Ep#95 How to Gain visibility when running Multi-Cloud

October 6, 2022

Episode Summary

Today we're talking about multi-cloud, single-cloud hyper. I don't know what type of cloud we're talking about just yet. Actually, we're talking multi-cloud, but there's so much that we have to get into. What's the first thing you would recommend to me before going or even thinking about it?
Guest: Chris
I mean, my first, uh, recommendation would be don't do, don't do it. It's, it's a little bit counterintuitive, like coming from my background, but, you know, my, my first, uh, recommendation would be don't do it. Like, make sure that it's running somewhere, begin like one step at a time, deploy it, make sure that it's, uh, running somewhere and it's running fine. Understand, what it's required in order to run. Uh, if you are interested in a high availability look, look into a multi-region setup before you span the boundaries of a cloud. And, uh, you know, when you are, when you are there, I think, I guess that you will understand if you really need a multi-cloud or not. So ma the maturity of the application and the maturity of the team deploying, uh, developing and maintaining the application is really important. You shouldn't go like from square zero to like square 100. You should take one step, uh, learn along the way and see if it makes sense to go all the way up there. So yeah, don't do it. That's the bottom line.

Sponsored by our friends at Veeam Software! Make sure to Click here and get the latest and greatest data protection platform for everything from containers to your cloud!

OLYMPUS DIGITAL CAMERA

About the Guest

Chris Psaltis

I'm an engineer who decided to become a researcher. I soon moved to software development and today I'm a happy manager and entrepreneur.

The common ground in all of the above is my drive to find out how things work and improve them.

#jonmyerpodcast #jonmyer #myermedia #podcast #podcasting

Are you looking to sponsor AWS re:invent from "Beyond the Expo"? Sponsorship opportunities are still available here!

If you're interested in learning more check out the AWS WAF Website.
Are you looking to attend an AWS Summit or maybe AWS re:invent, more information here!

Episode Show Notes & Transcript

Host: Jon

Please join me in welcoming the co-founder and CEO of Mist.io. Chris Psaltis to the show. Chris, thanks so much for joining me.

Guest: Chris

Thank you for hosting me

Host: Jon

All. Chris. Today we're talking about multi-cloud, single-cloud hyper. I don't know what type of cloud we're talking about just yet. Actually, we're talking multi-cloud, but there's so much that we have to get into. Chris, how about a little bit of backstory on yourself and Mist.io?

Guest: Chris

Yeah, so, uh, as you said, I'm the co-founder of Mist.io. Uh, at Mist.io, we are building an open-source multi-cloud management platform. And by that, I mean that, uh, our end users can come in, uh, connect everything, uh, to Mist.io and then manage their entire infrastructure from one control panel, the Mist.io panel. So, uh, the company, uh, was, uh, founded in, uh, 2013. We, uh, started working, uh, on the product a little bit earlier, though. Uh, we were, uh, trying to scratch our own needs. We were running a consulting agency back in the day. Uh, we were building and maintaining, uh, open source based systems for customers around the world. We were a small team and, you know, everybody was somewhere else, uh, on the cloud, on-prem in colo, you know. Uh, so we started building this tool in order to make our lives easier. And then at some point, we said, Look, if we need that, maybe there are others, uh, out there, uh, who need it as well. And that's how it all started.

Host: Jon

When you created Mist.io or the capabilities doing it and then actually bringing it out as a product. Before that, what were you doing?

Guest: Chris

So, uh, before that, we, uh, were running a consulting agency. We were, uh, building, uh, and maintaining systems for customers around the world. So they were practically all over the place on a public cloud, uh, Colo, uh, p name it. So we started building, uh, this, uh, tool in order to make our lives easier, first of all, you know, uh, make it easier to manage all this ous infrastructure, wherever it was. Uh, so that's how it started. You know, after a while we said, Look, maybe there are other, uh, people out there who, uh, need a similar solution. And, uh, yeah, that's how the company started.

Host: Jon

Let's talk about multi-cloud. And I know that's kind of a, I don't know, people don't jump on that word right away. Like it's individual cloud providers are like, No, no, no, there's no other cloud that exists, But they are starting to admit that there are multiple cloud providers out there. Now, we most people only talk about the top three, aws, Azure, gcp. There's others that are available without, you know, within a global reach. But what does multi-cloud mean to you? And is it really possible?

