54
Speaker
bigger ticket announcements maybe that you didn't hear of or that you did. And hopefully that'll be interesting for you. Oh, that's a great plan of action.
00:05:05
Speaker
I was like really product manager. I need to put it on the back. Okay. One funny story before we get started. So when I was at Lenovo, um, I had actually printed out stickers for my product manager because he always kept saying like, Oh, it's, it's in the backlog. It's in the backlog. So that's what the sticker said. And I found those stickers recently. I was like, yes, I need to put them on my laptop. I feel like I need that for my life. Oh yeah. It's in the backlog. Yeah.
00:05:36
Speaker
We got this. We got this. That's actually funny. I'll take one if you got extra.
00:05:44
Speaker
This episode is brought to you by our friends from Nethopper, a Boston-based company founded to help enterprises achieve their cloud infrastructure and application modernization goals. Nethopper is ideal for enterprises looking to accelerate their platform engineering efforts and reduce time to market, reduce cognitive load on developers, and leverage junior engineers to manage and support cloud-mandated upgrades for a growing number of communities, clusters, and application dependencies.
00:06:14
Speaker
Nethopper enables enterprises to jumpstart platform engineering with KAOPS, a cloud-native, GitHub-centric platform framework to help them build their internal developer platforms or IDPs. KAOPS is completely agnostic to the community's distribution or cloud infrastructure being used and supports everything including EKS, GKE, Rancher, OpenShift, etc.
00:06:51
Speaker
All right, well, I think without further ado, because we do have a number of news articles to get through. Why don't you kick us off with
00:07:00
Speaker
Yeah, sure. So I want to start with GitLab and an acquisition that they made during the KubeCon timeframe. So GitLab acquired a startup called Oxi. I think that's a play on bullseye, maybe? Yeah, definitely not Oxi.
00:07:15
Speaker
And the acquisition was more focused to advance GitLab's application security capabilities, right? So there's a term, a new acronym that I learned personally called static application security testing or SAST. So basically ox-sized capabilities. Another acronym.
00:07:33
Speaker
Again, I don't know how popular this one is. I just found out it in the press release that went out. But yeah, Oxi will help accelerate SAST roadmap inside GitLab. So they have a SAST scanner already, but this acquisition will help them streamline vulnerability management and remediation for developers, I think, as part of the CI CD pipeline. So cool acquisition and exit for any startup in the community ecosystem. So I just wanted to share that.
00:08:01
Speaker
Yeah, I mean, at the end of the day, right, we all have these tools like GitLab. If it's not GitLab, it's something else. And we're probably pumping through tons of applications, tons of images. And at the end of the day, like this is such an important part of the pipeline for us to like actually have these vulnerabilities and violations kind of given and visually kind of there for either the developers or security teams. So I'm all for it. I didn't hear of oxide before
Funding and New Projects in Cloud Native
00:08:27
Speaker
this. Yeah, same.
00:08:28
Speaker
All right, so my first one is going to definitely be from the WASM side of things. I've been digging in a lot there recently, both personally and in my day job. And if you haven't heard of Fermion yet, they're part of the ecosystem and one of the more popular frameworks to kind of get started. Their framework for kind of developing WASM applications is called Spin.
00:08:54
Speaker
And, um, previously I have worked with some of the Trinity shims, um, that allow you to run wasm applications through some of the integration points like wasm time and things like that, or was image. And, um,
00:09:09
Speaker
that all had sort of like a project on GitHub that you could go play around with. Well, during KubeCon, they came out with something called SpinCube, which was their biggest, their big announcement, which is really a Fermion platform for Kubernetes and running Wasm applications on.
00:09:25
Speaker
The cool part here is that it combines a lot of the things that were already being developed in the open source to the shims and integration points, but also combines this with a platform that's run by Fermion. If you wanted to get a fast start to developing and deploying these things on Kubernetes,
00:09:46
Speaker
Wasm's always had the benefit of being able to use a lot of the tool chaining that we've already been used to in the Docker ecosystem, even though that's not necessarily maybe the most efficient way to run it, but it does make it a really easy way to get involved and deploy these things on Kubernetes.
00:10:03
Speaker
to see SpinCube because it really kind of sets the tone, I think, for hey, Wasm and Kubernetes are sort of here to stay. We're kind of putting our money where our mouth is and where the ecosystem wants it to go. So if you haven't gotten to check that out, I think it's really interesting. Accessible, right? Yeah, for sure. I don't have to figure out how to run my Wasm workloads or develop it. Now with this platform on Kubernetes, I see that they also have an operator.
