EP. 23 Can Multicloud Optimization Revolutionize Manufacturing? Joe Johnson

00:00 - Introduction to the podcast and guest Joe Johnson

Matt Pacheco

Hello, everyone, and welcome to the Cloud Currents podcast, where we explore the innovative technologies and strategic approaches driving the future of cloud computing.

I'm your host, Matt Pacheco, and I oversee the content team at TierPoint, a managed cloud and data center services provider. Today, we're joined by Joe Johnson, the VP of Cloud Operations at Revalize. With two decades of leadership experience across tech startups, Joe is a true cloud pioneer. He'll share insights into optimizing cloud strategies through infrastructure automation, leveraging AI and emerging tech, and cultivating cloud talent.

At Revalize, Joe oversees the company's mission of transforming software firms into industry leaders through technical innovation. He previously architected Call One's transition to a cutting-edge cloud communications provider by adopting cloud native practices. And Joe has a wealth of expertise spanning DevOps, cost optimization, multi -cloud deployments, and then much, much more that we'll get into in this episode.

We're thrilled to have him on this episode to discuss architecting for the cloud challenges manufacturers face in the cloud era and the future of this dynamic industry. Thanks for joining us today, Joe.

Joe Johnson

Absolutely. And with that introduction, can I hire you to just walk into rooms before me from now on? Repeat that for me.

Matt Pacheco

I would love to. That's awesome. No, it's, it's, it's excellent. Your background is really interesting. I'm really excited to talk with you about some of these interesting topics that our listeners will get a lot out of. So, really appreciate you coming. so let's, let's start this thing off. Can you walk us through your career journey and what first sparked your interest in cloud and technology?

Joe Johnson

Excellent saying.

Joe Johnson

Sure, yeah. So I started off on the wrong side of the aisle in consulting, you the evil word that no one is allowed to use anymore. Working with technology consulting and management consulting, first for banks and credit unions, then working with law firms and corporate legal departments, and then just general corporate management consulting before I came across the aisle to the internal side and came on board with an MSP based in Chicago that was focused on the SMB market for colocation and managed services. To me, I remember way back in the day when was, know, virtualization was first born and, you know, the thought of, my god, I gotta put, you know, 16 gigs of RAM in a server just to run two servers. That's more than two servers. Why would anyone ever do that?

And going from that journey to now, you know, the scale of the hyper scalers and the amount of stuff that they can jam into to you of rack space in the data center is just incredible. Coming into that MSP, we had a lot of VMware. had a lot of hypervisor based. We started the journey of becoming a hyper scaler, building something out of our own on KVM using the on -app platform was real interesting. I was not involved in the on -app project, but watching it from the front lines really sort of gave me my first introduction to what was possible with these platforms, what we could do by embracing that model of hyperscale inside of data centers and pulling out as much as we can from the hardware itself. From there, moved on to call one where we were a regional, very traditional copper-based telephony provider in the Midwest market focused in Illinois, Ohio, Indiana, Wisconsin, that sort of region that was just sort of championing it a bit to become a VoIP and UCAS and truly digital provider and had the customer base, had the customer chops to build that experience with a really talented team And we rolled out a pretty impressive UCAS platform based on the MetaSwitch architecture that we first built in our own hardware in a colocation center and then really go full force into the cloud and adopting a VMware cloud for the platform, outsourcing all of our hardware into another service provider and letting them manage sort of the day -to -day tasks and headaches that come with managing your own infrastructure. So we could focus on what we do best, which was providing VoIP and voice experiences. We sold that company to a larger one, which then sold to a larger one based in Europe. So that was an interesting time and had an opportunity to come on down.

The thought of spending a second winter with COVID lockdowns in Chicago was like, no, let's throw a dart at a map and go somewhere warm and ended up on Florida. So Revalize was just starting out. I actually didn't even interview with Revalize because the company name was still secret at the time. So I came on board to AutoQuotes, which was the first company that they bought. And my first day was shortly after the Revalize name was released. So it came on to help build out and combine the really strong leadership and market dominance that we had in a number of verticals on CPQ and configuration products for manufacturers and bringing them to the next level in the cloud, which has been sort of the last three year journey.

10:20 - The evolution of cloud adoption from on-premises solutions

Matt Pacheco

Really interesting. And we're going to get into like some of the experience in technologies that you're using, that you've used throughout your career and what you're using currently are Revalize. I have a quick question about your, your experience overall. So you've worked with both small businesses and then enlarge businesses. how have your perspectives evolved, specifically like balancing innovation with pragmatism, in those roles.

