HS055: Cloud Is Not The Only Future

A special guest today, Keith Townsend joins Johna and Greg to discuss the advantages and disadvantages of cloud computing compared to on-premises infrastructure. We start by clarifying the definition of cloud and emphasizing the significance of software-defined infrastructure and automation in on-premises data centers. Keith shares his perspective on hybrid infrastructure and making the case for a combination of on-premises and off-premises resources as the future of enterprise IT.

•  addressing challenges such as the transition to object storage,
• workload placement policies, and
• the scarcity of skilled professionals in cloud computing. As hosts,
• questioning how established vendors are responding to the cloud revolution and exploring the complexities of off-premises cloud solutions
• the benefits of edge computing.
• highlighting the considerations of using SAS, IAAS, and PAAS in workload placement policies.

Links

The CTO Advisor | Keith Townsend – https://thectoadvisor.com/

Link: Nemertes | Strategic Consulting and Research | IT Consulting – https://nemertes.com/

Transcript

Johna (00:00:00) – Welcome to heavy strategy where the where we welcome unanswered questions more than unquestioned answers. And today we’re going to really take on one of the canonical premises of our time, which is cloud is better than than on prem. All three of us have somewhat nuanced views. And by all three of us, I mean Greg, myself and our guest, Keith Townsend. Keith, you want to say hi to everybody?

Keith (00:00:25) – Hey, how’s it going? Packet pushers. Audience Good to see everyone again.

Johna (00:00:29) – So we are going to talk about cloud, whether it’s good, bad or indifferent. Keith has some interesting ideas. Let’s just let’s just. Let’s just qualify. This cloud exists on prem and off prem, right? Greg Could we not? Because even though that’s the case, number one, every on prem data center is not also private cloud. No. And number two, it just blurs the distinction we’re talking about. Yes. You need automation and orchestration for your on prem data center. We get it. People who get it are doing it right.

Transcript

Johna (00:01:01) – But for the purposes of this particular podcast, can we just specify that we mean when we say cloud, we mean off prem and otherwise it’s on prem and we hope you’re doing all the right things with orchestration and yeah, we hope you’re already got software defined infrastructure and automation and all. Oh, if you’re not, then probably you don’t want to be listening to this podcast. You need to go fix your on prem.

Keith (00:01:21) – And you know what to to. To be fair, I’ve talked to enough folks who said that they have a private cloud and it turns out that they just have a vSphere cluster that’s running vSphere 5.5. We’ll have something in here for you to for context, vSphere 5.5 is like 12 years old. So the it is there’s a there’s a lot of range. There’ll be something in here for you folks that have no automation that that are that have been told that you have to go to the cloud and the cloud is the solution. So you’ve invested in the private data center. This is kind of a warning to you.

Johna (00:01:59) – And I just want to back up what you said there, Keith. Back when we were doing primary research about cloud versus not cloud, we used to ask people, so are you doing private cloud? And oh, yeah, we are. And then you look at it and you’re like, But you’re not doing any management, you’re not doing any orchestration, you’re not doing any automation. How is that private cloud?

Keith (00:02:16) – The definitions have been just so wonky that I agree. Cloud State conversation off Prem Cloud.

Johna (00:02:25) – So Keith, maybe you can give us your perspective. I know just before we kicked off, I gave you my perspective of your perspective, but perhaps our audience can get your actual perspective.

Keith (00:02:35) – We’ve made this bet at the CTO advisor. We based our practice. We’ve based our content marketing strategy off this premise, that hybrid cloud or hybrid, I call it hybrid infrastructure to capture everyone into this group that you’re going to have on prem assets, off prem assets and something in between, let’s call it Colo.

Keith (00:03:01) – For the sake of this discussion, let’s call on prem both co-location, you know, equinoxes and cuts of the world on prem, that that is going to be the future of enterprise. It because it just makes sense. I think if you’ve been in it for more than ten, 15 years that you’ve seen these cycles and you understand that there is no concept in which we pay down all of our technical debt and this stuff goes down, this, this, this, this stuff goes around in cycles. My best example is talking through a use case in which, you know, we talked to someone organization that was born in 2012, 2013, born, born in the cloud. They built the app. They got a lot of value out of the app. Now the app is just in maintenance mode. They’re just in what not to say. What Greg likes to call rent extraction. There’s no longer a need to invest in the app, but you’re spending too much money on it. What do you do? Do you replatform it or do you bring bring it on on prem? That’s the premise is that there is no such thing as all in on public cloud.

Keith (00:04:21) – There is no such thing. It’s all in in private data center.