00:10:29
Speaker
that makes it easier for deploying these resources. And they have a kubectl plugin as well. So I think these are steps in the right direction. Yeah, no, for sure. And there's a combination too with everything that's happening with Gen AI as well. Spin is also doing a lot for basically using Wasm as sort of a universal runtime to run models.
00:10:53
Speaker
So on their platform, you can do things like have models built into the platform and run like inferencing jobs in Wasm as Wasm modules. So there's kind of a bleed there. I think it's really cool stuff. So I've been paying a lot of attention to it and just really like to see that they came out with this project.
00:11:13
Speaker
Interesting. Like, okay. So now I want to continue the thread around running models locally. I know we had the great episode with Johnny, but, uh, I was at Nvidia GTC the same week as coupon and, uh, Nvidia made a, uh, an announcement called nims, uh, and it provides like a con it is a containerized inference microservice that allows or provides a streamline path for developers to deploy, uh,
00:11:37
Speaker
AI models in production and use them with your new or existing enterprise applications. So Nim is basically a Docker container. So if it's built for portability, it enables model deployment across various infrastructure stacks, including Kubernetes. So you can actually run that container on your laptop where you are building your applications, or you can also run it in Kubernetes or on any one of the NVIDIA DGX systems in DGX Cloud or DGX.
00:12:02
Speaker
on-prem as well. They are pre-built containers and head charts that are packaged and optimized for specific models. We'll add a link to the blog that they published, but it's a simple way, right? Pull down your model and then package it up as a container and then just run it as you would any other container, but it allows you to download and run some of the LLM models locally. I think Nim is going to be really powerful, especially it has the backing of NVIDIA and all the work that they are putting in behind AI.
00:12:32
Speaker
I'm excited for NIMMS for sure. Yeah, I know. The NIMMS stuff was really fun. When I first looked into it, when it was announced though, it seemed like you couldn't actually get your hands on it just yet. Is it still that way? Oh, that's interesting. I went to actually deploy it and run with it.
00:12:49
Speaker
I don't actually know. It could be available. I'm just, I'm not sure, right? But like, because I haven't tried to run it myself, but I think based on the blog, if you have like an NGC or NetApp GPU cloud subscription or an account, I think you can download the models from there and then run it locally.
00:13:11
Speaker
Gotcha. Well, we'll have to follow up with that. And if there is a link for sort of like how to get started, I know the blog sort of shows how it's used, but in there it's kind of nebulous in terms of like, here's where you go to actually run the thing on a Kubernetes cluster if you have one ready to go.
00:13:28
Speaker
which is more what I was looking for. But really, really cool stuff. I know Nvidia has been up to, it seems like every day they announced some new partnership, new project or whatever. So this is something that I think we can all kind of get our hands on and sort of start touching things ourselves because
00:13:51
Speaker
This whole area is moving so fast that you really do have to, if you're a practitioner, have to get hands on and kind of learn how these different layers and how these AI pipelines work to kind of at least start putting your head around it from a practitioner's standpoint. Obviously, if you're just kind of a consumer, then...
00:14:07
Speaker
Just go for it. Just get a subscription to like perplexity or open AI or any of these things. You get access to models. Okay. No, I think, as I said, right. AI is the buzz. And that was, that was, I think the primary theme of, for KubeCon as well. Like everybody was talking about AI, AI, AI, maybe some platform engineering in the middle, but mostly AI. All I can see is Oprah, you get an AI, you get an AI.
00:14:33
Speaker
Like last year, I remember like two or three of the conferences I attended, like the vendor conferences, Jensen, the CEO of Nvidia was on the keynote and talking about how they work closely and how the part, the vendor is the best thing ever. I expect like with the push that CNCF has, do you expect to see like a Jensen, even a virtual, like a Zoom session or a virtual recorded thing for the KubeCon keynote in Salt Lake City? That would be cool.