Joe Johnson

It's really interesting the way that the industry seems to pull 180s constantly back and forth over time. Because if you asked me 15 years ago, I had lot of customers we had. It's very interesting. It's 27 cousins all from the country of Jordan who came to Chicago. They all owned beer, wine, and liquor stores. And we did their accounting software and their point-of-sale system.

If you had thought of putting something like that in the cloud or a data center or anything else, you know, 15 years ago, you'd be like, that's crazy. It's so cheap to put a server under the counter at the store and run everything there. Why would anyone do that? And now I got to have internet and I have to have backup internet and I have to have all this other stuff that I need to get to the cloud. And it's gone the complete 180 where it's so cheap and easy to run everything in the cloud now to run that infrastructure.

And internet is so prevalent and cellular backup with 5G is so easy that it's gone the opposite. It's cheaper for them to maintain a virtual server that manages everything and they connect to it remotely than it is to run everything on -prem. And hell, most of the products that we sold them are now replaced with cloud -based products and SaaS -based products. So you're not selling Microsoft point of sale anymore. You're competing against all those virtual SaaS -based point of sale systems instead.

Or you're looking at like QuickBooks online with QuickBooks point of sale that just integrates directly with each other. So it's completely altered it. And in the medium sized business, when I started, it was, you know, green terminals with typing stuff into spreadsheets with serial mucks. And then it turned into the PC and then it turned back into VDI. And now it's coming back to PCs, but using cloud services. like it just whiplash back and forth over and over again as things adjust come and go and change so drastically over the years.

12:51 - Challenges manufacturers face in migrating to the cloud

Matt Pacheco

Interesting. You had mentioned the evolution of moving from on -prem to the cloud and how easy it is. Are you still seeing resistance with some of the companies you work with to do that?

Joe Johnson

It can be, I get it too, because a lot of it's a matter of access and control. When you're working with manufacturers in particular, they have long -standing legacy ERP systems they've worked with for 30, 40 years. And they've built the entire suite around their access to that ERP. So they've built their own BI platform. They've built their own reporting system. Oftentimes, there's legacy applications, i .e. Microsoft Access front ends, they've built the connect directly to SQL to pull data and systems and to have the CPQ or the PLM system no longer part of that little cocoon they've built inside of their data center. It's a lot of control and access they have to give up. And modernizing the access method. So for us, our primary CPQ platform uses GraphQL to provide database access to our backend. We don't provide SQL access anymore in our hosted product.

So to convert all of that stuff over would mean throwing sort of the baby out with the bathwater and completely starting from scratch. It's scary. It's terrifying to think about all of that access and all of that way of always having worked. Cause I see it myself. You I started as an exchange and SQL guy. I ran my own servers. I did my own OSs. To give that up and just use Office 365 seems like cheating in a way to me, but.

Once you start to use it and you realize, I haven't thought about email delivery in 12 years. you just don't, it doesn't come to mind anymore because the world has changed so much and you just don't do that. The other side to that though is also the cost factor. The biggest driver that you see from a lot of businesses to go to cloud is that I don't want to have to maintain the cooling and the power and the backup internet and all of that sort of stuff.

But if you have to maintain it for five other enterprise tech systems, taking one out doesn't make a lot of financial sense to really anybody because you're still going to have to pay for the cooling and the power and the backup internet to keep the ERP running and the Exchange server and the SQL server. So oftentimes it's the discussion about, but when does the other shoe fall? You're going to have to move one of them first. Why not us?

15:25 - Common mistakes companies make when establishing cloud strategies

Matt Pacheco

And that makes a lot of sense. So when these companies are looking to move to the cloud in general, what mistakes do they typically make when first doing that or trying to establish their cloud strategy?

Joe Johnson

I think that is a very loaded question because nobody wants to be the first one to make the mistake. But the MSP that I work for, for example, has a very famous customer who very famously went all in balls to the wall cloud and then made a complete 180 and came fully back out of the cloud. And that's all of their applications, all of their servers, everything came back to an on-premises solution just because it made more sense financially for them after they made the move.

