The Power and Novelty of Cloud Native

New innovations can quickly become IT sweethearts, but what’s best for the business remains the overriding factor for deployment. Data and storage analyst Steve McDowell explains the desire for and challenges of cloud native technologies.

By Ken Kaplan

By Ken Kaplan July 13, 2021

The emergence of cloud native was born from developers needing to flexibly deploy resources and reconfigure them as needed.  

“That's why we invented containers,” said Steve McDowell, senior analyst of data and storage at Moor Insights and Strategy. “They're very ephemeral, very temporary in nature. They were designed for that. Now we're in the process of adapting those to more stateful and enterprise kind of workloads.”

Cloud native is made up of exciting but immature technologies, he said. 

“These are some powerful toys we have now,” said McDowell. 

Still, there is plenty of COBOL code in the field. 

“If it's not broke why risk breaking it by moving it to something that may benefit the IT guy, because at the end of the day it's not about the IT guy,” said McDowell. “It's about the business.” 

RELATED

Clearing the Fog Around Cloud Native

In this 2nd part of an interview with McDowell, he offers insights on containers, hybrid multi-cloud, nascent cloud native technologies and the rise of a new generation of IT engineers. 

Tech Barometer Cloud Coverage is a series exploring the people and innovations behind cloud native technologies. 

Transcript 

Steve McDowell: Data has gravity. Jason Lopez:

Steve McDowell is senior analyst at Moore Insights and Strategy. This is tech barometer, cloud coverage. I'm Jason Lopez. When data doesn't live close to the workloads that consume it, it starts running into issues. Let's say you're editing a podcast. You need your source audio and video to be in sync. And that's kind of an illustration of what it's like with workloads, especially when they're in different geographies, that latency could be a problem. It could hurt responsiveness and performance. 

Steve McDowell:  To respond to that we see the cloud guys stepping up. Amazon introduced something called local zones. Exactly because of this problem. And it went to LA first for a reason because that's where the media and entertainment industry is. And they're probably the ones that are most sensitive to latency like this. But you also see this with applications, right? If I'm running a big retail database, I'm doing price lookups from a barcode scanner, right. That round trip to that database, right. It needs to be performing. And if I got a little overzealous three years ago and said, I'm going to move everything to Amazon's cloud, right. Maybe I'm experiencing some issues that are rebounding. And we see a lot of that. We see a whole lot of that. And that's one of the drivers behind this whole explosion of as a service on prem, right? So you got to HP GreenLake, Dell on-prem cloud stuff. I can get the cloud economic model and flexibility without, you know, all of the downsides of having my workloads on my data somewhere else. And Amazon is coming in without post CC, all of this activity around trying to bridge that gap. To circle back to your question, right? Cloud native gives me the flexibility to move those applications around, right? Because it encapsulates my workload in a way where I don't have to go manage a server and install all the dependencies and all of this stuff to move it. Right. So now I have this flexibility to put it where it makes the most sense. 

Jason Lopez: And you said you had another answer to that question too. There was a second one. 

Steve McDowell: Uh, the other answer is, you know, we're seeing containers used as application delivery mechanisms and you know, an example of that, right. I run a small database here in my house. I don't want to go manage a Postgres database. Right. But my NAS box has the ability to pull down a Postgres container. I'm not doing anything. I just push a button and it's here. Right. So it becomes an application delivery mechanism as well. So we're kind of seeing bnoth things.

Jason Lopez: Well, I wonder if you could talk to that idea of being able to decide which on the fly, being able to have that optionality, you know, where you have to have your application and where you have to have the data.

Steve McDowell: Yeah. And again, I mean, a lot of this container stuff grew out of the dev ops world where it was exactly that, right. You know, I'm an application developer, I'm running a coding shop. You know, when I check in code into the source code control system, there's a trigger and it says, go start a container and run all the test suites against it and then tear that container down. That's why we invented containers. Right? They're very ephemeral, very temporary in nature. They were designed for that. Now we're in the process of adapting those to more stateful and, uh, enterprise kind of workloads. But you're exactly right. And if you look at things like Kubernetes and there's a lot in there around rule-based management, IT guys want this as well. It may not be I'm going to move a workload from Amazon's cloud to my on-prem, but it may be that, you know, the server's getting a little overloaded. I'm watching all these metrics, let's trigger some rules, right? I mean, we've been marching toward the autonomous data center for a long time and containers, man, what a great enabler to move that to the next level.

