In the next five years, more applications will be written than were written in the previous forty years… Think about supporting that. There aren’t enough humans on the planet for us to be able to manage those, the way we’ve done them in the past, and it’s going to have to be done automated.
Jeff Deverter, CTO at Rackspace, and Max Clark dive deep into the future of multi-cloud and serverless environments, and Jeff explain how Rackspace empowers their customers to better navigate what’s coming.
INTRO: [00:00] Welcome to the Tech Deep Dive podcast, where we let our inner nerd come out and have fun getting into the weeds on all things tech. At Clarksys, we believe tech should make your life better, searching Google is a waste of time, and the right vendor is often one you haven’t heard of before.
Max: [00.18] Hi, I’m Max Clark and I’m talking with Jeff Deverter, who is the CTO of Solutions and Services at Rackspace. Jeff, thanks for joining.
Jeff: [00.24] Max, I’m glad to be here – thanks for having me on your show.
Max: [00.29] I’ve actually been looking forward to this one! So Jeff… Looking at your background, you start in cable and operations and you move into banks, financial services, and you know, from there – I won’t plug one but I use it and I love it – but… From there, you end up at Rackspace, you escape a little while and boomerang back… You’ve kind of been around the horn here, from physical infrastructure, to cloud infrastructure, to physical, to cloud, you know? I’m kind of curious, just to start off talking about what that’s been like for you and how this has evolved over the years?
Jeff: [01.01] Sure! Yeah, the resume does look a little curious when you look at how it’s strung all together, but as I look back on it, there’s some common threads of what caused a lot of it. So, when I jumped into technology – by the way, before that I actually had a pre-career in music, but that’s a whole other podcast, we’ll visit about it at some point! But when I jumped into it, I was really focused on, predominantly from a technology perspective, was the Microsoft stack. It was – very initially – helping build very large server environments that ultimately the US government uses, part of military readiness from a medical perspective, a super interesting project. From there, I did what a lot of folks in technology do, and that is getting into a job that’s going really well, you learn something new, and that creates an opportunity, and off you go to that opportunity! So, I was stuck in this cycle there for a little while, where every couple of years there was a hop involved, and it was always off to something interesting and different. So you know, I was over into banking – obviously in the technology aspect of it – where I was, you know, leading a team, and that’s where I really started to dip my toe… Less from an infrastructure perspective and now in the collaboration suite, specifically Microsoft SharePoint, that was the technology of the day, and that caused me to then make another move, and off I went to another company there in San Antonio – still in the financial services sector, and went a whole lot deeper in the SharePoint space. That created additional opportunity, and that’s where I joined Rackspace – that was in 2008. And for the – you know – couple of three, four jobs before that it was every… you know, year and nine months, two years, there was a jump involved. I realized once I made it to Rackspace – and I went there to launch an offering for them around single tenant, or what we called dedicated SharePoint offering for customers – and I was about two and half years into that job, and I went home from work one day and I talked to my wife, I said, “Michelle, you know how I leave jobs every two years because I get bored or I get an itch and there’s opportunity? Well, I forgot to get bored and I’m having too much fun! So, I’m not going anywhere for awhile!” And it wasn’t until ten years later that I had made a little change, just because it was – I really felt like I fell into a great place, and that great place – Rackspace – was great about investing in what we call ‘Rackers’, the individuals, and I had such an opportunity to build there, and that’s what I love to do. I love to build things, I love to build organizations, I love to build support teams, I love to build new offerings. And so initially it was SharePoint, then it was an exchange offering, a single tenant exchange offering, then it was Skype at the time, then it was the private cloud offering, then the public cloud offering, then the CTO stuff and – again – I’m about ten years into it until I’d gotten to a point where I wanted to make another change. But that change was driven because where Rackspace was at the time, I didn’t have a chance to build anymore… Just in its evolution, and its leadership and what was going on, so… I had an opportunity pop up, and off I went for another year and nine months. Which was going fine, it was great, but I did something I’ve never done before, and that was go back! I boomeranged back to Rackspace, because they had an opportunity – actually I was, I got the call… I was on a business trip and I called my wife, going, “You’re never going to believe who just called me about a job!” She says, “Who?” I said “It’s Rackspace!” She goes, “Well, what’s the opportunity? By the way, you’re crazy, but what’s the opportunity?” And I explained, “It’s the CTO of the organization that builds all the services and products that we provide to our customers, and that means that as CTO, I get to cast the vision for what the next couple of years is going to mean to Rackspace, and collaborate with teams and really cast and put that vision together, and dream… Now, there are other people who are detail-oriented folks who are going to make all that become a reality…” And she goes, “That sounds perfect for you!” I said, “Yes it does, when do I start?” So, at the beginning part of January, right after the holiday break, I’m back to Rackspace. I came, and boy am I glad I’m back.
Max: [04.51] So, as the CTO of Solutions and – I mean, it’s Products and Services, but really Solutions and Services now, right?
Jeff: [04.57] That’s right, thanks for correcting me!
Max: [04.58] How much of this is visionary, how much of it is implementation, how much of this is product design? Like, what is your day job, you know, in this role with Rackspace, and what do you have oversight over?
Jeff: [05.10] Super question. When I came in I was really meant to… The first job I was given was to create the vision deck that we would use for the next two years of everything – it would be presented from the CEO, all the way down. I actually started working on it the week before I started working for the company, which was a first for me as well. Not long after I got in and working with the company – about three weeks in – there was an opportunity to – that they needed to help shore up what we call the offer management process. And so, while the job initially was all visionary type stuff, it was all – not just looking for the opportunity or problems and creating a solution, and then getting consensus around that, which is the heart of what the CTO role was, it was around also then the offering management process, which is how we take products and services and solutions through the process of institutionalized it at Rackspace, so that the entire company can then deliver on the capability, which was very different than the whole vision thing, because this was a detail oriented job, this was making sure that the Is were dotted, the Ts were crossed, that we had alignment… Not just from ‘is this the right technology’, to ‘can we support the technology’ or ‘are we aligned from a legal perspective’, ‘are we aligned and trained from a sales perspective’, all the way through all aspects of the company, not just here in the US, but all around the world, in our EMEA region and APJ. This was – I was meant to come in and just help shore this thing up, create a new structure for it, what its vision would be, and then build a team that would then carry it – which is what I’ve done and handed it off to an extraordinarily talented individual inside of the company, and now he runs that. Now I’m back to just the day job of CTO, which is really where I wanted to be in the first place. So, to answer your question, it is predominantly vision type work, but it is also about building consensus, which is what I find the hardest part to do, because I like to go around and I like to think I can tell somebody something once and they’re going to know it and remember it. But as we know, humans don’t necessarily work that way. So, there’s a lot of repetition, there’s a lot of roadshows, a lot of team meetings, but that’s all great – I love to talk to people.