And that's the Basecamp product. So Basecamp was very famous in the cloud exfiltration community of pulling out of it and going back to an on -premise solution for all of their products. For them, it made a lot of sense because when they had that total control from the hardware layer all the way up to the application layer, they were way more efficient in the methods that they used for deployment, for management, for security, for everything. And I think that's the approach that people need to take is it is not an all or nothing proposition. It is a piecemeal, bit by bit, staged, slow progression. And if you do that, like rip the band aid, put everything in, it's not cost efficient at all.

Hosting legacy VMs in the cloud is very expensive compared to on-premise or co -location or a managed VMware from some other company that provides it. Bare Metal services are super cheap in comparison. And a lot of that is because you have to purchase cloud for full resource utilization. If need a 32 gig RAM VM, I'm buying 32 gigs. And I'm paying for 32 gigs. In the VMware world,

That 32-gig server might only use 32 for an hour a day, and then it's using two for the rest of the day. So the VMware can be oversubscribed. And that's not a concept that the cloud handles when it comes to virtual machines. When it comes to app services and other native cloud things, it's really good at that. But now you're talking a complete redesign and re -architecture of all of your applications to be cloud native. And that's not also a fast or a cheap process for anybody.

Matt Pacheco

So you got into cost being one of those challenges too. I'd like to ask you, what are some advice or best practices you recommend for effectively managing cloud costs at scale?

Joe Johnson

It's another tough one because there's a lot that can go into it. So, you know, in our particular internal use cases, we, took our eye off the ball for a bit and we had about a three-month period where cloud costs went up about 40 % because no one was paying that close of attention to it. And things were getting implemented. Servers were being deployed. No one was tracking it and keeping on top of what we were utilizing and more, you know, we bought 18 companies in a two-year period.

So there's 18 places to go monitor everything. And that's a challenge. think, you know, not many companies face it at 18, but companies face it with acquisitions and mergers that they have to deal with stuff that I didn't deploy. So I don't know it inherently off the top of my head. So I don't know if I can shut things down. So for us, it was really focusing in on first auditing every single thing we have in the cloud, and then two assigning ownership and management of those devices to specific groups so that we know as a cloud operations team, who do I go yell at when something is very expensive? Or who do I ask, hey, you still need this, hey, you still need this, hey, you still need this on a regular basis to make sure we're not oversubscribing all of the systems we have. The second thing that we really focused in on was for non -production systems, usage -based hours. So we have, for one of our major products, we have four development environments for it.

We have two teams in the US and we have two teams in India who work on that product. So what we've done is we've set up the two US environments to run during US business hours, a little bit extended because let's face it, most developers work nights. So we run essentially from 8 a to 8 p weekdays and then all Pacific Standard Time because that's where the team is based. And then the other two sets we run from 8 a to 8 p India Standard Time for the same group of people to work over there.

And we found that that covers 99% of our use cases. So when there's a big push at the end of a scrum or a big push at the end of a particular build process, we can go in and adjust and edit that schedule and give them the flexibility to turn everything back on if they need it for a little bit longer or one night a week or two nights a week or whatever it has to be. And that's really cut a significant amount of our cost across development and non -production environments.

20:33 - Overview of Revalize's hybrid and multi-cloud approach

Matt Pacheco

Great advice. So it's safe to say you have either at Revalize, if we shift the gear, your company has either hybrid or multi -cloud approach. Can you talk a little bit about your cloud approach?

Joe Johnson

Certainly, yeah. So we inherited everything. I mean, it's all products that we've acquired from companies that we've purchased. So there's a lot of opinions and a lot of thoughts into how that was deployed that are not always similar between company to company. it's been a challenge our last three years to kind of bring some standardization and common threads across all of our various products.

But what kind of stood out to us immediately was if we leave AWS and we move everything over to Azure, that's going to be the day that CorpDev comes to me and says, okay, we just bought three more companies and they're all AWS. Like I'm going to take the big godfather scene is I get out, they pull me back in immediately. Same as in reverse, Azure to AWS. For us, we're about 70 % Azure and about 30% AWS today.

So it wouldn't make a lot of sense to move 70 % of our stuff over to a different cloud provider unless there was a really compelling financial reason. But that's what it comes down to is, is the, is the cure worth the side effects of what we're going to do? One exception we had to that is we had a single product in Google cloud that the former owner of the company moved to Google cloud because they offered him a bunch of rebates early on that made it half price compared to AWS.

