 Welcome back everyone to Kube SuperCloud 22. I'm John Furrier, your host, got a great influencer, Cloud Arati segment with Sarbiet Joal, cloud influencer, cloud economist, cloud consultant, cloud advisor. Sarbiet, welcome back, Kube alumni. Good to see you. Thanks, John. Nice to be here. Now what's your title? Cloud consultant analyst. Consultant actually, yeah. I'm launching my own business right now, form formally soon. It's in stealth mode right now. We'll be out. Why just call your cloud guru, cloud influencer? You've been great. Friend of the Kube, really powerful on social. You share a lot of content. You're digging into all the trends. SuperCloud is a thing. It's getting a lot of traction. We introduced that concept last re-invent. We were riffing before that as we kind of were seeing the structural change that is now SuperCloud. It really is kind of the destination or outcome of what we're seeing with hybrid cloud as a steady state into the, what's now they call multi-cloud, which is kind of awkward. It feels like it's default, like multi-cloud, multi-vendor. But SuperCloud has much more of a comprehensive abstraction around it. What's your thoughts? As you said, as Dave said, he says that too, that the SuperCloud has that abstraction built into it. It's built on top of cloud, right? So it's being built on top of the CAPEX, which is being spent by likes of AWS and Azure and Google Cloud and many others, right? So it's leveraging that infrastructure and building software stack on top of that, which is a platform. I see that as a platform being built on top of infrastructure as code. It's another platform, which is not native to the cloud providers. So it's a kind of cross-cloud platform. That's what's here. VMware calls it that cross-cloud. I'm not a big fan of the name, but I get what they're saying. We had a segment on earlier with Adrian Kalkroff, Lori McVity, Chris Hoff, all part of the cloud already like us, ourselves and you've been involved in cloud from day one. I remember the OpenStack days, early cloud, AWS, when they started, we saw the trajectory and we saw the change and I think the OpenStack in those early days were tell signs because you saw the movement of API first but Amazon just grew so fast and then Azure now is catching up. Their CAPEX is so large that companies like Snowflake, like why should I build my own? I just sit on top of AWS, move fast on one native cloud, then figure it out. Seems to be one of the playbooks of the super cloud. Yeah, that is true and there are reasons behind that and I think number one reason is the skills gravity, what I call it. The developers and or operators are trained on sort of one set of APIs. And I would said that many times to out-compete your competition, you have to out-educate the market and we know which cloud has done that. We know what traditional vendor has done that in 90s it was Microsoft, they had VBS number one language and they were winning. So in the cloud era, it's AWS. They're marketing efforts, they go to market strategy, the micro nature of releasing the micro sort of features if you will, almost every week there's a new feature. So they have got it, you know? And other two are trying to mimic that and they're having little trouble like. Yeah, and I think GCP has been struggling on compared to the three. At native cloud on native as you're right, completely successful. Azure caught up and you see the Microsoft, I think is a great selling point around multiple clouds. And the question that's on the table here is, do you stay with a native cloud or you jump right to multicloud? Now multicloud by default is kind of what I see happening and we've been debating this love to get your thoughts because, you know, Microsoft has a huge install base. They've converted to Office 365. They even throw, you know, sequel databases in there to kind of give it a little extra bump on the earnings, but I've been super critical on their numbers. I think their shares are, there's over, clearly overstating their share in my opinion, compared to AWS as a native cloud, Azure though is catching up. So you have customers that are happy with Microsoft and they're going to run their apps on Azure. Yeah. So if a customer has Azure and Microsoft, that's technically multiple clouds. Yeah, true. And the strategy, it's just an outcome. Yeah, I see Microsoft cloud as, you know, friendly to the internal developers, you know, internal developers of enterprises, but AWS is a lot more ISV friendly, which is the software shops friendly, right? So that's what they do. They just build software and give it to somebody else. But if you're in-house developer and you have been a Microsoft shop for a long time, which enterprises haven't been that, right? So they have, Microsoft is well-entrenched into the enterprise, we know that, right? For a long time. Yeah, and the old joke was developers love code and just go with a lock-in and then ops people don't want lock-in because they want choice. Yeah. So you have the DevOps movement, it's been successful. Now you got DevSecOps. The real focus to me, I think, is the operating teams because the ops side is really, with the pressure is