Max: [07.09] I mean, you touched on this – Rackspace is a large, global company, and when you talk about institutionalizing and maintaining consistency… You know, Rackspace has operationalized and scaled its teams globally and you know, and that does really two things for you, right? On one side, it makes you a little less flexible in terms of customization in some regards, but the other side of it is, when a customer calls in any time of the day and talks to anybody, anywhere in the world, about anything, the support experience is very firm and consistent, and so… And that’s, in many cases, a better trade, you know? You get into a situation where – you know, Rackspace is also known as… The company started as a bare metal or dedicated server company, and it’s evolved up. We’re here to talk about this evolution now into public cloud, and clouds right? And so, can you give me a little bit of the history of Rackspace’s approach to public clouds, and… I think it’ll be good to talk about where this started and where it is today?
Jeff: [08.11] Sure, sure. So, Rackspace… This sounds really funny to some people today, and Rackspace built a multi-billion dollar business on the fact that they could turn a server on for a customer in fourteen days. We cringe at the thought of that today! Fourteen days, can you imagine?! But back in 2005, 2000, the fact that you could get a server at all was impressive, or rent a server. I mean, that’s the whole concept of a managed hoster, which was what we were in the beginning, it was what we invented, we invented that whole market. We could turn on a server or a whole environment for a customer in fourteen days. That would be an environment that would have all the right firewall stuff in front of it, all the infrastructure that we would design, custom for them, all the storage that would be behind it, and a support team that would wrap around them 24/7, to be able to solve any of the problems they had around that infrastructure. Obviously there was a huge gap in the market from a… Maturity is not even the right word, but even from an awareness of what was needed… So, think about a big organization, because this actually lends to a really important point as the cloud growth came along, and that is in a large organization, you had an IT team, and that IT team’s goal – mission – was to provide a consistent level of experience for their computing infrastructure, for the rest of the business, whatever that business might be. You know, selling high-styled wingtip shoes, let’s call it that – it’s a shoe company. And, they wanted to do it consistently, they didn’t want the phone to ring on the weekend, they wanted it to work, they wanted single vendors… But they were wrapped around by a procurement team, because large enterprises leave all the procurement to a dedicated people that can get the best part for the lowest price. You know, if you’re inside of marketing and all you’re trying to do is get a website up so that you can talk about how great your company is, your individual products are inside of your department, and now you’ve got to go to IT and you’ve got follow their checks and fill out their forms, and IT has then got to order infrastructure, and that means it’s going to be dealt with by procurement… Then procurement has to go and find some gear, and eventually it’s going to get shipped by some company: Dell, HP, whomever, and it’s going to show up in a datacenter. Well, they’ve got a backlog of things, who knows when it’s going to get out of a box and get into a rack, and once it’s in the rack it goes on to the kick team, and the kick team’s going to put an OS on there, then it’s got to get harden for the company standards – who are so unique, and so special… And you get the point – it’d just take forever for infrastructure to get online until people found Rackspace! That was the opportunity that we found, and that was, you know what, we can get a server online faster, we could support it in a way that met their outcomes, their needs, and make it available to them. That scratched an itch in the industry, and we grew, grew, grew. We had some huge competitors at the time! If you remember back, so the company started in ‘98, you remember the beginning of the 2000s; there’s a bit of an economic downturn, and we saw a lot of our competitors go away. The people who were fighting really hard would suddenly say, “Hey, you want to buy my company?” So, we came out of that one with a lot less competitors – really the ones that were left were some pretty big telecos who had bought up a lot of our regional competitors. But then in 2008, here comes another economic downturn, and we came out of that super strong because this is a great industry to be in when things get hard, because either companies are looking for better, more economical ways to make it happen and we were able to do that. But it was also about this time the bookseller became a web seller, you know? We think about Amazon in this context. I remember the CEO at the time was walking around handing out a printout of what Bezos was up to with AWS, and he said, “This is something we’re going to have to pay attention to.” So that was in 2010, 2009, I guess that conversation happened. So, we didn’t pay a whole lot of attention, up until the point we thought, “You know what? Maybe we need our own cloud.” Rackspace built the cloud! In fact, we built it with NASA, which was really kind of cool. When Rackspace does something big, they do it really big. It’s just awesome. So, off we go to find NASA – we find NASA – we build our first public cloud, we built this thing called OpenStack, and we take a very revolutionary approach in making it open to the world, we consume this thing; we’re also part of the consortium that runs it… But at this point, we are squarely competing with public cloud vendors. Microsoft is starting to think about what Azure might be, AWS is all-in, and we’re competing with those guys. Is that a great company to compete with? I mean, the logistical geniuses and, we’ll just call them well-funded. And I was there, and it was – it was a really interesting thing happening at this point in the evolution of the company, and that was – the first thing that happened was that when the company started, they built all their own servers. So, literally ordering components, white boxes; they would order – the racks that they went on originally, they were bread drying racks from bakeries, because they could lay them flat, and they found they could get the most of them in there. They didn’t have rack rails, you know? They didn’t have any of that stuff! It was just a white box, like which you would have under your desk. And so then they decided… Alright, we’re getting out of the building our own department, and partnered up with – I think it was Dell at the time – and so now we’re buying servers and a thing happened in a Racker’s brain that… “Oh, maybe it wasn’t me putting memory sticks in this thing that made it special, but we’ll go with another company, but something’s different here.” So, we go with Dell to provide the infrastructure, and then we built our own public cloud. And we started to have people in pockets of the company saying, “Look, it’s not the infrastructure that makes us unique and special, it is what Rackers do with that infrastructure that solves customer business problems. We should support Amazon, we should support Microsoft, we should support – what would become Google.” And those were some fights – I mean… I think there may still be some hardcore folks in the back corner going, “Argh, I like our servers better!” But people had a lot of problems with that because it was scary, because Rackspace made a decent margin on that infrastructure, because we would have to change everything about what did, from a costing model, to a delivery model, from the teams that would support it and the way that we would support it… The biggest thing – I’ll tell you next – the biggest concern people had was, how can we provide an SLA on infrastructure we can’t control? How do you do it?