Jason Lopez: So what about the idea of the high-tech aspect of some of the stuff in cloud native, you know, technology for the sake of technology?

Steve McDowell: It's funny, you asked that because you know, for all the benefits we've talked about, it's still a relatively immature set of technologies. We talked recently to a CIO at a casino in Las Vegas, and he said they actually keep a metric internally, the turnover rate of those under 30 versus those over 30, because the young IT guys come in and everybody wants to play with the toys. I mean, this is not a new phenomenon, but these are some powerful toys we have now. I'm not going to change the architecture of my casino because another architecture might be more efficient. Right? Oftentimes there's a reason that we have so much COBOL code still working, right. If it's not broke, why risk breaking it by moving it to something that may benefit the it guy? Because at the end of the day, it's not about the IT guy. It's about the business. The it guy is selling data services back to his business. So yeah, the CIO wants to be more efficient, but Joe business owner, man, you better not lose the transaction or give him downtime. But no, it is very generational. It's very generational. I talked to another, I'm actually writing a paper right now for one of the big OEMs and their take on this is your business is too critical to chase after the cool toys. Right. They're trying to sell their legacy technology, you know, and in some sense, they're right. It's going to frustrate the younger guys, but at the same time, right, those are the guys that can go to smaller companies and prove out this technology. And then it will become enterprise. So early days for this cloud native again, in five years, it's going to be baked into everything and that's going to be the younger generation pushing us to it. For sure, for sure.

Jason Lopez: Well, it seems like they're, you know, they're building their careers and in some ways they want to chase this new hot thing. They need to know it, but it sounds like they also need to be grounded in, you know, some of the fundamentals.

Steve McDowell: Yeah. 20, 25 years ago, when I started my career, I came in with all sorts of hot ideas. I grew up in the era of the mini computer, but I was a Unix guy. And I came in I'm like, Unix is the way today. It's all used, but that took a long time. And I think the same thing's going to happen. These are the toys they have to play with. These are the ones they're learning on. And as they have opportunity to use them, they will. And they'll prove it. One of the challenges with cloud that IT folks are facing is not a technology challenge. It's a cultural one. We make it so easy as an industry now to spin up an instance, you have all of this, shadow IT happening within organizations. It's nothing to take my corporate card and create an Amazon account. Actually, if I just go create an Amazon account, I get a fair amount of resources for free for a year. And people are putting corporate stuff cause they can spin up a virtual machine because it is to restricted ever taking too much time. You see the young guys doing that, you see business people that are tech savvy, join that. And that's really, I think, pushing IT very, very hard to go look at some of these technologies, right?

Jason Lopez: Well, on another tack, what's up with the hybrid cloud. We hear that the hybrid cloud, isn't really a hybrid cloud. It's not really here. It's a multi-cloud I'm wonder if you could help us sort this out, Steve. 

Steve McDowell: Yeah. And I think this, this, this whole area is deeply intertwined with the cloud native technologies. We've been talking about. We live in a hybrid multicloud world. There's no question. I don't know that we're seeing a lot of it shops deploying architectures that are spanning cloud vendors necessarily. I mean, some of that happens, but that's not a model that we see when we talk about hybrid multicloud. What we're really talking about is a range of experiences, kind of a scalability of experience from on-prem through maybe on-prem as a service or a managed service, like an HPE GreenLake or an Amazon outpost all the way up through the public cloud. Cloud native gives us the tools we need to migrate workloads between those fairly seamlessly. I don't know that I see a lot of inter cloud from like Amazon to Azure, for example, but I do see a lot and we are seeing increasing levels of, uh, you know, it's a continuum from on-prem through private cloud, you know, define private cloud is it could be like an IBM private cloud or an Oracle private cloud or it could be HP GreenLake managed experience, which looks a lot like cloud then all the way up to the public cloud providers.