Guest: Chris

Yeah, first of all, it is possible. Uh, people do it all the time. It's, uh, like the predominant, uh, paradigm for, uh, running infrastructure these days, even though, uh, not everybody is doing it in a conscious and, uh, well, well planned way, let's say. So, um, what we, what we mean by, uh, multicloud, although, you know, the term tends to be a little bit, uh, overloaded, uh, is a mix of heterogeneous infrastructure either, uh, by like a public cloud vendor, uh, or, you know, some sort of, uh, private infrastructure, uh, that can help, uh, users consume infrastructure in a, in a elastic way. Uh, so, uh, that could be, you know, some sort of, uh, hypervisor like, uh, VMware, or it could be like a private cloud, like open stack or some sort of, uh, Kubernetes, uh, cluster or anything like that. So any mix of all those. So, uh, multi-cloud probably can be, uh, seen as, uh, an umbrella term, uh, including, you know, all sorts of, uh, heterogeneous setups, uh, both hybrid, like public, private, but also only, uh, public, uh, or only private, or, you know, any, any other mix like, uh, private infrastructure, private platforms running in a public cloud, like, you know, uh, a vSphere class that are running on a public cloud or something like that

Host: Jon

For multi-cloud. You indicated that companies might not be implemented in the most effective or efficient way. What are some of the things that you're noticing with the multi-cloud environment?

Guest: Chris

You know, in, in most cases, uh, multi-cloud isn't like a strategic decision to begin with. It just, it just happened. And, uh, you know, by far, the number one reason for that are some sort of historical, uh, reasons. Let's say, uh, you know, we were running, uh, applications on VMware, and then we started using AWS, and we created some new applications there, but the old ones never, uh, were migrated from our on-prem stack to the public cloud. So, you know, right out of the box, we already have two platforms, and, uh, if you count any staging or any dev environments, then these tend to be, uh, multiple. Uh, so what we, what we usually see is that overall, you know, applications are very tight to the underlying infrastructure layer. And when a new generation comes out, then the old ones are not dead right away.

Guest: Chris

Uh, migrations are expensive, are complicated, are tricky, and in most cases they never happen. So, you know, you end up matching all sorts of different generations and types of infrastructure, uh, matching your application needs. And in a few years, even without taking any mergers or acquisitions into account, uh, your, your multi-cloud already, right? Um, so, uh, yeah, so that's, that's the predominant, uh, case that we, we see by far. Uh, obviously there are also, uh, cases where some companies are very conscious about that. So, you know, step one, they recognize that they are probably already running in a multi-cloud, uh, setup, and they try to do something about it. And, um, in, uh, in many cases, the correct way is not put everything on aws, wherever it is, like movie that there in many cases is about tidying up, uh, what you currently do, um, making more organized and, uh, more efficient. So, you know, embracing the complexity. Let's say

Host: Jon

The example you gave with VMware to AWS or to another cloud provider, is that private cloud going to public cloud? So that's multi-cloud, or is that hybrid cloud?

Guest: Chris

Yeah, you name it, <laugh>, whatever, whatever.

Host: Jon

I think the terms or the sy synonymous the same thing, because those who have a data center, right? They're like, This is my private cloud environment. No, that's your data center, but however you wanna say it. Now they have a multi-cloud using private and public,

Guest: Chris

And, you know, you, you didn't take into account ads yet. So it can get really, uh, really complicated really quickly. So we tend to prefer the multi-cloud term as like an umbrella term. Okay? Obviously like hybrid cloud is more, uh, specific, but then, you know, also hybrid cloud cloud has become a little bit of a, an overloaded term. So, you know, for example, uh, what's outpost? Is it like a hybrid cloud? Is it a public cloud? Is it an edge cloud? What is it like, you know, Uh, so there are all sorts of offerings like that, which, which blur the boundaries between what you would normally call, you know, public cloud versus private cloud versus home pre, or versus whatever.

Host: Jon

When folks decide to go to the cloud, they have the on premise, they decide to go to cloud. It's usually done because of the services, but there's another indicator on why they want don't want to use one single cloud provider. Do you think there's a fear of vendor lock in with going with one, then with one public cloud? Now, for a quick interruption, a huge shout out to our friends at Veeam for sponsoring this episode. Veeam backup for AWS can easily protect all of your Amazon EC two, rds and VPC data. Wait a second, they can protect my VPC data too. Yep, that's right. Simplify AWS backup and recovery while ensuring security and compliance. All right, now back to our episode.