Max: [14.11] I mean, that’s a constant IT question, right? It’s funny – I’m laughing listening to you talking and explaining this, because all infrastructure at some point becomes commoditized, right? I mean, servers become commoditized, at that point it just becomes scale and margin. How can you compete at high scale with very low margins? And we’re seeing networks becoming commoditized right now, and these white box networking networking devices coming out at the same time, and for a… You know, if you look at public cloud, public cloud is commoditized! There’s differentiators and different options, but to some degree people don’t think about it so much, in terms of… “I need this logo,” I think about it like, “I need this many compute, I need this much RAM, I need this many functions to execute, I need this much storage,” right? I think the mentality is changing, but still supporting and operating these things at scale and all the time, that becomes very interesting, and it’s also become a very big shift for an enterprise and how they manage their IT infrastructure. When you go into a public cloud environment, things are very different. Years ago, Rackspace was a lot more inflexible about how to approach and support people, right? That was, I think, the dipping the toes, and now when we talk about Rackspace support and how you actually layer offerings to people, that’s become a lot more flexible, you know? You’re a little bit more like, “What do you need from us and we’ll plug those holes for you,” right? It’s a slightly different approach?
Jeff: [15.38] Well, the phrase that became really popular, once we realized, “Okay, we had some customers in AWS who’ve got some customers in Microsoft… The world didn’t end, our business is still growing,” and then we decided to really run after it. The phrase that we used as kind of the ‘calling card’ around Rackspace was “Up the stack and out of the rack,” because – just like you said – the infrastructure was commoditized. We just can’t care about it as much – we have to care that it works, but can’t hang our hat on that being the end all, be all. If you go back to the beginning of the story, we could provide a server with literally nothing on it but an operating system, in fourteen days, and that had value. If I just provide a server today, that has no value – we’ve got to have security around that, we’ve got to have all kinds of other things naturally baked into the whole story, and it can’t just be about the server – it has to be about the customer outcome on the other end. You see, even now, like in the early days – here’s a little secret – in the early days, we would sign a customer to a one, two or three year contract, and it wasn’t a stated goal, but what ultimately happened was – a lot of times – we would respond to stuff, but there wasn’t a lot of proactive reaching out to a customer, because not a lot would happen on that infrastructure in the intervening one, two or three years… Just not that much changed – I mean, all the development cycles were waterfall, it barely saw updates, and those updates rarely needed more infrastructure, customers’ account bases weren’t growing so fast that they had to throw tons of things at it. We had a few customers grow, and we just added more servers, but generally we didn’t talk to customers all that often. Today, we have to talk to them every day. Let me rephrase that – we get to talk to them everyday, because we designed into the whole process now – as opposed to just you know, “Here’s an architecture design, we’re going to deploy the servers, have fun storming the castle, we’ll see you in a few years.” It’s now… We’re going to design into that the optimization process, from the very beginning part of the conversation, because if you’re going to deploy out into the cloud, whether you change or not, that cloud is evolving underneath your feet on a literal moment by moment basis, and what awesomeness we deploy today may actually be out of – not as perfect as it could be tomorrow, just because the cloud got better underneath it, and there’s a better way.
Max: [17.41] We shouldn’t talk about this in terms of a specific cloud, because you are cloud agnostic – I mean, you have a practice around Amazon, you have a practice around Google, you have a practice Azure, right? So, it doesn’t matter where the customer wants to be – and you still have you still maintain your own private cloud practices. I mean, that’s –
Jeff: [17.56] Which are growing, interestingly enough!
Max: [17.59] I mean, as it relates to enterprise transitions to public cloud, or an enterprise is in public cloud and they come and they say, “Okay, we need to do things differently.” I mean, what is Rackspace – I mean, how do you… How do you explain this in a short span? I mean, what is it that Rackspace is doing for these companies? Let’s start broad, and let’s get into the weeds.
Jeff: [18.20] Yeah, so it’s great that you mention that and it’s great that you very correctly say, you know… And when you go into that environment, how do we have a conversation with them around what the future looks like? Because when you … In the old days when we would have to do an update at some point, it was – think about it going from VMWare X to VMWare Y, or this old piece of hardware to this new piece of hardware… It’s a very waterfall event. But when you adopt a cloud – and again, fill in the blank on the cloud of choice – when you adopt cloud, you’re not adopting just a real estate location. In some cases you’re not even doing that. You are adopting a methodology… To say, “I am cloud,” or “I have cloud infrastructure,” then what you’re saying is you’re adopting a more frequent update process, you are adopting a commitment to upgrading your application to utilize more efficient mechanisms, and we’ll define what efficiency means in a minute, for how that compute happens, whether you get down to the point of containers, or even serverless; how your storage works. You know, in the old days at Rackspace when someone would say, “I want redundant storage across multiple datacenters,” you’re just 2x’ing or 3x’ing your cost, because not only do you have to double – literally double – the physical infrastructure that’s holding the storage in two different locations, you’ve got to get special software to make it stay in sync between them. Now it’s just an extra line in a script or a checkbox in an interface to say, “Uh, replicate this in three or four or twenty-seven different locations, and make it available for fill in the blank number of days, weeks, months.” And so, Rackspace has to take all of that into account, and here’s the other thing Rackspace has to do, and this is very different: is the move to the cloud – and again, that’s not just real estate, that’s methodology – is one that if… If a service provider like Rackspace doesn’t go into it with eyes wide open, we have to continually show value – even after that architecture is in place, and it’s been built and the customer’s been moved in. We’ll lose that customer because it just sort of… Chunks along and works. And, the amount that they would need Rackspace involved becomes less and less, which is why Rackspace now focuses a lot more on applications, and even application development, and teaching customers how to become – air quotes for those who can’t see them, because it’s not video – to become “cloudy”; and that’s ultimately the goal. So, we talk about that in cloud maturity, we talk about multi-cloud, we talk about all these sorts of things.