00:14:59
Speaker
Yeah, that would be kind of cool. I mean, just to kind of showcase the technology as well. You know, one thing I did here, even though this was the largest KubeCon ever with 12,000, I think was the number. What I heard was even though they put so much emphasis on AI and they gave the tracks and the talks that were AI related to the big rooms, as they expected, a lot of those rooms were
00:15:26
Speaker
empty, not empty, but like you did not fill up. So, um, I don't know. There's, there's the, there's kind of a two sided thing where
00:15:35
Speaker
people that went and sort of said, the hype is there because I feel like that's what the market wants versus what people were actually going to and listening to was maybe a little different. Interesting. Yeah. I think that's a good perspective, right? Like everybody wants to talk about AI to show that they are relevant, but then do users really want every everything around AI right now? They may be still experimenting with it or they might still be in the development phase. They don't need everybody, all their vendors to give AI solutions. To a certain degree, it's a little bit of a toy to them, right?
00:16:05
Speaker
they're kind of playing around with it. Anyway, just an anecdote that I thought was interesting because I think often what the market hypes, especially in AI, might be a little bit different from the reality of many of us, even though a lot of our companies are definitely investing in it. That's no doubt, but it's early days, that's all.
00:16:25
Speaker
I'm going to keep going with the AI theme. Why not, right? Because we're going here. The CNCF announced the AI working group's new cloud native
00:16:38
Speaker
artificial intelligence white paper. That's a mouthful, but basically it's an AI white paper that is really an overview of how those sort of AI and ML tool chains and techniques, how they kind of work with the cloud native technologies, what's offered there, what are the challenges, what are the gaps, what is sort of the development and usage of these things look like. A lot of people might be,
00:17:04
Speaker
sort of hearing even from this podcast with Johnny of like, oh, you can run it on Kubernetes or I know it is, but how does it work, right? These white papers and especially from CNCF might give you a little more in-depth reading for... Yeah.
00:17:21
Speaker
how this is all put together. I think it's a good resource. The CNCF usually does a really good job of recruiting the right people to write these things and whatnot. I haven't read through the whole thing myself, but just glancing through, it seems pretty good. No, I think, as you said, glancing through, they also have a neat little cloud native AI diagram.
00:17:42
Speaker
I think it's what you would expect from the hardware all the way up to applications. And I like how they are still keeping the predictive AI piece around. I know we have spoken about how predictive AI, even though gen AI or generative AI is getting a lot of the marketing buzz and the dollars, predictive AI is actually the ones that people are.
00:18:02
Speaker
utilizing and building real applications. Gen AI is still like people are figuring out like what outside of chatbots is something that they can do like or what maybe building rack models right so I think yeah I like that they have included both of those and basically serve to all of the hardware partners so Intel and video arm Google Cloud and AMD everybody who provides the GPU is on that list so
00:18:23
Speaker
Yeah. And if you've seen the CNCF map of projects, and it's quite complicated, well, they include a specific mind map for all the AI-specific stuff. And it's already gigantic, and it's new, right?
00:18:38
Speaker
I think we're keeping with the trend there, but there's, I think the mind map is super useful because it's hard for us to really pick up on every project that's out there. So anyway, we're moving on. Okay. More, more AI stuff. You know, like I'm just moving things around because we are on this AI. Keep it rolling. So a startup in the ecosystem, right? Akon.io. I think they, they announced, they were, they, they, they have been around for the past 12 or 18 months. I know I saw them at KubeCon Chicago with a, with a, with a specific booth and
00:19:08
Speaker
The founders for that, for Acorn are really popular. They are the guys that built Rancher, so Darren and those guys. But when they started, the whole thing was, oh, instead of a Docker file, we'll have something called as an Acorn file where you can define how your application should be deployed. And once you apply or use that Acorn file against your Kubernetes cluster, it automatically deploys all the pods and services and load balancers that you might need.
00:19:31
Speaker
That was great. But now with the AI boom, they are pivoting. So they are now moving their focus completely to an open source LLM stack that's based on their own GPT script technology.
00:19:44
Speaker
The pivot is still kind of only a couple of months old, so I don't think they have a full-fledged product yet. I know if you follow Darren on Twitter, which I do, or X, he likes to share a lot of how he's building the thing, so it's an interesting follow. But I think the problems that they are solving for are things like how you can answer your questions about your own data, about in your own intranet, on your own databases, on your own applications. It's building everything
00:20:10
Speaker
still open source, still built for LLMs, but instead of sending your data to OpenAI or Microsoft, how you can run these proprietary stacks or handle your proprietary data inside your own data center. Then how can you implement these LLMs in resource and network constraint environments? Then how can you go beyond that chat interface and apply LLMs to your existing application? Definitely a new focus, a vendor that I'm excited about because they always do great things. This is a vendor that I'm keeping my eye out on.