Guest: Chris

Yeah, this is, this is a very common, uh, a common issue that we see. I don't know if it's like 100% justified in all cases, to be honest. Uh, you know, in some cases, embracing a single vendor can help you move faster and, uh, get to results much in a much better way than doing it like across provider. So it kind of depends on the application and the needs of, uh, the end user. What are you trying to achieve? What's the business problem? And would multi-cloud make sense in that scenario or not? Because, you know, spanning one application across multiple clouds, that's usually a very bad idea. Uh, ous pricing and, uh, the, the latency alone can, can kill everything. Uh, so usually it's a really, really bad idea. Uh, but in, in some cases, it would make sense to, you know, spin up applications, different applications, solving different problems in the infrastructure stack that makes more sense.

Guest: Chris

You know, if it's like a, an internal service, why, why put it on the cloud? You know, I, I'm betting that you have rack sitting idle in your office as we speak, so, you know, why not put it there? Uh, zero latency, zero cost, zero everything. Uh, so it depends a lot. But yeah, obviously vendor lock is a big, is a big issue, is a big question. Looking at the exit is always a good idea when you are, uh, looking to embrace, uh, new technology. And this goes beyond, you know, the specifics of a public cloud. Uh, it, it goes all the way down to your, you know, mobile device or, uh, uh, what os are you choosing? So it's, it's a very similar situation. You have to weigh in, uh, the ba the, you have to weigh the benefits and, uh, decide what's best for, uh, for your case. Many people bring up vendor locking as a reason for doing multi-cloud, But yeah, I don't think this would be a very good idea because this kind of requires to span a single application across multiple clouds. Otherwise, you know, uh, it's, there's no real avoidance of the hope the vendor lock in, right? But I am afraid that in like 99% of the cases, this just adds complexity and overheads and costs that you shouldn't be taking. Uh, so yeah,

Host: Jon

I, I think the business reason behind using multiple vendors or multiple clouds is key, but I, there's definitely no cost benefit on spliting your application between multiple clouds, like you indicated the egress, the application, trying to communicate back and forth. If it's not installed in the right region, it's global, it's banned. But I think the other thing with like a vendor lock is it's an old mentality. They don't wanna be locked into a specific, There are so many tools out there, and I wanna touch on those in a couple minute that make it cloud agnostic, that you can go to any cloud provider, but, but the reason that you go to a single one is for the cost savings, the benefit, and the depth of services that are available to you. Now, one cloud provider might provide, uh, you know, certain tools or certain services that are more beneficial to your business needs than another one. And that's really where you should be gearing towards it. My next

Guest: Chris

Question, Yeah, exactly. That's, that's best of grid thing is, uh, a trend that we're seeing and, uh, which is very, very interesting. And I don't know if it will work out in the end, because again, it depends on the applications and type of services that you need to leverage and how these communicate. Uh, but a very, very common thing, however, that we do notice is, um, people, people gravity. Uh, we, we usually talk about data gravity or something like that, Uh, but there is also a lot of people gravity. Where do people feel more, uh, confident? Uh, are they like AWS people? Are they Google people? Are they Azure people? And if they are to deal with this infrastructure the next day, why not let them choose what's best for them? Uh, why, why not allow them to, to use the, the tool that's, uh, that's more helpful, uh, and, uh, they can be more efficient with it. So yeah, people is another big, uh, parameter here. Mo probably the biggest one, uh, to be honest, it's probably less technology and more about the people.

Host: Jon

Actually, you touched on my next question, talking about the people. What about the skills required to be multi-cloud? You cannot have one person who is very efficient and a top senior engineer for multiple clouds. It is possible, but it is not efficient because here's what happens. You're designing something for one provider in a specific way, then you have to switch gears mentally and figure out all the technology, learn everything. You're spending more time learning than actually doing what about the skills? Isn't that a deciding factor?