Max: [20.46] So, cloud is… Cloud maturity, and being “cloudy”, right? This is more than just taking and virtualization infrastructure in your office or in your datacenter and then putting it into a public cloud, right? And, when we look at the different public cloud platforms themselves, you have, you know, base infrastructure – so you have compute, you have storage, you have these building blocks, and then they all offer managed services on top of it, whether or not it’s a managed day-to-day service, or a managed objects store, or a managed, you know, columnar storage, or something along those lines, right? And people don’t, I think, understand where they’re actually spending money in their clouds, and what actually causes pain for them as they make these migrations. I mean, the comparison I’ve always asked… You know, should I run this method of compute in Amazon, or is it cheaper in Google? I mean, it’s almost negligible and who cares, because that’s not where you’re spending money. You’re spending money in your support infrastructure and in supporting that cloud, you’re spending money in your egress costs, moving data in and out of that cloud, you’re moving data in your data science, or your machine learning… You know, pipelines. These things become very expensive, but not necessarily, ‘is it cheaper to run this compute here, or this compute there, or this compute in this other place?’ And this is something that Rackspace, you guys help a lot of people with these decisions, and also maybe optimizations or – what’s the word I’m looking for – you know, migration of workflows, off of one environment to another to optimize that.
Jeff: [22.13] So, we think about cost optimization a little differently… So, as a customer – as a company goes down the road into cloud, and cloud evolution from a maturity perspective, you know, they have a decision to make, and I’ve been having a similar conversation, the same conversation, for about the past four years, and that’s where somebody says, you know, “Should I be worried,” I hate this conversation… “Should I be worried about vendor lock-in?” Everybody’s worried about vendor lock-in, and it’s sort of right, you know? If you’re all committed on Dell, and Dell raises their hardware, then you know, how easy or hard is it to move over to HP, or whatever it might be? But when you think about vendor lock-in from a cloud perspective, what this does – so, what they’re wanting to do is if all of a sudden a favorite cloud provider of choice raises their rates, they could easily move to another cloud provider, but what this does… First of all, nobody is going to raise their rates, right? They’re afraid of it. But what it does is it forces them to the lowest common denominator of capability. If their only concern is lowering cost through being concerned about vendor lock-in, then they are – they should absolutely go back and buy servers and put them into their datacenter or come to Rackspace for private cloud technologies, because they’re not going to evolve. So, if you only are focused on that, then you’re not going to use the higher order of things. You rightly said it before, and that is that the cloud now is more than just infrastructure, and there’s all sorts of things that clouds are building on top of it, and when you talked about how Rackspace, you know, was dealing with commoditized infrastructure… They realized that VM is a commoditized piece of infrastructure, and so what they’ve done is they’ve done two things: they’ve one, found ways to make VM non-commoditized, think about your high-compute machines, you know, if you’ve got GPUs in them that’ll do special stuff – so they have specialized things they can charge a handsome price for. But for the other stuff, they don’t care – if they just want your stuff, your storage there, especially if they get your data, then they’ve got you. But once you’re in, if you start to peel away those monolithic applications that you originally moved in, and start to use the higher order of things, what it does is it gives your business power – it gives your business… It puts your business back in the driver seat, because IT is no longer sitting there in hand-in-hand with procurement going, “Where can I find the cheapest cloud?”. I’ve been in those conversations, and they’re miserable. But now you’re in a conversation where IT is working hand-in-hand with the business, to help direct them to create a technology suite, which serves them and puts them into the driver’s seat from a cost governance perspective, from a technology governance perspective… For instance, if a customer is able to take their application and break it down into serverless, and what serverless does for them in this scenario is they don’t have to think about the infrastructure, the cloud provider does all that. It patches it, they make it available, they make it redundant, they do all the things. And what they do is then they charge you by the drink to use that service. Well, what that means then is IT can now very much measure how many transactions to do one order of function. Let’s say it’s a – marketing is coming along and says, “Hey Mr IT, I want to put up a Super Bowl ad, and we – we’re putting up a Super Bowl ad, and you need to put up this page to capture all of our awesome viewers who are going to come and look at it!” Well, they can deconstruct that webpage, you know, how many transactions are going to be involved, and they can measure that against what they’re being charged by the cloud provider for those transactions, and they’ll turn back to the business and say, “It is going to cost you x number to actually do that, is that worth it to you to make this decision?” See, it actually puts control back in where they wouldn’t have even had any idea what it would take from an infrastructure to put online to be able to answer the call from a Super Bowl ad type of scenario. So, I think it’s a long way to answer your question to say, as a company evolves, and they stop thinking about infrastructure and start thinking about capabilities in the cloud that they’re in, or choose a cloud – not on who has the cheapest VM but who has the technology and the reach; geography – that ultimately meets their needs, and then go all in. My recommendation is to go all in.
Max: [26.10] A conversation I have often relates to core infrastructure, building blocks, versus managed services, you know? For example, Kafka, right? Are you going to run Kafka on your own infrastructure, or are you going to run Kinesis? You know, are you going to run Cassandra, or the various – there’s too many database tools, or are you going to run something that’s managed for you, you know, DynamoDB, and you know, the way that I relate it now in a lot of ways is more about velocity. There’s – and I had a CTO that was much smarter than I years ago explain this to me in a sense of, you know, if you have finite resources in terms of – usually it’s people – if you have finite resources in people that can implement processes for your business, do you spend money in trying to bring more people into the environment, or do you take a shortcut and you go buy something off the rack? You know, and this was – so you’re talking this build versus buy decision, always. And clouds do a really good job of providing this infrastructure and saying, “Oh, you don’t have to figure out how to run Kafka, you can just turn it on and run it,” and a year and a half down the road when you realize it’s costing you a lot of money, well maybe then you’re ready to operationalize it and move it back into your own infrastructure. But these are all decisions also an enterprise would have to make, but with a partner like Rackspace, it could be, “Hey, we’re going to run a managed service in the cloud, and okay, we’ve now identified this as something that we don’t want to run anymore, so can you operate this for us and adjust how we’re getting billed from -” and you know, maybe that’s… We don’t want a per-transaction, we want predictable monthly spend, you know? We want to know this is going to cost us forty-thousand dollars a month, every month for the rest of the year, and we can scale it up or scale it down. I mean, that’s… That has a lot of value – you talk about value creation, I mean… Helping people make these things definitely has a lot of value.