So obviously made financial sense then as soon as those rebates expired, we're now paying the same as AWS and we have exactly zero GCP experience in our team. And then the engineering team that does the coding was not a big fan of GCP. So that made no sense to maintain one product there, made no sense to keep everything there. So we worked with AWS and moved everything back early this year, late last year to AWS. And that was a pretty successful project to move.

We had a great amount of help from AWS. I really appreciate what their team did for us.

Matt Pacheco

So you had mentioned that there's a balance between AWS and Azure. I think you said 70 -30, something like that. How do you determine the optimal distribution of workloads across those different cloud environments?

Joe Johnson

So for us, it's mainly brand specific. Because one brand is already in one cloud provider, if we turn up a new server or a new product for that brand, we tend to keep it. It's my mess, but I keep all the piles of mess in one place so that they're all together. If it's something completely new, say we acquired a company that had an on -premise solution or a co -located solution that needed to be moved to the cloud, for us, it's really been a matter of going to both of them saying, OK, here's what we need.

Give me your best and final, tell me what you'd pay for. And that's often brought in partner dollars to pay for the actual migration itself. So my team had less work, less overhead to work on. Engineering had less work, less overhead to work on. Or it's a product or a company that we've purchased simply because it's related specifically to an existing product we already have. So we made the acquisition of KCL CAD last year, which is a visualization tech that relies heavily on the infrastructure and technology of AutoQuotes, another one of our brands. AutoQuotes was hosted in Azure, so that meant when we bought KCL, it made the most sense to put it into Azure so that there would be low latency or no data transfer costs to move data back and forth between the two systems.

Matt Pacheco

It sounds like there's potential for a level of complexity with those different cloud platforms in a hybrid environment or a multi -cloud environment. would you say, I say complexity and that could be your answer if you want, but what are some of the biggest challenges you find in managing an infrastructure like that on a global scale?

Joe Johnson

I think the biggest one is finding responsibility of who owns the solution when it is in a different region or a different cloud provider than normally we have. So what we found is we have to maintain a balance between AWS and Azure knowledgeable engineers on the cloud team. But it's much the same with our product engineering and even our just our product management teams they've had to have people who have specific experience in the tech stacks of those products. So for us, it's generally fallen into the buckets that our Java based technology has all been Azure and our PHP and Node has all been AWS with one notable exception. One Node application is written and run in Azure instead, but that's because all of its Java apps are in Azure as well. it's...

It's really more a matter of talent management and having to sometimes double bench so that we have experience in both sets. Some of that I blame on marketing because AWS has invented a marketing name for every single thing that has existed for 40 years in technology and decided to call it something different because they can. Not that Microsoft's not guilty of that too on a few counts, but at least they call it a virtual machine and not EC2. Get off my high horse for a second.

The second thing is the complexity of keeping it. We've adopted a really strong kiss mentality, keep it simple stupid whenever we can because it can get infinitely more complex than we've made it. And having a strong set of controls around what we name things, where we place things, and how we procure them has made it lot easier for us to keep our arms wrapped around all of that.

26:51 - Challenges of operating clouds across multiple regions in manufacturing

Matt Pacheco

Great answer. Thank you for the insight on that. I would like to shift gears to, we're still talking about cloud, but talking more about the manufacturing aspect and cloud challenges as they relate to manufacturing. So with many manufacturers operating globally, can you elaborate on the challenges operating clouds across multiple regions and countries have on this industry?

Joe Johnson

Sure, yeah. I think what you see is, obviously over the past 30, 40 years, a great deal of manufacturing has shifted overseas. So what we found is that we'll have a strong sales presence in EMEA or Asia Pac or North America or South America that are using our tools that we develop.

But the problem is that our tools have to integrate with the ERPs and the manufacturing systems. We have to offload the bills of material that we generate for our quotes to the actual manufacturers. Some of our stuff has been in visualization automation. So we've got a lot of tech around 3D visualization where, say, a gear motor company has the marketing gear motor design where they show you the pretty picture of what your motor is going to look like. But the actual thing that goes into manufacturing has all of the thousands of gears that go into that gear motor also defined. And our system has to generate that, or at least automate part of that generation process to make life easier for the engineers. So they're not dedicating an engineer to every single customer sale that they make all the way across the board. So what we found is that our sort of SMB level has always been a dedicated single application server application.