Guest: Chris

Yeah. Yeah. I think you said it already. It's about learning, right? Yep. Uh, even in the context of a single provider, things are changing so quickly and so often, like you need to be learning all the time. So, uh, I would prioritize software skills, let's say like capacity to learn, uh, and, uh, communication, because this is also really important, especially when you have to deal with, uh, multiple people in different roles. And, uh, you know, you have to communicate what's going on and, uh, make sure that you're making the right choice. So it's, uh, these software skills, which are usually the problem in, uh, <laugh> in such, in such cases, learning, uh, the capacity to learn and, uh, communication.

Host: Jon

Who's deciding to actually go multi-cloud? Is it a top down? Because I don't see many engineers saying, I want to do multiple clouds, unless they can really tie in the efficiency of the business application, or they want to test something out in the poc, but who's deciding?

Guest: Chris

Uh, it's both top down and, uh, uh, the other way around. Uh, so it depends on the organization to be honest. You know, with more traditional and more, uh, rigid organizations, usually it's like a top down approach. We, we sit on a table, we, uh, discuss, uh, what we, what needs to be done. We come up with some sort of plan, and then, uh, it goes, uh, to the, the right people to implement it. But with more, um, agile organizations who are smaller organizations, we also see, uh, bottom bottoms up approach where, you know, uh, teams are more free to choose whatever tool they want to work with, uh, as long as they can maintain what they're building. So, uh, we also see this happening a lot. Uh, usually it's more like in engineering focused, uh, organizations, uh, who, uh, who follow a more loosely, uh, coupled approach, let's say. So both,

Host: Jon

What's your recommendations on actually choosing a cloud agnostic tool or provider?

Guest: Chris

I wish there was a silver bullet <laugh>. Yeah. Um, and like we at Mist.io we're trying, uh, to solve part of the problem because solving the entire problem, Yeah, I don't, I don't think that it can be done. Uh, it's, uh, and you know, anyone who says who can, that can solve it is probably lying or is, is very, very, uh, uh, ignorant of what lies <laugh>. So there's no silver bullet, unfortunately. Uh, the, the problem is so complicated. Uh, the, the needs vary a lot between, uh, users. So it's practically impossible to find a general solution for everything. Like, you know, ETT is, uh, rose to, to prominence promising something like an abstraction layer on top of the infrastructure. But you know, what happens when the next ETT is, uh, comes up? Uh, yeah, we've, I, I guess you've seen that happening, uh, in your career already.

Guest: Chris

If you remember, like the open stock days, everybody was crazy about open stock and the open stock will be the future, and everything will be running on open stock and things like that. And then, you know, containers happened and Kubernetes happened, and I'm guessing that the next thing will happen in the next few years. So what, what are you doing in these situations? Like, um, technology in the infrastructure space is mostly so fast, Meaning you are, you are bound to have this heterogeneity for, uh, forever, uh, even more so now that ads is coming in the mix. So solving the problem in a general way is practically impossible. What you can do though, is that you can combine services, tools, uh, from here and there, depending on your needs, and try to stitch together a platform that makes sense for, for you. And, uh, that's why we've been, uh, open source since day one.

Guest: Chris

I think like open source solutions are very critical into that, uh, solutions which you can adopt, modify, edit, make sure that you know what's happening in the background. Everything is transparent. Uh, no black boxes. So yeah, this Lego blocks, let's say, that can create, uh, a platform, uh, out of nothing. And, uh, I would also advise, like finally, I would also advise against doing it yourself. It's specifically because it's such a complicated and big problem, engineers tend to, you know, say, why, why get something when I can build it from scratch? I think like this type of mentality could work in some cases, but not in this one. And we've seen, uh, several large organizations fa failing in that in such approaches, exactly, because they thought they can build everything, uh, from scratch themselves. Uh, you know, it looks good on paper, but it's a lot of work.

Guest: Chris

It's complicated work, and it's not what drives your, uh, business forward. I mean, unless you are an infrastructure company that is. So I would just suggest just integrating different pieces here and there, you know, some sort of cloud management platform like the one we are building with, uh, some sort of, uh, you know, orchestration tools like Terraform or configuration management, danceable or, you know, sa like AOPs or whatever, you know, uh, just pull things together and, uh, try to build something that's great for, uh, your case, uh, with us little effort and us little integration and maintenance work, uh, as possible.