Jeff: [27.50] Yeah, so Rackspace is filled with a ton of smart people, but some of the smartest in my opinion are what we call Sales Engineers, or Sales Architects. And so, think about the – breakdown the whole delineation of who’s involved in that whole sales and support structure… You’ve got somebody on the front end who’s going to be that smiling person, who is the front of salesperson, but they’re one hundred percent supported by those Sales Engineers, and it’s the Sales Engineer’s job to sit with the customer and listen to their business challenges: what is the customer trying to solve? And we’ll try to – we’ll also ask, you know, “What’s your favourite technology,” along the way, but they’ve chosen a cloud – even the people who say they haven’t chosen a cloud, they have a favorite – and we’ll wrap them with a Sales Engineer who knows that cloud inside and out, and here’s where I think one of the massive pieces of value that Rackspace brings to the table, and that is on a day to day basis, we’re architecting and designing with the latest cloud technologies and staying ahead on what those latest cloud technologies – those solutions for the customer. It’s impossible for a customer – any customer, on size – to be able to really stay abreast of everything that in AWS is deploying on a day to day basis, or either… The other two providers. But these guys do, and they stay trained, and they’re self-organized, and they’re doing some amazing work, and so they actually can help make those low-level types of decisions. Well, what are the goals for your data storage? You know, help decide, do we need to architect some datas storage that will meet those needs, or can we use an as a service and let that kind of happen. And just as there’s super-smart people on that Sales Engineer side to architect the stuff, they’re handing those workloads off to a professional services team to do the build, and then they’re handing that work off to a support team to then manage on a day-to-day basis. I can see you’re about to raise your hand Max and go, “But it’s in the cloud, what do you have to manage? Doesn’t, you know, Amazon or Microsoft take care of all of that?” Well, they’re going to make sure that the infrastructure’s running, but we’ll make sure that the application – the solution – stays functional, current, secure, all those sorts of things.
Max: [29.58] You talk about, you know, what infrastructure support comes from the cloud vendor, and a lot of cases, especially as you get larger as a – consumption, a cloud provider will force you into support agreements with them. But, that’s not necessarily like a golden ticket into support resolutions with the cloud vendor. So, in an engagement with Rackspace… Rackspace staff and support are triaging and managing and troubleshooting, and doing everything that needs to happen before escalating and saying, “Okay, we know this is an infrastructure problem, Amazon or Google or whatever, we need you to fix this for our customer.” But you have leverage and scale now, because by the time you talk to AWS and you call up on the phone, it’s like, “Okay, we know you’ve already done all these other things,” you know, you can skip a lot of lines. What is that like with that infrastructure control shifted, right? You’re not talking about… Go to the datacenter and touch a box, now it’s escalated to an infrastructure provider – but that is a value, and that is something that you do offer and that people do need.
Jeff: [31.00] Yeah, it’s very much a different angle. Now, granted a couple of the providers are now started to offer a managed server, and let’s just define real quick what ‘managed service’ means. It means that you’re going to pay a company, maybe even an infrastructure provider, some money a month – however they decide to charge for that – and what that means is, they’re going to do some sort of extra service on top of the infrastructure that’s just not part of when you go and buy it. In some cases, it means something as simple as… You can get to a higher level technician – not coming in through tier one, but coming in through tier four – and if you’re super big, or you pay a ton of money, then you maybe even have a dedicated person, or a technical account manager, who is someone that’s familiar with or aware of what your infrastructure is trying to do, and becomes a good touchpoint to then rally all the different people inside of that cloud provider to fix whatever the problem might be. But a lot of times Max, where they’re drawing the line is, “Oh, that’s your application? I can’t touch your application. My infrastructure works!” Or, “I think that your VM is misconfigured in this way,” and they’re not actually allowed to log into your VM. You’ll still have to go and effect that change. When we go with a provider like a Rackspace, you know – this is actually… If we skip out our infrastructure, you know, we loved our white boxes in the early days, and we loved our own servers, but in the early days, the value is the fact that while we provided that server, we provided and made sure that that server – and in some cases the application itself – had a guaranteed uptime, and when you called and said that it doesn’t work or we sensed that it didn’t work, then we were already on it and making that change, and ensuring that… Not just that the disk is spinning or that the c-prompt is blinking, but that your application was doing what it was designed to do. We would implement monitoring techniques that would do transaction or a synthetic type monitoring to make sure it was one, behaving, two, behaving in a timely manner and things along those lines.
Max: [32.58] So, I mean if you have SAP HANA, right? I mean, that’s an application that’s pretty understood in terms of support delivery, and you guys can operationalize SAP HANA… Pretty straightforward. When you’re talking about a customer who’s written their own application – and in theory has a CI/CD pipeline, and is doing regression testing… I mean, even so – that environment’s going to change pretty frequently. That changes your operational structure in order to support that application. I mean, how do you operationalise a customer’s application to help them support themselves in whatever environment it’s in?
Jeff: [33.33] So that’s a really good question, and I’ll answer it in a couple of parts. The first is how we have done that – and there’s a couple of ways we’ve done it. One, we’ve gotten really involved… When a customer is a mature customer like that – and by the way, if they’ve got a CI/CD pipeline and they’re doing daily… Minutely, whatever the timeline is, builds – this is an advanced company in the grand scheme of cloudy type companies, and we have an offering that can actually come alongside of them and provide support along that way, and support the pipeline and understand the code. Now, that is – to be transparent – those are people being involved in your process, in your technology. But Max, that’s not going to scale. You used the word before, and I wrote it down on my piece of paper because I wanted to come back to it, and that was the word ‘velocity’. In the early days, speed was enough for Rackspace: server on in fourteen days, good. It wasn’t – it just didn’t get any more complex than that. But today, speed without direction may just ascend you into a brick wall, and that’s what – of course, velocity is speed with direction, and that’s where our focus is: how do we help a company accelerate their digital transformation, their corporate transformation, through the use of technology? And we are actively working on projects right now for what will happen in the next three years. I was watching something by Microsoft earlier today, and that was… They were talking about an application development thing, and the short story is – the future of Rackspace is to be embedded and involved in the customer’s custom development lifecycle, in providing a CI/CD pipeline, in being involved and connected into theirs from a monitoring and awareness perspective, applying data models to that, being – applying a machine learning algorithm to that, and helping them to understand what good deployments are programmatically, so we can get to a point where AI Ops really does become a reality. Because, the amount of applications that we’re going to be faced with in the next five years is stunning, the explosion that’s happening. You can look at it from a couple of different angles: one would be the amount of storage, I think we’re going from seventy-five zettabytes to a hundred and fifty seven zettabytes – or a hundred and seventy five zettabytes by 2025. So, you look at it from a data perspective, the same thing holds true from a device perspective, I don’t remember the exact numbers… But the stat I heard this morning – which was in a Microsoft broadcast by the way, but it was – they were citing a Gartner statistic, and that was: in the next five years, more applications will be written than were written in the previous forty years… Think about supporting that. There aren’t enough humans on the planet for us to be able to manage those, the way we’ve done them in the past, and it’s going to have to be done automated.