And our enterprise grade one has been a true multi -tenancy, multi -server load balanced, but what they haven't offered is regional availability. So you pick, I am a North American instance or I am an Asia pack instance or an EMEA instance of the application and everyone goes there. And the problem is you can't fight physics. There's just a limitation to latency across the world that you run into. So we've tried to make it so that the application itself is more reliant on the local latency of the systems that are talking to each other so that the end user, it's just the rendering of the application in their browser, especially when we deal with Chinese manufacturing behind the great wall. So I'll say this preface is my husband's from China, been to China many times. Internet within China is amazing. Like stuff is super fast. We're getting gigabit, multi gigabit on 5G service, but it's as if the entire country shares one cable modem to the rest of the world. So
Joe Johnson

There's lots of problems with speed and latency when you go from China to outside of China or vice versa. And that's been a real challenge for a lot of our customers when they're trying to get CPQ applications or God forbid, 3D visualization technology to the Chinese staff that they have. A couple of things that we're working on is in specific to the application architecture. So we have these monolithic legacy Java applications that we've purchased.

We're trying to take those and really break them down into microservices so that it's easier for us to deploy a scalable multi interface, multi region application to replace what we have today in the legacy application. And particularly in the CPQ market where there's a major impact to our manufacturing customers because of their globalization efforts. That's where we're focused in right now is to containerize and multi region those applications as fast as possible in a responsible, obviously, and user -friendly way. And it's been the biggest push we've had in our Configure One product. We've launched Configure One Cloud as a classification for that product that's really looking to make it more of a SaaS application than an enterprise application, which it traditionally was. So what it's including is stuff like, we're going to maintain the infrastructure. We're going to patch it. We're going to keep it up to date. We're going to put it behind a WAF, DDoS -protected, all the stuff that

We've always done with our hosted applications, but what we're also doing is upgrades to the software, deployment of new versions of code, implementing new features on the overall product stack are all going to be part of the hosting fee of that application now. Whereas traditionally you've bought version 11 and you're stuck with version 11 and then you buy version 12 when it comes out. And by the way, then there's the professional services that go with upgrading you from version 11 to version 12.

Now it's like the Salesforce model. You've got the 2024 first half update, the 2024 second half update. You get your release notes, you get your what's different. You have a sandbox environment to go test any changes and then you just get new code. And that's really been a big focus in order to make it easier for those people so that we can deploy those new features and make it better and not have to go through a major re -architecting or redeployment on behalf of those customers.

31:53 - Security implications of using various cloud platforms

Matt Pacheco

Yeah, that's really interesting about the multiple regions. And you talked about what I'd like to get into next, the security aspect of it as well. And you touched upon this with DDoS and WAF. But can you talk about the implications of having customers in various different regions with various different cloud platforms and then the security implications of all that?

Joe Johnson

Yeah, and that's the thing that works most to keep me up at night. I had a lot more hair before I got into information security side of things, but I think the biggest risk nowadays to any business is the information that they put out there on the internet. And let's face it, it doesn't matter if it's in the cloud or if it's on premise or if it's on a locked box inside of your office, it's just as vulnerable the minute it's put onto the internet as any other.

I worked a lot with manufacturers early in my career too, and I remember a plant in Kansas that I went to to go visit on behalf of one of my customers. And the manager literally had a lock box on the wall behind his office that had the internet cable. And he would unlock it and the foreman would unlock it and they'd open it and they'd plug the internet into all of the connected devices to run updates on it.

And when they were done running updates to all the CNC mills and everything else, they'd unplug the internet from the machine network and they'd lock the box back again. And that was how they stayed secure. The internet was only on when they needed it. That's only way you can stay 100 % secure when you need to. In the real world where the rest of everyone works, you face these challenges and a lot of it is around scale.

Small companies, manufacturers in particular, they don't have the scale necessary to provide information security to the level that it needs to be to protect their data. And what we're looking at is that as we scale larger with our hosted environments, we've got 200 something Configure One customers hosted in our cloud today. So that's 200 times the resources and infrastructure that any one of our customers could have, even our largest customers.

On FPX, we have one global manufacturer that does something like $3 billion a year of quotes through our platform. That's amazing. that's more than, you know, those 200 customers combined on the, on Configure One side, but they do it because we're able to scale to meet their needs and provide that same service in, you know, 47 countries in 14 languages all across the world and keep it secure for them in a way that they..just even they don't have scale to cover.