Host: Jon

I think building it yourself is a great idea, but in the long run, it's not cost effective. It's difficult to manage and maintain. An example is a lot of the stuff that you build yourself or your company, right? I mean, it, there's nothing wrong with Bill using open source that builds some things for yourself, but you have to look, take in the accountability that you have, not only tribal knowledge to that application. So now that person, if they leave, it gets a little different, where if you do get something off the shelf, but tailor it to an open source, uh, the type of integration or plugin, those are some of the benefits of it, really kind of going through it. But the cost around it of a multi-cloud environment, we touched on it in the beginning. I mean, how do you manage the cost of two environments and you're not getting the cost savings on it, whether it's public cloud or it's public and private cloud.

Guest: Chris

You tell me <laugh>, I'm trying to do it, and it's, uh, you know, it's, it's really, really hard and it's much harder than it seems on the surface, um, because it's not just about pricing. You know, pricing is already hard, trying to figure out how, you know, I don't think any

Host: Jon

Cloud provider makes pricing as simple as it should be, because they started out small and it was easy to project and give you some pricing. And now when you have hundreds of services and you have to predict, and you've got developers deploying out infrastructure every single where using services and multiple accounts, it just pricing gets to be hard. Now, try doing that for multiple cloud environments.

Guest: Chris

<laugh>. Yeah, Yeah, yeah. So I think like, there are three things that you should be, uh, very careful, uh, about, like, first of all, having visibility, you know, this, everything starts from there. If you don't know what you're running, where is it? Uh, what's, uh, it's price point, then, uh, you cannot do anything. Everything begins from there. Uh, visibility is step one. And then, uh, the, uh, the second part is proactively controlling cost. Uh, you know, how do I avoid spending too much before I spend it? Uh, and there are many, you know, low hanging fruit that you can, uh, you can pursue there without lots of complexity or, uh, uh, many sophisticated like tools or anything like that. You know, for example, the, the one example I usually, uh, tell people is, if you want to reduce your, uh, cloud budget, just call your account manager.

Guest: Chris

You know, it's like, no, no brainer, uh, zero technical effort. And, uh, I'm sure that you will get something in the end of the day. Uh, so, you know, just start from there. Like, just pick up the phone or send an email or something, and that's it. Uh, so, uh, dumb solution, but it can, uh, save you a lot of money really quickly and with zero effort, the most importantly. And then, you know, there are other things that you can do like, uh, you know, proactively, uh, cleaning up your environments, uh, when they're no longer, uh, needed, you know, every week or so, or, uh, having some sort of accountability. Who's, uh, VM is this, or, uh, uh, so, you know, I don't go out to Slack asking, Hey, I found this extra arts instance. It's called Test 1, 2 3, and, uh, it's been,

Host: Jon

And it been running for 365 days. How's it going,

Guest: Chris

<laugh>? Yeah, yeah, yeah, yeah, yeah, exactly. Cus PM is this, does it do anything? Things like that. Uh, so avoid situations like that, but you can also constrain, uh, who can provision what, you know, uh, why should everybody be able to provision external, large instances and not just, you know, medium ones or small ones or, uh, something like that. So there are a lot of steps that you can take there in order to proactively control costs. And I think this can easily save you like 30 to 40% of your, uh, of your budget, at least for dev and QA workloads, which are more volatile, more dynamic. You know, production workloads are a little bit different story. They're better understood, they're well mentored, uh, people know what's going on. So it's a little bit tricky to save money from that. Uh, where we see a lot of, uh, waste is, you know, these, uh, not absolutely required services or dev environments or things like that.

Guest: Chris

And then, you know, as long as you have visibility and you have some proactive controls in place, then I would look into the more interesting, but more complicated, uh, aspect of, you know, how do I optimize, uh, how do I optimize spending by, let's say, purchasing, uh, reserved capacity or by, uh, scaling my workloads here and there, or moving VMs from there, from here to there, or whatever. You know, these more, uh, these are really, really important. But, you know, if you don't do the first two, then you're probably, uh, won't save a lot of money by optimizing something that you don't know. You know? Uh, if somebody says his AWS bill increasing 10% month of a month sounds reasonable. You know, we are increasing our business. The infrastructure cost is increasing, but is, is this, is this true or not? Or you just have, you know, more and more infrastructure sitting idle. So, you know, it's, uh, I think you should definitely begin from the visibility part, then go to the proactive step and then finally begin optimizing something that you already understand really well.