Jason Lopez: So if you have a private cloud, let's say, you know, on-prem why would you be using Amazon, Azure, Google?

Steve McDowell: I don't know that I am using all three of those, but each of those are going to have a different benefit to me, right? If I am using multiple clouds, it's going to be for a specific reason, or it may be Google has some geographic coverage that Amazon doesn't. And I want to deploy an instance in this part of Malaysia, but by and large, I'm using public cloud for again, the elasticity, right? The flexibility of experience. If I have a workload that you know, is going to change over size, change over time, or I have a cluster of workloads that are know maybe short term, or I don't want to spend the CapEx required to go build out a rack of servers and I can go lease them from Amazon, right. pay my, you know, $9 a month for my UC two instance. It gives me that flexibility, the things I'm running on-prem are going to be the things that, again, either for regulatory reasons, I don't want to put in the public cloud or it's existing legacy stuff. I don't want to touch, or it needs the kind of performance that I'm only going to get on prem, or it's tied to a data warehouse that is just too expensive to move off prem because you know, your biggest cloud costs are not compute. It's done moving data back and forth, Amazon making all their money on egress charges. You know, the why is it's about a flexible software driven architecture that cloud enables, which I'm going to get from an Amazon or an Azure in blending that with kind of the performance I'm going to get with on-prem and I'll put private cloud with the on-prem stuff. Right? So I'm going to talk about, on-prem often I'm talking about, you know, your traditional kind of server VM architecture. I want to talk about private cloud. I start to get the cloud experience kind of push button infrastructure that the public guys give me, but deployed with on-prem services. And then from an economic perspective, it's really where the asset service guys come into play. I get the cloud economics without the CapEx, right? It looks like a data center.

Jason Lopez: Well, in all of this, we see a lot of competition between, you know, especially the top service providers, you know, the cloud service providers, how has that competition between them benefiting all businesses that want to use those services? What do people have to be careful of?

Steve McDowell: So public cloud has removed IT and compute as a barrier to entry to anything. There was a time when you were a startup, you're raising $20 million because you need to buy $10 million with the servers. Those days are gone, I'm getting, it's not quite rent to own, but I'm just paying for what I need and when I need it. So, you know, I can fire an experiment up that requires a, you know, $5,000 server, but I'm only spending $89 because I'm going to run it for a month. So it's removed IT as an obstacle for deployment and that scales up to enterprise IT, right? If I want to go do some machine learning, right, I can go spend $50,000 to build a big, giant GPU equipped server, or I can go to Amazon and rent one for the weekend, right. And run my model. It's all about push button deployment, flexibility of service. Now the gotchas is how do I manage that with the rest of all of my IT resources? And it all comes back to the management plane. And I want on my screen, my software defined data center, and I want to be able to push a button and control whatever. And the more kind of infrastructures, if you will, that I have to support under that umbrella, the harder that becomes. Right, right now, I've got my ma my AWS console. I've got my on-prem console. I got my Nutanix private cloud console. What I really want is Steve's IT shop console. That's the challenge and that's the struggle. And that's where there's a couple of players that are trying to solve that in a big way, but it's not quite there.

Jason Lopez: Steve McDowell is senior analyst at more insights and strategy covering data and storage. This is Tech Barometer Cloud Coverage, I'm Jason Lopez. Tech Barometer is produced by the forecast and you can find us at theforecastbynutanix.com.

Jason Lopez is executive producer of Tech Barometer, the podcast outlet for The Forecast. He’s the founder of Connected Social Media. Previously, he was executive producer at PodTech and a reporter at NPR.

© 2021 Nutanix, Inc. All rights reserved.  For additional legal information, please go here.

Related Articles

5-desktop-as-a-service-daas-common-mistakes-and-how-to-avoid-them
Technology

5 Frequent DaaS Mistakes to Avoid

Desktop as a Service is an increasingly vital component of the digital workforce. End user computing senior technologist Ruben Spruijt warns against these common pitfalls.