serving, I want to get your reaction because we're seeing kind of the script flip. DevOps worked, infrastructure as code has worked. We don't yet see security as code yet. And you have things like cloud native services, which is all developer goodness. So I think the developers are doing fine. Give them a thumbs up and open source is booming. So they're shifting left, CICD pipeline. It's issues around repo, monolithic repos, but devs are doing fine. It's the ops that are now have to level up because that seems to be a hotspot. What's your take? What's your reaction to that? Do you agree? And if you say you agree, why? Yeah, I think devs are doing fine because some of the devs are going into ops. Like the whole sort of movement behind DevOps sort of culture is that devs and ops is one team. The people who are building that application, they're also operating that as well. But that's very far and a few in enterprise space. We know that, right? Big companies like Google, Microsoft, Amazon, Twitter, those guys can do that. They're very tech savvy shops. But when it comes to, if you go down from there to the second tier of enterprises, they are having hard time with that. Once you create software, I've said that, I sound like a broken record here. So once you create software, a piece of software, you want to operate it. You're not always creating it, especially when it's in-house software development. It's not your core sort of competency to, you're not giving that software to somebody else. Or they're not multiple tenants of that software. You are the only user of that software as a company. So or maybe maximum of your employees and partners, right? But that's where it stops. So there are those differences. And when it comes to ops, we have to still differentiate the ops of the big companies, which are tech companies, pure tech companies, and ops of the traditional enterprise. And you're right. The ops of the traditional enterprise are having tough time to cope up with the changing nature of things. And because they have to run the old traditional stacks, whatever they happen to have SAP or financial whatnot, right? Thousands of applications, they have to run that. And they have to learn on top of that. New scripting languages to operate the new stack, if you will. So for ops teams, they have to spin up operating teams for every cloud specialized tooling. There's consequences to that. Yeah. There's economics involved, the process. If you are learning three cloud APIs, and most probably you will, you'll end up spending a lot more time and money on that. Number one, number two, there are a lot more problems which can arise from that, because of the differences in the API, how the APIs work. The worst is if you pick one primary cloud and then you're focused on that and most of your workloads are there, and then you go to the secondary cloud, number two or three on as-need basis, I think that's the right approach. Well, I want to get your take on something that I'm observing. And again, maybe this is because I'm old school, I've been around the IT block for a while. I'm serving the multi-vendor. Vendors kind of, as Dave calls the calisthenics, they're out in the market trying to push their wares and convince everyone to run their workloads on their infrastructure. Multi-cloud to me sounds like multi-vendor. And I think there might not be a problem yet today. So I want to get your reaction to my thoughts. I see the vendors pushing hard on multi-cloud because they don't have a native cloud. I mean, IBM ultimately will probably end up being a SaaS application on top of one of the CapEx hyperscales, some say. But I think the playbook today for customers is to stay on one native cloud, run cloud native hybrid, go in on one cloud and go fast. Then it gets success and then go multiple clouds versus having a multi-cloud set of services out of the gate. Because if you're VMware, you'd love to have cross-cloud abstraction layer, but that's lock-in too. So what's your lock-in? Success in the marketplace or vendor access? It's tricky actually. I've said that many times that you don't wake up in the morning and say like, we're going to do multi-cloud. Nobody does that by choice, right? So it falls into your lab because of mostly because of MNAs. And sometimes because of the price-to-performance ratio is better somewhere else for certain kind of workloads. That's like far and few, to be honest with you. That's what my read is, that being a developer and operator of many sort of systems, if you will. And the third tier is that the, which we talked about during the VM world, I think 2019, that you want vendor diversity. Just in case one vendor goes down or it's broken up by Fed or something and you want another vendor, maybe for price negotiation tactics or whatnot. That's an optimal mentality. Yeah, yeah. And that's true. They want choice. They want to get locked in. You want choice because, and also like things can go wrong with a provider. We know that we focus on top three cloud providers and we sort of assume that they will be there for the next 10 years or so at least. And what's also true is not everyone can do everything. Yeah, exactly. So