Host: Jon

Doesn't all three of these things between visibility, control and then optimizing get really difficult when you have multiple cloud providers? And then how are you doing this in a hybrid type environment if you're using multi-cloud?

Guest: Chris

So, yeah, it's, it's certainly more complicated because, uh, you know, if you have one platform, you, uh, you ring and repeat, you do the same process again and again, even if you have like a hundred AWS accounts, you do it again and again, and again and again. So it's, it's much simpler. Like the, the more platforms you have, the more, uh, complicated it will, uh, it will become. So, you know, I, that's, that's why we think it's really critical to have like some sort of centralized view of, uh, what's your inventory like, Uh, and also be able to, uh, apply such, uh, proactive controls from, from one place. Uh, and also, you know, uh, make sure that you can do these optimizations from as fewer places as possible. So, you know, try to a, as much as possible, limit the search space, uh, let's say. And, you know, there are, uh, there are tools out there like ours, uh, who can help you with that.

Host: Jon

Well, let's talk about your tool a little bit. I, I know it wasn't one of our topics for the discussion today, but how does your tool help with multi-cloud or these three controls?

Guest: Chris

Yeah. So, uh, first of all, in uh, terms of, uh, visibility, uh, the only thing you need to do, uh, is to connect your cloud account, uh, to Mist.io, uh, like some sort of API credentials. And then in real time, miso discovers the resource that you have running in your cloud account. So you know, in a few seconds, you know, at least what it is that you're running and where is it, and, uh, what's, uh, the price point for its thing. Uh, so, uh, very, very quick win in just a few seconds. Uh, so that's for, uh, visibility. Now, for, uh, for control, there are obviously, you know, you could, uh, perform all common operations that you normally do from your cloud portal, like starting stopping resizing VMs. Uh, there are automation, uh, workflows that you can apply. You know, for example, if I detect that, uh, V is Id, uh, stop it.

Guest: Chris

Or there are also additional constraints, like, you know, limiting the types of sizes or images that end users can provision, uh, setting expiration dates, Everything that gets created over our api, uh, gets an ownership tag. So without the user doing anything, without setting labels, without doing anything else, uh, you already know who created this, uh, resource and, uh, to whom it belonged. And then in the optimization part, we, uh, we allow you to, uh, like quickly detect some, uh, some issues, uh, which, uh, could be problematic. Let's say for example, you know, uh, detecting that you have, uh, provisioned a hundred core, but you are, uh, actually utilizing 10. So what do you do about that? Or, you know, you detect that, uh, VM with, uh, uh, with two courses working at 10% capacity for the last, uh, two months or so. So, you know, scale it down in order to save some money. So things like that.

Host: Jon

Some of the things that you mentioned in indicated, uh, of the three controls that you're doing, is it for all cloud environments? And when I say, oh, I'm talking like multi, a hybrid, uh, maybe edge, cuz that's actually an interesting one that we might have to talk about at some point, but does MIS io handle all those types of capabilities?

Guest: Chris

Yeah, yeah, yeah. And uh, you know, obviously there's, there are no price points for, uh, private infrastructure. Uh, you know, you can plug in your AWS and it will have the plus prices there, but, uh, you can plug in also your vSphere, but it will have, uh, no prices there. We do allow you, however, to, uh, set custom price catalogs so you can apply, uh, price points to your, uh, internal infrastructure as well. Uh, but you can also override public infrastructure, uh, costs. So, you know, let's say that you have negotiated a discount with aws. Good luck with that, by the way. So, uh, you can apply that, uh, to Misto, you have a better view of, uh, what's going on. Uh, we are integrating with more than 20 platforms out there ranging, you know, from bare metals, hypervisors, container hosts, all the way to private clouds and, uh, public clouds.

Host: Jon

Chris, do you think it's gotten a little more difficult now that originally operations used to be the ones that are deploying out infrastructure, whether it was on premise, and now when they're going the cloud, it started out with operations, but has it become difficult in the new DevOps mentality where developers are deploying out the infrastructure and managing it themselves to handle this in multi-cloud and handle the cost and visibility?