00:20:38
Speaker
Yeah, I mean, acorn was definitely I always felt like it was a little late to the game in terms of like application composition when Docker compose ages and ages. So I don't I don't really hate the pivot. I think it's really cool. And it is tackling a very specific problem about how we interact with these LMS. Right. And there are languages and kind of program languages out there that allow you to kind of
00:21:03
Speaker
script these things. So I'm all for this kind of pivot. And if we make it easier to interact with these LLMs, ultimately at the end of the day, we can get more out of them. So I haven't checked out GPT script myself, but it seems pretty interesting. And I think it's exciting. Good for them. Yeah. That's what I say. Nice.
00:21:27
Speaker
Um, cool. So, uh, the next one I have is the project Kato, which, um, if I sang that right, I think I sang that right.
00:21:38
Speaker
Kato is a project from Microsoft on the Azure side. It is the Kubernetes AI toolchain operator. Again, just flow in this AI theme here. I think it's one of the last ones I have on my list. I think I'm done. If you're sitting here, we're like, is this going to be all AI news? We're almost done, I promise.
00:22:04
Speaker
This AI toolchain basically allows you to automate the AI and ML inference model deployment in a Kubernetes cluster. We have also been chatting with... What's his name again? I think it's a her, but yeah, it's a product manager for Kato inside Microsoft.
00:22:23
Speaker
We're hoping to have her on the show. So if you're interested in learning more about Kato, hopefully in a future episode, we'll be able to chat more about the project and kind of get into some of the details. But really, it really targets some of the more popular large open source models, like Nalcan or Llama 2. And if you're looking at kind of a way to get going on a Kubernetes cluster pretty quickly,
00:22:49
Speaker
Kato operator might be just the tool for you. I don't know. I'm excited, right? Because Microsoft does so many things with AI, like they are the cloud vendor who started the whole thing, right? With open AI and publishing these models as a model garden or something. So I think Kato should allow people to use the models that are available on the Azure model garden with their applications running on AKS cluster.
00:23:16
Speaker
Yeah, it's a good point. This was specifically, I think, designed for AKS and it's in preview, I think. Yeah, public preview. Yeah. Yeah. Public preview as of March or something. Yeah. So we'll get more details when we have the people who are actually working with the project on the board. But I think good shout out here to wrap up our AI discussions. Yes. Are we done? I think so. That was more than I actually thought we had. Yeah.
00:23:40
Speaker
But it built a good flow, I think. I think it tracks with, you know, KubeCon. So I think it makes all sense. Yeah. Okay. For me, the next one is Red Hat, right? They always make a series of announcements at any KubeCon events. For the Paris one, OCP or OpenShift container platform is 4.15 is now generally available.
00:24:03
Speaker
If you guys don't know, they do a YouTube live stream, a month in advance, talking about all the capabilities and they have all their PMs. It's like an hour, hour and a half session where every PM talks about the features that will be in OCP 4.15. I'll try to find a link and add it to the show notes. But since it's now GA, we can talk about some of the features.
00:24:22
Speaker
If you are an OCP customer, you can now, and you want it to run OCP or OpenShift on AWS Outposts or AWS Wavelength for your Edge applications, those things are now fully supported. So you can run your OCP worker nodes on Outposts or Wavelength sites, just worker nodes, right? Because those Edge environments are resource constrained. So you can run the control plane on AWS and then have those worker nodes running locally at the Edge. So that's now supported. If you are using
00:24:48
Speaker
OpenShift Virtualization, or you are contemplating using OpenShift Virtualization, their storage solution, OpenShift Data Foundation, now supports Metro DR capabilities. So SyncDR between two OpenShift clusters in a stretched configuration is now supported as well, or it is generally available. And then a couple more things, OpenShift Service Mesh can now support ARM-based architectures as well. So now that ARM
00:25:15
Speaker
ARM architecture for service mesh is also GA. And then one final thing is, if you are running VMs using OpenShift word, dynamic reconfiguration of the NICs that you might have for those VMs is also supported. So those are just like three or four things that caught my eye. Obviously, we have a whole long list of things that they announced. But yeah, OCP 4.15 is now GA.