you have to pick the provider based on all these sort of three sets of high level criteria, if you will. And I think the multi-cloud should be your last choice. Like you should not be giving up for that by default, but it should be by design, as Chuck said. Okay, so I need to ask you, what a super cloud in my opinion look like five, 10 years out? What's the outcome of a good super cloud structure? What's it look like? Where did it come from? How did it get there? What's your take? I think super cloud is getting born in the absence of having standards around cloud. That's what it is. Because we don't have standards, we long or we want the services at different cloud providers, right? Which have same APIs and there's less learning curve or almost zero learning curve for our developers and operators to learn that stuff. Snowflake is one example and VMware stack is available at different cloud providers. That's its infrastructure as a service example, if you will. And Snowflake is a sort of data warehouse example and they're going down the stack while they're trying to expand. So there are many examples like that, right? So it's, what was the question again? Is super cloud 10 years out? Yeah, 10 years out. What's it look like? What's the components? Yeah, I think the super cloud 10 years out will expand because we will expand the software stack faster than the hardware stack. And hardware stack will be expanding of course with the custom chips and all that. There was a huge event yesterday was happening from AWS Silicon day. And that's an eye-opening sort of movement and the whole sort of technology consumption, you know, I feel well. And they had the differentiation with the chips and their supply chain kind of hurting right now. We think it's going to be a forcing function for more cloud adoption. Because if you can't buy networking gear, you're going to go to the cloud. Yeah, so super cloud to me in 10 years it will be bigger, better in the likes of HashiCorp. Actually, I think we need likes of HashiCorp on the infrastructure as a service side. I think they will be part of the super cloud. They are kind of sitting on the side right now. Kind of a good render loss in transition kind of thing. It's like Kubernetes, we'll just close out here. We'll make a statement. Is Kubernetes a developer thing or an infrastructure thing? It's an ops thing. I mean, you know, people are coming out and saying Kubernetes is not a developer issue. It's ops thing. It's an ops thing. It's orchestration, it's under the hood. So you're in this infrastructure as a service integrating this super pass layers, Dave Vellante and Wikibon call it. Yeah, it's ops thing which actually, which enables developers to get that as a service. You know, like you can deploy your software in sort of different format containers and then you don't care like what VMs are those. And serverless is sort of rising as well. It was hot for a while. Now it's like a little state, but I think serverless will be better in next three to five years. Well, certainly the hyperscale is like AWS and Azure and others have had great capex and investments. They need to stay ahead. In your opinion, final question, how do they stay ahead? Cause AWS is not going to stand still, nor will Azure. They're peddling as fast as they can. Google's trying to figure out where they fit in. Are they going to be a real cloud or a software stack? Same with Oracle. You know, to me it's really, you got the big race now with AWS, Azure, Azure's nipping at their heels. Hyperscale, what do they need to do to differentiate going forward? I think they are in a limbo. They, on one side, they don't want to compete with their customers who are sitting on top of them, likes of Snowflake and others, right? And VMware as well. But at the same time, they have to, you know, keep expanding and keep innovating. And they are kind of, they are deciding, they're debating within themselves, like, should we compete with these guys? Should we launch similar sort of features and functionality, or should we keep it open? And what I have heard, as of now, that internally at AWS, especially, they are thinking about keeping it open and letting people sort of build. And you see them buying some, you know, the Cerner with Oracle that bought Cerner, Amazon bought a healthcare company. I think the likes of MongoDB, Snowflake Databricks are perfect examples of what we'll see, I think, on the AWS side. Azure, I'm not so sure. They like to have a little bit more control at the top of the stack with the SaaS, but I think Databricks has been so successful with open source. Snowflake a little bit more proprietary and closed than Databricks. They're doing well on top of it. And MongoDB has got great success. All of these things compete with a higher level services. So that advantage of those companies not having the CapEx investment and then going multiple clouds on other ecosystems, that's a path for customers. Stay one, go fast, get traction, then go. That's huge. Actually, the last sort of comment I wanna make is that, Dave also, and you also, you guys include this in the definition, SuperCloud, the likes of Capital One and