Guest: Chris

Yeah, it's certainly, it's certainly a much bigger issue. Uh, it's certainly much bigger than it used to be. Uh, and I would add to that, that, you know, the, the footprints have have grown a lot. It's not just one server anymore, It's like ten one, one hundred, 1000. So, uh, it's certainly much harder. And the ops teams, they, I don't know how they called them nowadays, it's like it was, uh, system admins. Yep, they were later DevOps teams. Uh, I think now they're called platform engineers, and I don't know how they will be called, but it's practically like the same people. Uh, so these people, they have a lot of trouble, uh, trying to somehow control the situation and why, that's why you see this rise of DevOps tooling. Um, and it's only going to get worse, to be honest. That's, at least that's what I think. So, uh, yeah, there's a, there's a lot of, there's a lot of pain out there.

Guest: Chris

Ops teams are even, even in an even worse position right now than they used to be because they have a lot of pressure to, you know, keep control, uh, while not performing the actual provisioning tasks themselves. So they need to be prepared, they need to have everything in place. They need to be testing and maintaining everything, so others can just come in, click a button and get what they need. So yeah, the, the, the type of the work has changed and, uh, but it hasn't become easier. Uh, it's, it's, it's much harder right now, I think mostly because of the additional complexity.

Host: Jon

Chris, before I wrap things up, I got a quick question for you. On multi-cloud, that's what we've been really kind of centering focus. I am a customer and I want to go to multi-cloud, right? I'm thinking, yeah, this is gonna be the best thing. I'm gonna use two of those. I'm gonna have high availability, my application's gonna run. You'll never see it go down. What's the first thing you would recommend to me before going or even thinking about it?

Guest: Chris

I mean, my first, uh, recommendation would be don't do, don't do it. <laugh>. It's, it's a little bit counterintuitive, like coming from my background, but, you know, my, my first, uh, recommendation would be don't do it. Like, make sure that it's running somewhere, <laugh> begin like one step at a time, deploy it, make sure that it's, uh, running somewhere and it's running fine. Understand what, what it's required in order to run. Uh, if you are interested in a high availability look, look into a multi-region setup before you span the boundaries of a cloud. And, uh, you know, when you are, when you are there, I think, I guess that you will understand if you really need a multi-cloud or not. So ma the maturity of the application and the maturity of the team deploying, uh, developing and maintaining the application is really important. You shouldn't go like from square zero to like square 100. You should take one step at eye, uh, learn along the way and see if it makes sense to go all the way up there. So yeah, don't do it. That's the, that's the bottom line.

Host: Jon

I would actually tack on that question. Who do you think's really driving more the multi-cloud or should be, or even thinking about it, startups or enterprises?

Guest: Chris

Uh, it's, it's hard to say, to be honest. Um, I think for some time it was, uh, mostly because of startups. Uh, it wasn't that enterprises weren't actually using some sort of multi-cloud setup, but I don't think that they, uh, they realized it very early. Uh, now they're realizing it, and I think now the enterprises are the ones who are, uh, pushing this forward. So it's a real PR problem that's really hard to solve. Uh, it's very critical for, uh, how well you run your infrastructure and how good business, uh, outcomes you deliver. So I beg, I believe that from, at least here and on, enterprises will be the, the main driver and obviously, you know, vendors will follow. And that's why you see, you know, many public cloud vendors lately speaking more about multi-cloud, when it used to be, uh, a word that was, uh, you know, off limits, let's say

Host: Jon

You weren't allowed to say it or a minute. So Chris, is there anything you'd like to leave the audience before we wrap it up?

Guest: Chris

No, no. I think, uh, that was, uh, that was really, uh, really great. Like, uh, had a lot of fun, uh, doing this and, uh, I hope it's going to be helpful for, uh, for your audience.

Host: Jon

Uh, definitely. It's great to hear from thought leadership around multi-cloud. I know it's been one of those words that people didn't emit, but now is prominent and it's helpful to understand what type of cloud environment they want to implement, whether it's an on premise, hybrid, multi and upcoming. And I think, Chris, you might wanna coin the term edge cloud, unless it hasn't been, or already because it's very interesting. Chris, thank you so much for joining the show.

Guest: Chris

Thank you. Thank you very much.

Host: Jon

All right, everybody, co-founder and CEO of Mist.io, Chris Psaltis, I'm your host, Jon Myer. Don't forget to hit that, like, subscribe, end notified, because guess what, folks, we're outta here.

 

Comments are closed.