34:48 - Emerging trends in cloud technologies and AI integration

Matt Pacheco

Yeah, security. And we know with the threat of ransomware, the threats and the opportunities to be disrupted are growing. So that's really interesting. Switching gears to cloud trends and emerging cloud technologies. I'd love to ask you about what are you most excited about when it comes to new technologies like serverless?

Joe Johnson

AI machine learning, integration into the cloud, immutable infrastructure, things like that.

I think the biggest change and the problem is that enterprise customers are seven steps behind the leading edge. And that means that a lot of companies like us have to be six steps behind because we've got to be one step out of the customers. But it also means that a lot of this stuff that comes out is as flashy and interesting, but not immediately applicable outside of the bleeding edge tech company, software company type stuff.

But that's not to say that it's not useful. As a feature in our product deploying an artificial intelligence engine inside of a CPQ, that's awesome. That'd be some of the coolest tech I've ever seen. Are we developing it today? No, it's not something in the product roadmap right now. It's something that's being considered. It's something we'll pursue next year as part of our product roadmap. But the benefit to the customer would be that we'd spend $100 million building it and they wouldn't be able to use it because they're subject to EMEA AI laws or USAI laws and they maybe don't trust it to provide the correct answers yet because it hasn't been proven and shown to work successfully. You get the Chevrolet dealership that got someone to convince the bot to give them a zero-dollar car because they didn't put enough security guidelines around the AI chat bot.

That sort of stuff is very scary to manufacturers, especially in a tool that's designed to quote and sell their products. And the worst thing from a customer standpoint of customer service is we quote you the wrong item. That's often why we exist as a market is that we're building tools for companies to quote the right product to the customer and give them the service that they need. And if that fails, like...

Why do we exist? They can use an Excel sheet to sell their product. They don't need us if the only thing we provide is the business logic and business rules that make it impossible to sell the wrong product. That's not to say though that we're not utilizing it in other areas. So we've seen a lot in software development, use of Gen .ai to do coding. It's not necessarily a matter of reducing the number of people, but increasing productivity higher.

I talked to a lot of the portfolio companies owned by our owners and we're seeing 30, 40 % productivity increases for engineers and software developers across the board. We're seeing massive amounts of security increases in environments because of the use of Gen .ai tools that go and analyze your cloud environment and find vulnerabilities and find misconfigurations in your systems using AI GIF.

AI gen AI driven tools to push that out. And then what we're starting to dip our toe into too is on the sales and business operation side. What kind of efficiencies can we drive in contract analysis? Customer sends us a contract in a different language. Can we translate it and then analyze it automatically to provide better solutions for our legal team? Salespeople, can we get follow-ups? Can it scrape their inbox and give them summaries of what's most important today?

Are there angry customers? Are there happy customers? Can they do sentiment analysis of things inside of their inbox so they can prioritize it and focus on the right thing right now versus just starting at the top and working their way down every day like most people do currently? And then we're pushing it really hard as part of the sales process through the use of some sentiment analysis tools. So when we do a demo to a customer, we use a product that records the meeting with everyone's consent and does sentiment analysis of what was most interesting and what was least interesting based on engagement with the people participating. So that we know next time we give the demo, spend more time in this area and less time here to further refine and develop the salesperson skill set and train them to be a better salesperson just using these tools. And that's been really, really cool stuff. As a nerd, that's the sort of stuff that I love to see is the creative use of Gen .ai in stuff that's...

Like, not even, doesn't make sense, not something you'd think of as part of this process.

Matt Pacheco

Yeah. I've seen, more of those, note taker, note takers join meetings and, and, and then the, output is just crazy with the, the sentiment is really cool. I know Fireflies does that. I've seen that there, a few other tools do it, but I love that sentiment tracking. really cool. so you covered a few of the interesting ways, from your sales perspective and from a, from a cybersecurity perspective of how.

Joe Johnson

Exactly.

Matt Pacheco

Companies are using AI and machine learning, when it comes to the cloud, what are some ways, and this has probably existed for a while with machine learning, what are some AI or machine learning techniques that could potentially streamline workflows and cloud workflows and automation? Do you see that in your world? Is that something you foresee becoming more prominent in the future?

Joe Johnson

I think so, and I think a lot of that comes from just the amount of data analytics that comes out of cloud use. We run an analysis tool called Vega Cloud that gives us some FinOps output from our cloud. So it does financial analysis of what we're spending our money on, where we could better spend money based on machine utilization, CPU, RAM, disk, that sort of stuff, places where we're using premium disks.