Sorner sort of vendors, right? So they are verticals. They're Capital One is in this financial vertical and then Sorner, which Oracle Bar, they are in this healthcare vertical. And remember in the beginning of the cloud and when the cloud was just getting born, we used to say that we will have the community clouds, right, which will be serving different verticals. Specialty clouds. Specialty clouds, community clouds. And that is actually, that is happening now at very sort of small level, but I think it will start happening at a bigger level. You know, the Goldman Sachs and others are trying to build these services on the financial front, you know, risk management and whatnot. I think that will be built in parallel. What's interesting, which you're bringing up a great discussion, we were having discussions around these vertical clouds, like Goldman Sachs, Capital One, Liberty Mutual, they're going all in on one native cloud, then going to multiple clouds after. But then there's also the specialty clouds around, specialty clouds around functionality, app identity, data security. So you have multiple 3D dimensional clouds here. You can have a specialty cloud just on identity. Yeah. I mean, identity on Amazon is different than Azure, huge issue. Yeah, I think we have to, at some point, we have to distinguish these things which are being built on top of these infrastructures service in pass, which is platform service, which is very close to infrastructure service, like the lines are blurred. From, we have to distinguish these two things from these super clouds. Actually, what we are calling super cloud, maybe it will be a better term, better name, but we can't, we are all industry pundits, actually, including myself and you or everybody else. Like, we tend to mix these things up. I think we have to sort of separate these things a little bit to make sense of that. I think that's what this super past thing is about because you think about the next generation SaaS has to be solved by innovations of the infrastructure services, to your point, about Hashi Corp and others. So it's not as clear as infrastructure, platform, SaaS. There's going to be a lot of interplay between these levels of services. Yeah, we are in this flux kind of situation. A lot of developers are lost, a lot of operators are lost in this transition. And it's just like our economies right now, like I was reading CNBC today, there's not having a headline that people are having a hard time understanding what state the economy is in. So same is true with our technology economy. Like, we don't know what state we are in. It's kind of, it's in the transition phase right now. Well, we're definitely in a bad economy relative to the consumer market. We've, I've said on theCUBE publicly, Dave as well, I'm not as aggressive. I think the tech is still in a boom. I don't think there's a tech bubble at all. That's bursting, I think the digital transformation from post COVID is going to continue. And this is the first recession downturn where the hyperscalers have been in market delivering the economic value almost like they're pumping on all cylinders and going to the next level. Go back to 2008, Amazon Web Services, where were they? They were just emerging out. So the cloud economic impact has not been factored into the global GDP relationship. I think all the firms that are looking at GDP growth and tech spend as a correlation are completely missing the boat on the fact that, you know, cloud economics and digital transformation is a big part of the new economics. So, you know, refactoring business models. This has come continuing and it's just early days. Yeah, I have said that many times that that cloud works good in the bad economy and cloud works great in the good economy, right? You know why? Because there are different type of workloads in the good economy, a lot of experimentation in a way that solutions go into the cloud you can do experimentation that you have extra money now. But in the bad economy, you don't want to spend the capex because you have money. Money is expensive at that point. And then you want to leverage, you want to keep working and you don't need an option. I think inflation is a big factor too right now. Well, Sharpee, great to see you. Thanks for coming into our studio for our stage performance for SuperCloud 22. This is a pilot episode that we're going to get a consortium of experts, Clouderaati, like yourselves in the conversation to discuss what the architecture is, what is a taxonomy, what are the key building blocks and what things need to be in place for SuperCloud capability because it's clear that without standards, without de facto standards, we're at this tipping point where if it all comes together, not all one company can do everything. Customers want choice, but they also want to go fast too. So DevOps is working, it's going to the next level. We see this as SuperCloud. So thank you so much for your participation. Thanks for having me and I'm looking for to listen to the other sessions. We're going to take it on the internet. I'm John Furrier, stay tuned for more SuperCloud 22 coverage here at the Palo Alto studios in one minute.