00:25:36
Speaker
Yeah, tons of little announcements in here. So if you're interested or use OpenShift, definitely worth going to take a look at all what they have going. I know there's been a lot of interest in OpenShift Virtualization. Oh, yeah. Even people are figuring out whether they want to stay with VMware or how that works if they can't, and where to go. I know Foodbert or even OpenShift Virtualization and OpenShift has been a huge
00:26:02
Speaker
focus for a lot of people. So definitely worth going and checking out and seeing that going over there.
00:26:09
Speaker
Okay, so my next one is from Solo.io. So if you're familiar with Solo, they're sort of in sort of the service mesh space. What they announced at KubeCon was support for Glue Networks, in Psyllium that is. So we had talked with, was it folks from Psyllium? I know we had a podcast about EVPF and Psyllium, the Kubernetes integration.
00:26:37
Speaker
for networking works on eBPF. And so Glue Network from solo.io is now supporting psyllium. So that kind of enhances what you can do out of the box with psyllium by default. So what Glue Networks is kind of saying is that you get a little bit better performance, you get a more
00:27:01
Speaker
Support you get more sort of insight and observability. I know with glue networks you get a nice dashboard And all sorts of things like that like supporting multiple Multi-tenancy and policies for zero trust security and all sorts of stuff If you're looking to kind of up level what you're doing with psyllium definitely go check out What the support from so I oh, I know a deep is an old Colleague of mine. Oh nice. They're what?
00:27:31
Speaker
What they're up to has always been really, really good stuff. So go check that. And if you, if you have to do an episode around it, I guess Ryan already has a contact. So yeah, I'm sure I do. We might, might go on. Um, I'll have to reach out to her. Um, maybe if, uh, some of our audience wants to learn more about what solo IO has to offer, um, reach out ping us on Slack or any social media, let us know you'd like to hear about that. And we'd.
00:27:57
Speaker
Probably, happily get them on. Awesome. I'll follow that up with like a funding round from a vendor in the observability space called Observe. That's what they're called. I don't know how they call it. Well, at least they're creative.
00:28:15
Speaker
I'm just surprised how they got their domain. I know the domain is observing.com, but still, Observe is a good name to register if you are in the Kubernetes observability space. They raised a series B round worth $115 million. That was the amount of money raised. Based on
00:28:32
Speaker
It's not like a brand new company, right? Series B is that they have found like a product market fit and they're now ready to scale as well. So they have a platform that allows you to aggregate all of your telemetry data. So logs, metrics, traces, anything else that you might have and store them as events in one consolidated database. So it's possible to have like one language or one way to manipulate all of that data without having to leave and context switch and go between different tools.
00:28:58
Speaker
So if you have to do any troubleshooting, if you don't have to do any root cause analysis, you can just use the observe UI to do everything. So go from a metric to a log or from a trace to a log without leaving observe is the whole thing. So a cool vendor to check out if you are in the need or if you are looking for observability solutions.
00:29:17
Speaker
Speaking of naming, I came across this project called Cosmos, which is my next article here, which was I found it through someone else who did a KubeCon review. Shout out to tetrate.io, their blog on insights and highlights from Paris.
00:29:34
Speaker
Turn me on to this project and it's called Cosmos with a K, of course, for Kubernetes. And the whole idea is it symbolizes limitless expansion of Kubernetes. So really what it actually is, is an approach to Istio multi-cluster networking. And they have something called cluster link, cluster tree, and a scheduler, which basically allows pods in different clusters or in services in different clusters to talk to each other.
00:30:05
Speaker
We've had a number of folks on the podcast that had different approaches to this and even entire companies that are built around this kind of thing, right? So this kind of creates a tunnel link between clusters and manages how all that works, but I think it mostly uses the
00:30:22
Speaker
cluster networking, APIs. That might not be the right name for it, but it uses a lot what's in there in Kubernetes already and allows you to work with pods and services and pods to service and all these kind of things with, of course, gateways so that you can connect those clusters and stuff together. So as far as being an open source project and being something that's talked about at KubeCon, I thought it was a really cool way to
00:30:52
Speaker
showcase this kind of thing that we've kind of come at from a vendor perspective before. So if you haven't seen this project yet, I definitely would go check it out. It's github.com slash cosmos dash IO. Interesting cosmos. Yeah. Cool project, Ryan. I need to check that out. I don't know much about it. Yeah, neither did I. I still don't, honestly. Although you're reading through it.