And we're only pulling enough IOPS that a standard disk would work instead. Things where, you know, it's on hot storage, but we use it once a month. So does it really need to be hot storage? Can it be cold storage? Those sorts of things. And provides analysis to that. The next generation of that I 100 % see being, you know, Gen .ai driven where it's analyzing that data, not just for us, but for thousands of customers across the cloud and taking the same lessons learned from everybody.

And providing better feedback on how to tweak that. And then as we, as a company, move more towards infrastructure as code, as we adopt Terraform and other automation technologies to do our deployments for us, rather than a human clicking next, next, next on the new VM field in Azure, we're also seeing that across Gen .AI tools that can do Terraform templates for you. So it can suggest a better way to build and deploy a new server install for a customer that maybe is more efficient and provides a better solution to that deployment process.

Matt Pacheco

Back to what you said about Terraform and infrastructure as a code. What role do those play in optimizing your cloud footprint?

Joe Johnson

For us right now, not a big role yet. What we're starting to do is use Terraform to capture current state and suck it out because we're looking at that as a very strong disaster recovery tool for us to be able to quickly recover infrastructure in different locations. Two of our product lines are 100 % container driven, database driven.

Application as code sort of thing where we have container services deploying the various microservices out there. So I could recover those applications very easily in a different region. Or if I had a customer in EMEA who wanted that product hosted in EMEA, I could turn it up like that using CloudFormation or Terraform and provide them the service that they need in a different region. So now that we're looking at it as a disaster recovery method,

If we can tie it in with the backup solutions we're using in the cloud to say, OK, this is a central US, here's the VMs, here's where the VM backups are stored and what they're named, and we replicate that to US East 2, I want you to change all this to US East 2, redeploy it, restore the VMs, and walk away and do everything, and go update all my DNS entries as a method. And Gen .ai is helping us greatly with that because you can essentially tell the tools in Visual Studio, hey, this is what it is, make it that instead, and it does most of the work for you. We're not quite to the point where you don't need to know how Terraform, like I can't do it. I don't know how Terraform works. I've written like two templates in it, but my team who know Terraform can do it and do it more efficiently because they're able to let the GEN .AI do the heavy lifting and then they can do the audit and make sure it all looks right before we actually test it.

Matt Pacheco

Thank you. Thank you for that. You had mentioned these technologies, and I figured I'd ask about it in your environment because it's really interesting. So to do all these things, whether it's what you mentioned, Terraform, Gen .ai, all these cool things in the cloud, and then also maintaining your cloud takes talent and that's, that's, that's a thing. It's a sticking point for a lot of companies out there because cloud skills are in demand, just not enough people to do it. So, so what strategies has, has Revalize employed for attracting and retaining specialized talent who could do these things.

Joe Johnson

Yes. So we have a really strong focus in what we call high potential employees. So typically these are people who are earlier in their career but have the knowledge and the skill set already, just don't have the resume to back up what they're looking for. And we focus on these individuals where we can bring them in, oftentimes a junior developer or junior engineer who's ready to be an engineer or developer and we can give them that opportunity and they come work for us. We give them that support of a senior working over them to help guide and run their work for them and get them pointing the correct direction.

But we can level them up very quickly because they have the drive and the personality to be someone who wants to push to that next level as quickly as possible. And we run that all the way up the chain. Everybody, we try to focus in on that high potential where we can, given that we have a experienced senior person to be their mentor within the organization. And that's given us a lot of stickiness in terms of talent, as well as attracting a really good pool of people who know this new technology because they're the ones that typically are passionate about it, because they're going off and learning and experimenting with it in their free time just to learn because they like to, and not because it's something they've done for 14 years and they have all this deep bench of experience in that particular time.

Matt Pacheco

Yeah, the world is constantly changing too. Like what is every two weeks there's something new, something new going on. So I really, that's really cool.

Joe Johnson

Cheers.

Matt Pacheco

We're sharing about it. We're getting into the final part of the episode. The time's flying today. This is where we have a little more fun and ask about the future and think of hypotheticals. I'll start with a nice hypothetical. If you could go back and restart your cloud journey, what would you do differently knowing what you know now?

Joe Johnson

That's a tough one because there's a lot of early mistakes, but I would say more heavily focusing into technologies that we now know are cloud scalable. So, you know, early in my career, a lot of what we did was very focused on like the C sharp monolithic fat windows 32 applications that that people use every single day. And it's great.