MID-ROLL: [36.13] Hi, I’m Max Clark and you’re listening to the Tech Deep Dive podcast. At Clarksys, we believe that tech should make your life better, searching Google is a waste of time, and the right vendor is often one you haven’t heard of before. With thousands of negotiated contracts, Clarksys has helped hundreds of businesses source and implement the right tech at the right price. If you’re looking for a new vendor and want to have peace of mind knowing you’ve made the right decision, visit us at Clarksys.com to schedule an intro call.
Max: [36.37] So, a few years ago I had a CTO tell me… I mean, basically he was at a company and he had to travel a lot. So, he was constantly on a plane between where he lived and where his office was, with the corporate parent. You know, and it was a long plane flight, maybe four or five hours, but he recognized a certain degree of stress: every time he was on a plane, what was happening with his infrastructure, running in a cloud provider? Was it serving traffic, were they able to conduct commerce, and serve their customers, and on-board… You know, everything that comes along with… “We have a business connected to the internet, what are we doing with it,” right? What he was looking for at that time was he was looking for a partner that could help manage his environment – not necessarily do application development – but manage and make sure things were running, like, who was responsible for keeping everything running? That partnership ended up becoming with Rackspace, it’s been very successful for them. And this leads me to a question about how you actually implement and what service blocks are, and how people make decisions, because there’s a lot of options now from Rackspace, and what used to be this all-or-nothing kind of, one-size-fits-everybody forced approach, is now become a very flexible, nimble… “Here’s all the different pieces we can offer you, which ones do you want?” So, how do you actually layer – what is a service block, how does this layer together, and how do people… Understand – “I want these things, but I maybe don’t need this other thing right now.”
Jeff: [38.06] Yeah, so in the earliest days of Rackspace, we had – our support model was what we called a denial of support model. We didn’t answer the phone – it was not good. This was the first couple of years, and at one point one of the original founders had accidentally answered the phone, and it was a customer who had been down for several days. This customer was livid, and through that highly energetic conversation, this founder realized… He hung up the phone, and he slammed his hands down on the card table, because this was not a fancy company at the time – very early Rackspace, first couple of years. And he said, “We have got to get fanatical about customer service.” This was against everything in the previous couple of years, because these were techies, these were technical geeks, and they were like, “If you can’t run a server, we shouldn’t sell you a server.” But they realized that they had to support it! And what that meant over years has evolved, we now call it fanatical experience, but deep, deep, deep within our DNA at Rackspace is the fact that we will do whatever it takes to serve that customer and make sure that they have an amazing outcome. In the early days, our support – the way that we charged for it was built into the cost of the server; you just paid for it. Then when cloud came along, we adopted as similar fashion, and I talked about this earlier, and that was we have to continue at Rackspace to figure out how to provide value, because – here’s the hint – in the early days we didn’t figure that out, and we had some customers leave us – because after they were deployed and things were just working, you know what, the cloud provider had monitoring… Hm. You know what, the cloud provider has back-up… Hm. You know what, the cloud provider, I could probably eventually get on the phone and help me – and serve it… While they did still see some value in Rackspace, maybe they didn’t see the value for everything we were charging them for, and so along comes service blocks! What we did was, we blew up the whole support model, and it was one of the smartest things I think Rackspace has ever done. What they did was, instead of saying, “Alright, if you’re going to go and buy a hundred dollars of infrastructure, I’m going to charge you twenty dollars a month to support it. A fixed, you know, variable model based on how much you consume.” And as opposed to that – and then saying, “Okay, everything that we would normally do for a customer, we’re going to break down into these bite sized, or fun sized bits of service and support that you can add and remove without penalty from Rackspace.” So, when you’re going through that… The initial move out into Rackspace, that whole ‘advise and design’, that’s a service block. You consume that, you get a tangible deliverable, and that feeds into the migration work – it’s another service block. You buy that service block to get the migration stuff, as your cloud starts to come online, you have foundational capabilities, we call it foundation services, to be able to do the blocking and tackling of having – managing things in the cloud: the monitoring, ticketing, just very basic stuff; helping you with tagging strategies -without doing a lot proactively, very reactive, but the more embedded you want Rackspace to be involved in that process, there’s a service block for that, and you can make as a customer that decision. You will not be given – in our cloud based environments – you will not be given a fixed-fee managed services capability, you’ll be able to tailor that from a capability perspective to best meet what it is you want to do. You get to a point where things are sort of just working and you’ve got other things to focus on in your business – scale some of that stuff back down, get it to where it makes sense, keep us doing the base stuff, and as soon as you need to go deeper with this, scale us back up. Again, all without penalty, which is unheard of.
Max: [41.24] Something that Rackspace does as part of your foundational service is that – I don’t think people understand – is you include cost monitoring, optimization and governance tools that are automated. I mean, this is something that’s a very sophisticated intellectual property around these things, but anybody who’s ever tried to dig into their AWS bill and figure out what they’re actually paying and how much and why and what for and other things – and people talk about it as like, “Oh, is this VM running, or is it the right size?” I mean, that’s the really sophisticated stuff; there’s like basic things of like, “I disconnected a VM but I didn’t get rid of block storage for it, so I have all these EBS volumes that are still around.” These things pile up and all of a sudden you’re like, “Well, that’s a significant amount of money each month.” So, you know, with your foundational services, you do include software, you do include systems to allow people to actually better understand what they’re spending: do they want to set thresholds, do they want to set alerts on these things, they want to set notifications, so hey – oops, somebody spun something up and forgot to turn it down, we can alert before the bill comes at the end of the month, or do we have a security policy set up properly and is this being tracked, and you know, are things changing that shouldn’t change? And even within that, if you look at what it would cost to acquire that technology and that capability outside of it, I mean – a lot of that cost, I mean, it’s incredible that it’s included – you really should talk about it more, because people need that. I mean, there’s a resistance to it for the whole sense of, “You’re going to help me figure out how to go into reserved instances,” or, “You’re going to help me figure out – you know – instance resizing, but I can’t do it,” so that’s not really of value, but the flip of it is actually having… Reporting and data and trend analysis.