00:31:18
Speaker
Okay, talking about, I think vendors or things that we have spoken about on the podcast, CubeCost, another vendor from our Kubernetes Bites alumni list, they announced a cool new feature in their 2.2 release. I know it had other things, but the one that caught my eye was more around
00:31:35
Speaker
how they can help you measure your carbon footprint for your Kubernetes workloads. Again, if your organizations have a specific goal to meet, if you want to figure out which applications or which business units are spending or using the most carbon credits that you have available,
00:31:50
Speaker
It gives you an option to now monitor those using the cube cost UI and it helps you conduct energy audits and then compare financial and environmental costs across different business cost centers. So cool new feature in addition to the cost management or optimization that they do, carbon cost monitoring feature is super cool. I know Red Hat had an open source or still has an open source project called Kepler, which kind of does the same thing. So I think we have multiple options.
00:32:17
Speaker
for if you are looking to monitor how much power your community source is using. Yeah, I mean, this is always something that we have to be aware of. And I feel like, you know, when you start to look at
00:32:31
Speaker
various use cases like Gen AI specifically and how quickly that can ramp up and how much capacity can be used, right? These tools at least make it front and center for us as technologists to kind of think about the carbon costs that we're putting out there because sometimes it's really hard to really wrap your head around it.
00:32:51
Speaker
And I think there are data center providers, right? Like I know AWS and Azure and Google Cloud are the big hyperscalers, but there are smaller GPU cloud providers like Caruso Cloud. I think I saw them or met them at Nvidia GTC and they have an interesting solution. So instead of
00:33:08
Speaker
building data centers where the population density is higher and power consumption might be an issue. They are building it near hydroelectric plants or some other form of electricity generation can also be nuclear plants, I'm not sure.
00:33:27
Speaker
in other regions, of course. But they're like running fibers or network between your CDNs and the data center is cheaper than figuring out how to get enough power to a data center. So I think they have an interesting way of doing things. Check them out if you are looking for a solution or a data center provider that's focused on providing you GPU capacity from these data centers that are located in random locations.
00:33:54
Speaker
Very, very nice. Very nice. So the next one I have is from Envoy Gateway. So their 1.0 release, big release for them, I think was March 14th, right before KubeCon. So 1.0 of this release makes Envoy Gateway for Kubernetes generally available, ready to use for production.
00:34:20
Speaker
which basically takes the Kubernetes Gateway API and manages Envoy proxy to provide all those types of things. So rate limiting, OAuth2, Gateway APIs, clients, backend, security, all these kinds of things built into this one point of release. So really exciting stuff from them. I know they've been doing a lot of great work over the days. So we probably should get someone on from
00:34:49
Speaker
maybe this project to chat about it as well. Yeah. I like, like in the 1.0 announcement blog, right? They're like, what does 1.0 mean for the project? And the first bullet is commitment to ensuring stability and fixes for any CVs that might come up in the project, which is, which is obviously a good thing. Like if you're trying to use ongoing production for your Kubernetes, you want to make sure that the community is supporting this project. So 1.0 is a big milestone. Congrats guys. Yes. Congrats.
00:35:14
Speaker
Next up, I have a few announcements that are focused around the Rancher project in one way or the other. Let me start with LoftLabs. We had Lucas on the pod earlier this year to talk about their vCluster solution and how it allows you to deploy these virtual clusters on a parent Kubernetes cluster or an underlying Kubernetes cluster.
00:35:34
Speaker
Now, they have extended that integration with Rancher. If you have a Rancher cluster, if you are using the Rancher UI, you can now deploy and manage your virtual clusters along with your Rancher cluster. You can have a big Rancher cluster and have tens or hundreds of virtual clusters. If you are using the Rancher UI to manage all your different Rancher clusters, you can manage your virtual clusters or V clusters running on Rancher from the same UI. I think it's a pretty cool integration
00:36:00
Speaker
I know that they did support EKS and AKS and GKE, I think before, but they never had integrations into the UI, right? So you always had to use your CLI tools, again, not a drawback, right? Like people in the SQL are familiar with using CLI tools and kubectl extensions, but having that integration built into the Rancher UI, I think makes it really easy for people that are new to the ecosystem. So enabling self-service, enabling that quick spin-up time of a virtual cluster, a great solution together from SUSE and LoftLabs.
00:36:30
Speaker
I think, as I said, this is going to be a theme of rancher announcements. The next one was from a cloud provider in the European region. They're most popular in the European region. I know they have US regions as well, but called OVH Cloud. So now they are entering the managed Kubernetes race or managed Kubernetes distribution race.