But going web -based, going Python or PHP or something that can be hosted online would have been a lot better. And I think I was kind of slow to some of the application stacks. at call one before, we used Python for everything. It's kind of like,

I don't use Python. Like these people, every other application they use is a desktop application. Why aren't we using C Sharp? And my team really went into, this is the power of building these solutions out this way where we can sort of reuse this microservice for multiple inputs from multiple applications and capture it all and then put a GUI in front of it that the user doesn't really do anything in the GUI. It just points at the microservice to run API calls.

OK, that makes more sense. And then as you learn more within particularly the telecommunications industry, it's all powered by APIs. Everything is an API. So all of our new customer turnups, all of our existing customer data we're pulling from carriers, all of it's coming via API from the carrier itself. And that really started to make a lot more sense. And I probably would have started doing that two or three roles earlier with stuff that we worked on just to make it easier to continue to support and provide that into the future. Second one would be Docker. I was kind of slow to Docker as well. I did discover Kubernetes at Call1 as well, and I'm a huge Kubernetes fan now. Again, someone who took concepts and ideas we've had names for for 40 years and threw all the names away and wrote new names, like what the hell is a pod? No, it's a container, but.

Either way, like it was, it was great. Like there's the self -healing, self -growing, self everything methodology of it where you just tell it, I need this application running this Docker image and I need three of them load balanced. And it just does it. Like, you don't, you don't care what it's called. You don't care what it's named. It's just out there and it's available. It's so cool to me to be able to pull that sort of stuff off inside of an application that I would have started using it almost from, you know, the first version that came out rather than sticking to traditional one VM per application deployed in the traditional virtualization environment.

50:48 – The future and advice for aspiring cloud leaders

Matt Pacheco

Where do you see the future of cloud computing heading in the next 5 to 10 years?

Joe Johnson

Ooh, I would say you're going to see a lot more of the individual components drawing more heavily on AI technology. I'm already starting to see some of it come out in the information security side, which is huge, because I look at like my expenses on the security, the single biggest expense is the human factor. Getting human beings eyeballs on glass to look at and respond to events is 85 % of what we have to pay for in terms of security.

Getting a computer system that can respond to and react and actually remediate those alerts as they come out is going to be the next big thing when it comes to that. From the cloud specific side, I think you're going to see a lot of that in the telemetry and the data that comes out of the cloud being able to pipe that into an AI system to respond to and react to it. And I don't limit that to just security. I would say availability and uptime as well. Auto repairing and auto fixing cloud issues based upon AI driven scripts and responses is probably going to be huge for a lot of people. But what did Netflix call it? The chaos monkey? Like that'll be the next big thing is, you know, Cast Monkey 2 .0, which will be AI driven to not only create problems based off AI, but then solve them.

Matt Pacheco

Very interesting. So last question for you. What's the number one piece of advice you give to aspiring cloud leaders and architects?

Joe Johnson

So I've coined a name for title for myself. It's not the CIO, it's the CIFO, the Chief Information Financial Officer. 50 to 60 % of my job is finance now. It's dealing with cost, it's dealing with budgeting, it's dealing with everything that goes into deploying cloud applications and maintaining them, but also paying for them as an organization.

And financial literacy and business literacy is probably the most powerful thing for anyone who wants to become a technology leader now. You gotta be best friends with your CFO and your VP of finance or FP &A because they're the ones that are going to make or break your success. For us, we run an agile accounting system. We're doing forecasting every week for the rest of the year to see where we're at make sure we're maintaining where we're going to be in terms of promises to the board for bookings and sales and even EBITDA and things like that. So it's become a much bigger part of my role than it ever used to be before to the point where it's now almost part of my title in everything that I do. So it's, I think the biggest focus area for a lot of people is it's not just the tech, it's the business and the business is what's gonna make you succeed.

Matt Pacheco

Great advice. And with that, I want to thank you for being on the episode and sharing your insights and expertise with us. This was really eye -opening and educational. I really appreciate you taking the time to join us.

Joe Johnson

Absolutely, Thanks, Matt.

Matt Pacheco

Yeah, thank you. So that's it for this episode of Cloud Currents. To our audience, thanks for listening and watching. If you're watching on YouTube, make sure you like and subscribe and check out some of our other episodes. But we look forward to seeing you again soon. Thank you.