Jeff: [43.02] We’re in this alone, exactly. And the fact that those tools were cross-cloud as well, so if you were a multi-cloud customer, you’re able to get visibility and that same kind of response and reporting across all of those clouds.
Max: [43.15] It’s interesting that you mentioned tag strategies, and anybody who’s been in public cloud for any amount of time and that actually wants to come back and be able to do divisional billing or reporting, all of a sudden you start talking about tagging, and you actually roll out tagging and there’s something of Rackspace’s onboarding of helping people go through and say, “Okay, here’s a tagging strategy and how do we actually tag resources,” because that influences what is your billing structure – how do you support all this other thing, how do you report against it, how do you say… You know? “Hey, I’m the IT department, I’m managing this stuff, I’m not expensive – marketing had a Super Bowl commercial that was really expensive last month, it wasn’t my fault!” That’s – these are all things that I don’t – I don’t think a lot of places think about until after the fact, and going into a process with Rackspace where you’re bringing this up at the beginning of it, and really trying to help drive into operational excellence… These are all good places for the enterprise.
Jeff: [44.12] Well – and another thing they don’t necessarily think about is, a lot of companies, I mentioned earlier before – where strategy to help from vendor lock-in is, they will have – okay, we’re going to make two clouds available to the enterprise. Go. Now, what happens immediately – boom, factions form! You’ve got the AWS camp, you’ve got the Microsoft or Google camp over here, and they may be sitting next to each other in the lunchroom but that’s as friendly as they’re ever going to be. God forbid they would agree on tagging strategy, on – even any of the other architectural elements, unless it’s a super large organization that can really heavy handedly do it. But what Rackspace realizes is, this is a challenge, and we’re in the process of creating some new technology which I am spending a majority of my time on these days, and I’m so excited about… But what it’s going to do is it’s going to operationalize multi-cloud for customers. It’s not going to – so Rackspace today, we do that by force, and we do that just through sheer human will and common account teams who connect the dots, and that – transparency here – they’re doing that manually. What we’re in the process of creating right now, is a commonality across clouds for customers. Not to where we want you to use a common cloud management portal, we don’t want to homogenize the cloud experience, we want to homogenize all of the things that you have to do solve for – that really have nothing to do with, “How do I make a server run better,” or, “How do I make my application run better?” Think of it like… Tagging is a huge one, cost governance is a huge one, usage governance is a huge one… Back-up! How do we ensure that I have common back-up and monitoring across everything, and be able to look at all that stuff from a single pane of glass, or have API access to that that I can adjust that and deal with it in my own internal systems? So, that’s the next piece that Rackspace is working on – we’ve got the MVP under construction at the moment, and it’ll start to see the light of day in the next few months.
Max: [45.59] Multi-cloud is a huge, consistent topic, and I think every serious public cloud deployment ends up in some version of multi-cloud and I feel like that conversation starts with, “We’re going to have redundancy against cloud vendors,” and then it shifts very quickly to, you know, “We have so much ingrained application in one that we can’t effectively go multi-cloud, because shifting data and synchronizing data between environments becomes the complicated piece of it.” But it’s… I’m on cloud vendor A, and cloud vendor B has a very interesting application that I want to run, you know? And the first ones that we really saw this a lot with was the machine learning pipelines, you know? Taking out of Redshift and going into Google in order to run Big Query in TensorFlow, right? That became a huge, popular thing. Now, AWS would argue that if the application was re-implemented, that Redshift would be just as efficient and what Redshift is doing now would be even easier, and that’s a different conversation, right? That comes into architectural decisions and application pipelining and all these different things that you guys can help with as well, but where do you see multi-cloud going over the next… Let’s call it two years, right? That’s a long enough horizon that we can have a serious chat about. What is the reality of multi-cloud?
Jeff: [47.19] So the reality of multi-cloud is, everybody will be multi-cloud. In the early days, unless you were a really large company, I actively tried to dissuade companies from utilizing multiple clouds, because you have to solve for so many things every time, all the stuff I just mentioned, and not to mention the internal expertise that you have to maintain, and train, and hire for, and manage, and all the above. But at this point, we’re at a point where companies are big enough, clouds are mature enough, interoperability is good enough, that the world will be multi-cloud, even if you think of it as something as simple as, “I’m using Office 365 over there, but I’m – GCP is my data cloud of choice, and I’m putting all of my other compute and web-pasting properties over in AWS.” That is a very real scenario that’s happening over and over and over again, out in the wild, and it’s only going to get worse because more things are going to be cloud based. Less and less will we see infrastructure being used, and less and less will we see companies – I believe – just dedicated to ‘how can I do it all in one’. Now, if you are a smaller company, I highly recommend you find a cloud and hold on close, and find one that can serve you all the way across, and that really is going to – that’s going to narrow you down more on… And when I say ‘all the way across’, I mean from collaboration to end user tools, to your cloud based things. You really have two choices here, it’s Google or it’s going to be Microsoft. Though as you get larger, you can think about multi-cloud. The other thing that I think is going to drive multi-cloud’s story all the more are containers, because containers are the great democratizer in my opinion. Because it’s this thing that I can pick up, and it’s going to move, and it’s going to run the same on my laptop, it’s going to run the same on my dev farm, it’s going to run the same out in the wild on a bunch of servers, and it can be built into the CI/CD pipeline, and I can broadcast a brand new application stack every single time, in cloud of choice, and I think that’s going to be a huge driver of multi-cloud.
Max: [49.22] I mean, containers are awesome, it’s an awesome piece of technology, if you’ve ever done application development and tried to figure out what your production release cycle looks like, containers make your life a lot easier, and for portability – as you said – they make your life a lot easier. But, a question for you: as more and more enterprises come into the fold with containers, are they already behind the trend a little bit? I mean, have containers already been jumped into serverless? Should we be talking and focusing on application deployment to a serverless infrastructure as opposed to container application development into, you know, whatever your container orchestration is?