00:36:54
Speaker
by partnering with SUSE and announcing a managed transfer solution. So if you are an OVH customer and you are missing out on any managed Kubernetes distribution, you have a solution now. I think it made sense that they are based out of the
00:37:10
Speaker
I don't know which country in Europe they're based out of, but they are a European CSP. So I think this makes sense to do the announcement and make a bank at KubeCon Europe. I think they were also part of one of the sponsored keynotes as part of the CNCF keynote, and they announced their partnership with SUSE and how they have a new managed service. So again, something to check out if you are an OVH Cloud customer.
00:37:33
Speaker
Yeah, and I believe in that article where LVH announced it, it specifically calls out that it also comes with that vCluster support. Oh, yeah. Yeah. That's a good point to highlight. Yes. Yeah. So they do connect. Yeah, it's fun. Like these people are actually thinking about things we were announcing. So like it actually works. I'm happy. And then finally, like since we're talking about Rancher, can't skip the announcements from SUSE, I know we had
00:38:02
Speaker
We had done an episode on K3S earlier this year. Now, SUSE also is announcing Edge 3.0. Edge 3.0 stack, which is a purpose-built stack to run in resource-constrained environment or remote locations with intermittent internet connectivity.
00:38:19
Speaker
Edge 3.0 from SUSE is underpinned with SUSE Enterprise Linux micro, so SLE micro version, and run certified Kubernetes distributions like K3S or RKE2. It's a fully managed stack or supported stack from SUSE, so if you are looking for
00:38:36
Speaker
deploying Kubernetes on your edge location, something to check out for sure. And then to follow that up, they also announced a 3.0 version for their Rancher Prime offering, which now streamlines cluster deployments. It offers centralized authentication for all different Kubernetes distributions. So you can use this one UI or one tool to manage all different Kubernetes clusters, EKS, AKS, GKE, Rancher, et cetera. It also includes options for integrating with, I think, new vector
00:39:05
Speaker
for container security and then also they have the long-haul projects. They also have persistent storage and data management as part of the Prime 3.0. Interesting announcement. They also focused on security. They are Salsa or secure-level software artifacts compliant. I would never. Since I have it in my notes, I know what Salsa stands for. If you ask me on the street what does Salsa mean, I wouldn't be able to answer the full form.
00:39:32
Speaker
So it's also sales are compliant and they have a shareable S-WOM. That I can answer. That's also making me hungry. Yeah. Nice. No, I think we should take a lunch break after this.
00:39:44
Speaker
And they have a shareable S-bomb, so you actually know all the different container images and packages that are part of Rancher Prime 3.0. So you can run your own scanning and figure out if there are any new CVs and you can fix them. So overall, great announcements for the Rancher ecosystem, everything from Loft Labs to OVH Cloud to SUSE. So I'm excited about that being or getting more users in this ecosystem.
00:40:08
Speaker
Absolutely. And I know we've been talking a lot about platform engineering in recent episodes, specifically Prime 3.0 does call out that it helps build platform engineering sort of teams deliver a platform as a service to their developer community. So it might help you out there as well if you're looking at kind of creating that for your teams as well.
00:40:31
Speaker
And with that, I'm done with my news, so I know what to use. Done. OK, so I have two more to go here. The first one is from New Relic, basically adding observability, sorry, support for observability and telemetry when it comes to Kubernetes hosts. So this basically takes the integrations with Prometheus and Hotel
00:40:58
Speaker
and allows you to kind of pull them into your New Relic workflows to give you sort of what they're calling a all-in-one observability platform for every engineer. Because obviously New Relic does a lot more than just Kubernetes. And honestly, I was a bit surprised that they haven't supported previous and in the past, although Hotel definitely is something that's coming more and more mainstream. So if you're a New Relic user,
00:41:25
Speaker
Go check that out. You can go and take advantage of that for Kubernetes workloads. The final one that I have here is from aviatrix. I probably am saying that right and maybe not. No, that sounds great.
00:41:42
Speaker
They announced a distributed cloud firewall for Kubernetes. So basically claiming that managing the security and operational complexities of firewall is a big pain point for customers. They are going to deploy
00:42:04
Speaker
this distributed firewall. And I'm all for this kind of project because I feel like it kind of, it's, it's still showing us that there's opportunities, especially for new companies to go out and tackle problems that maybe don't have to do with AI or platform engineering, all these things are maybe tie into those. Um, uh, but also this kind of addresses multiple aspects, right? So this now that Kubernetes can run virtual machines and containers, you have, you know,
00:42:32
Speaker
quite a complex array of applications and IP addresses and all these things. So managing your firewall as sort of a distributed software platform as well makes a lot of sense. So go check out Mitch Connors did an interview with The Cube, I think, in the article we have. You can learn more about what they're up to for sure.
00:42:55
Speaker
Interesting, dude. What I realized as you were talking about aviatrix is we didn't have a lot of security vendors in the news surprisingly. You remember a couple of years back? Unless we missed them, which is a very real possibility. Bobbin and I can only do so much research not being physically there.
00:43:14
Speaker
That's actually a great segue to say, if we did miss anything big or on your radar as a listener, let us know. And we'll include it in some of the upcoming episodes to kind of backfill what we missed. Obviously, if we tried to tackle all the news, you'd be here for hours and hours and hours, and I doubt you would all want to listen to us. Oh, come on. You're not listening to us, man.
00:43:39
Speaker
Just saying, you know, I think this is a good balance, right? Like there is only one podcast that I listen, which is over three hours. But other than that, like, even in you do that, do you listen to it in spurts? Yeah, spurts for sure. Yeah.
00:43:55
Speaker
And I don't listen to the last 30 minutes where they share analysis or things like that. So it's a podcast called Acquired. They break down companies. I think they used to break down technology companies. So it was really cool to listen how the other companies started and what are the different evolutions that it went through. And then at the end, in that three-yard episode, the last 30 minutes is them analyzing modes and figuring out how the company will survive, things like that. So I don't listen to that part. But overall, a good episode. Which makes me think, if you don't listen to that part, and we all have a part we don't listen to,
00:44:25
Speaker
podcast. We should put out a poll and figure out where people are skipping on ours, right? Is it the beginning where we like catch up or is it, hopefully it's not the news because this is not the episode for you. Why wouldn't you want to hear about our new jaws and what are the mountain biking trips? Come on, man. Okay. You know, we got to make it fun at the end of the day, uh, cause Bob and I really, we're here to have some fun. All right. So, uh, again, um,
00:44:55
Speaker
If you're thinking about attending Kubernetes Community Days, KCD in New York City on May 22nd, we've been kind of promoting this for the last few episodes, the promo code KubernetesBytes gets you 10% off your registration. So again, that's Kubernetes Community Days or KCD New York City, May 22nd. I still haven't verified it. It's a medicine, dude. Come on, let's do it. I'd like to say that.
00:45:20
Speaker
I know. Don't look it up every time. I think the tickets are getting close to being sold out. So, like, Ahmed is one of the lead organizers, right? Somebody who we had on the podcast earlier this year. And he just keeps sharing updates. So, man, it's turning out to be a great event. So, if you don't want to miss out, you support Community Spikes and register for the event. Come on.
00:45:41
Speaker
Yes, please do if you're local to New York City. I know a lot of smaller events are struggling this year, and I think that's due to economics, travel and expense, all this stuff, but go show support for your local meetups and events. Please do.
00:45:59
Speaker
Lastly, please, if you know somebody that might enjoy this podcast and learn a few things, please share with them. Let them know about the show and a bonus if they've reviewed the show. Give us five stars and share a screenshot. We'll give you a little shout out on the show.
00:46:21
Speaker
And if you're a US based because we do have the limitation we'll send you some swag or stickers or something like that But please we're trying to get the word out for the show a little bit more this year and do some more interesting things So we'd love to hear more from you our listeners and so please share the podcast for us
00:46:37
Speaker
Yeah. And use our social channels as well, right? If you don't want to share one-on-one or if you want to do something else in addition to one-on-one, just tag us on LinkedIn. Like just today, I think I saw the CTO for Wizz security give a shout out to our Wasm episode that he listened to that to learn more about Wasm. I was like, that's great. Like he now expanded our listeners to his network. So I know that like tag us. It's just Kubernetes bytes everywhere. So yeah, everywhere.
00:47:07
Speaker
All right. Well, Bobbin, that was a lot of fun. Hopefully we'll be both at KubeCon in November in Salt Lake City. We're going to be there and get some real in-person news this time. But that brings us to the end of today's episode. And I'm Ryan. I'm Bobbin. Thanks for joining another episode of Kubernetes Bites. Thank you for listening to the Kubernetes Bites podcast.