Jeff: [49.56] Right, so I don’t think – I think all companies are behind right now… That’s a starting block; second piece is… It depends, is my other answer, because if vendor lock-in is a concern – and I throw a lot of rocks at lock-in, but containers are a good way to get around that. The other piece is that it is a great way – in my opinion – to deconstruct monolithic applications into the next logical building block. So, I can take a VM and now I can start to containerize applications, and it does – you get economy as a scale, let’s be real about that. There are benefits to utilizing them from a cost perspective, from an availability perspective, even if you’re thinking about on-prem private clouds, or a Rackspace private cloud, or bursting or utilizing resources out in a public cloud. So it is, in my opinion, a logical stepping stone to serverless, but while I do believe that some workloads will remain and make sense to leave in a VM-type construct, more of them will be left in the container construct, but then still tons will be in the serverless world. And yeah, their explosion from it… And it’s – to come at the conversation, Max, from a different direction, so we’re thinking about these big applications, and how do I turn them into and make them cloudy? Well, think about what’s happening, and the revolution that is happening right now, in the low-code/no-code application suite? All of that is a hundred percent enabled by serverless, so they get to completely bypass the whole VM world, because you know, they’re utilizing an application suite or a cloud provider of choice, but they’re completely circumventing the whole… VM/container conversation, and going straight to things that are enabled through serverless technologies in the cloud.
Max: [51.44] I got into the internet originally with a modem, so everything that is happening in the world – on a constant basis – it just never ceases to amaze me, and you know, we’re spoiled! I mean, we have this LTE cell phone modems that are insanely fast in certain cities, and all these things, but… You know, when you look at what’s available, and we can get into like, really deep tech conversations, but I go back to what you were saying a lot in the beginning of this, which is: how do you enable an organization to do whatever it is that that organization needs to do, right? It’s like – tech choices, we can get into battles, and you can have cloud A versus cloud B, or server vendor A versus B, or network vendor this, or you know – or container versus serverless, but we really still have to connect all this back to… What is the company – what is the organization actually trying to do? Like, what is its mission, how does it serve its customers, you know? That’s – and I feel like that does get lost a lot, you know? People get in – fall into the tech, and they lose sight of what is actually the goal of all this stuff.
Jeff: [52.45] Folks like you and I love to get lost in the tech, it’s just what we’ve done forever. I mean, you’ve probably bound two different phone lines with modems just to get a little faster than 566, or 144, if you remember those days, or 9600 – but that’s the thing… When a company goes through this transition, it’s not about the technology. Technology is one of the three legs of the stool: it’s people, it’s process, and then it’s technology, and the easiest piece of this is the technology. We get enamoured by it because we like the shiny tech bits, but the people and the process is the hardest thing, and the – I love using the word, because I think it really applies here – and that’s the revolution that’s happening inside of large enterprises, and it’s happening there first because they can afford it. And what they’re doing to enable – and this is, I don’t know if it’s a Microsoft word – but I just hear them using it the most, and that is the ‘citizen developer’. And they’ve got some great tools in their PowerApps platform to be able to do this, but Gartner’s number – I’ve obviously just watched a webinar earlier today – but Gartner’s number is that in the next four years, sixty-five percent of enterprise apps will move to no-code/low-code, and that is a massive transformation of – the only way they can do that by the way, is you can’t use developers; you have to use the business analyst, you have to use the people who have been in the business forever and understand the business – and the tools are sophisticated enough that you can literally drag and drop a story, or just write a couple of lines of T-SQL to get what you need out for a report… To be able to create a meaningful solution that will never touch corporate IT – can you believe that? Never touch corporate IT. The other quote that I’ve heard that is just stunning is that Ford Motor Company now has more computer engineers and coders than they have mechanical engineers. Car companies aren’t car companies anymore. I mean, we know that from Tesla’s perspective, but the other ones – and that revolution is happening everywhere, and that’s what’s exciting about being in tech right now.
Max: [54.42] And you get to sit at an interesting intersection with all this, right, because companies that are not tech companies that have to figure out how to become tech companies, or have to figure out how to operate tech for their companies, or maybe it’s a further extreme, that they want to be in a position where they don’t want to be tech companies at all, they just want a partner to support and enable their business, right? I mean, this becomes an intersection. I mean, Rackspace originally was a logistics company, predominantly, right? Could you buy the root equipment and build it, and put it into a facility and connect it to the internet? How efficient were you at performing those steps, right? And then, could you answer the phone and support people and help them log-in and change their passwords and install an application or whatever? You’re still – for the most part – you still have that logistics kind of… You know, brain going in, but now it’s about server infrastructure and more about applications and where applications are running, how the businesses actually interact with them.
Jeff: [55.33] Yeah, yeah – I mean, you look at the evolution, and it was from: we have our own servers, happy with that, then – okay, on demand computing is probably going to be a thing, we’ll build our own server so you can come on-demand compute with Rackspace on cloud. And then it was, well really, the service doesn’t make us special – Rackers make us special, and what we do for customers, so let’s go support things on the public clouds. And then it wasn’t just, well we’re supporting the infrastructure, if the infrastructure’s so darn important, then maybe those applications are… So we got really deep in SAP and Oracle and SalesForce, so supporting specific applications as deep as we did everything else, and then Rackspace’s latest pivot just looks to where to we’re going, is when we acquired Onica – announced it last fall, finalized it this spring, and that’s all around custom development. It’s around how do we help companies do go through that transformation? Not just from a technology perspective, but from a people and a process perspective. Some companies just hire us to come in and be referees, if you can believe that – to help grow them and their capabilities, and help make sure that they like each other at the end of the day.
Max: [56.37] Look Jeff, this is fantastic, thank you so much for your time, I promise – if you’re willing to do it – I’d love to have another one, when we can get into the religious debate of monolithic versus microservice based application development –
Jeff: [56.47] Oh, yes!
Max: [56.48] – and then whether or not that should be deployed in containers or serverless, I think we could probably – but I will, you know, we’ll have to give ourselves a lot of time to get into that chat.
Jeff: [56.57] No kidding! Hey, I have really appreciated being on today Max, thank you so much. We really appreciate your partnership, we appreciate you just being part of the Rackspace larger family, and everything that we get to do together.
Max: [57.10] Thank you Jeff, talk to you soon.
OUTRO: [57.16] Thanks for joining the Tech Deep Dive podcast. At Clarksys we believe tech should make your life better ,searching Google is a waste of time, and the right vendor is often one you haven’t heard of before. We can help you buy the right tech for your business, visit us at Clarksys.com to schedule an intro call.