Johna (00:04:25) – Well, so three observations on that. The first one is I think it can be proven that you are you are empirically correct because the hyperscaler companies all have private cloud. So by definition, you know, the Googles and the Facebooks, they’re not using public cloud services. I mean, they are possibly to to supplement, but their core operations are on their own data centers. So that’s proof point number one. Observation number two, though, is that the emphasis here is on enterprise. When a company is small and lacks the resources, the trade off of more expensive cloud for the. Or the agility that they can get as they’re getting launched may make sense. And I think the third observation, and this is where I’ll hand back over to you, because I’d love clarity from you and from Greg. Your argument is fundamentally that cloud is just too darn expensive for a and steady state operational mode. Did I? Did I capture that correctly?

Keith (00:05:21) – So I wouldn’t say the cloud is too expensive for steady operational mode.

Keith (00:05:24) – There’s this IDC quote that I like to pull that says that 63% of the respondents said they spend more on public cloud than they anticipated on service. That could be bad in my case. In my experience, that’s mainly bad. But if you’re getting the value out of public cloud that you want it to get, get, you know, the before you got on me and Greg, we’re talking about people resource issues. CIOs claim that staffing is still their number one problem. And if you can buy public cloud to help mitigate some of that staffing issue, then paying too much for too much. I’m doing air quotes. If you if you’re paying too much for private cloud or for your infrastructure, but you’re putting your smart people on problems that actually drive the business, then you may be perfectly okay with that because it’s solving another resource constraint. So this is again, not an all or nothing. I’m not I don’t like these big, huge statements that say the cloud is too expensive. It is just another tool.

Keith (00:06:35) – Sometimes you need to use an expensive tool. Sometimes you know, the back of my screwdriver will serve perfectly fine as a hammer.

Johna (00:06:43) – It’s going to highlight something like that. I think that off prem cloud is always going to be there. And to a large extent, one of the reasons that we see so much hype about it is because the people who are involved in off prem building in a in someone else’s infrastructure like to talk about it because it’s new and it’s fresh and exciting. And if you’re someone that’s in the industry and you look around you, everybody’s talking about that, not talking about the old stuff anymore. And that’s because the old stuff is known. We know where we are. The innovation cycle has largely ended. I mean, we’ve just coming out of the end of a 15 year innovation cycle where the whole nature of on prem it has changed. We’ve got software defined infrastructure for on prem and then but at the same time we’ve got this cloud thing. You know, normally we wouldn’t have two sets of change cycles like that happening at the same time.

Johna (00:07:31) – Normally in it you can barely get one. Like if you look back and say, let’s say back in 1999 was the last real change cycle into the year 2000, and from then until 2012, 2014, nothing. The whole idea of development and applications and infrastructure was completely stagnant. The only thing that we saw year on year was an increased clock cycles in the CPU, slightly larger storage, slightly faster networking, blah, blah, blah. And then all of a sudden, somewhere in the early 2010, there was a step change in what the capabilities around physical hardware. We started to break down the walls of storage to move away from block and file to something else. You know, object oriented databases started to break down from structured query language to really diverse. We started to see the emergence of things like Redis. And now the latest one is vector databases that they’ve come out with a whole idea, and then the whole idea of outsourcing. We realized that outsourcing that stuff didn’t work. But what we did, if we went to a managed service by a company that actually worked for it, because I still regard off Prem Cloud as a managed service, right? It’s really not much more than that when you really you sort of pick it apart.

Johna (00:08:38) – Sometimes it’s even less because there’s not that much management, but carry on. The thing about off Prem Cloud is it only manages a single thing. So if you want to buy a place like a key value store or read a database, what they don’t do is integrate it for you The way only reason they can run these such large things is that they don’t actually do the integration. That’s still the client risk. So if you go to the cloud, you might have somebody who’s doing your database or doing S3 for you or doing the, you know, the net gateway so that you don’t have to do it yourself in an appliance or whatever. But they don’t integrate it. You still have to stitch all of these services together yourself, monitor them all, do the reconciliation. And a lot of people underestimate cloud reconciliation for invoicing and purchasing. What you’ve actually done is changed the problems that you had before, which were generally not accounting focused, not cost focused, not they were they were different with just another set of problems.

Johna (00:09:35) – And you most people come out the other side with exactly the same headcount, except now they spend most of their time picking apart an bill, trying to find out why they’re paying three times as much as for what they did on on prem.

Keith (00:09:46) – I would say the problem is even worse than that. So what we discovered is that, you know, we tried to make this move from block and file to object. And companies rose up to put block on top of object because we still need it block storage. So we end up with these ridiculous levels of abstraction to fit our applications to the mold of the infrastructure. And the point of infrastructure has always been to serve the application and we’ve kind of flipped this. So now we have the accounting problem, Now we have the finance operations, you know, fin ops model that doesn’t fit legacy infrastructures. So we have to figure out how to manage the our our. Apps that pays the bills with new apps and new initiatives with the same number of people, which, you know, we’ve abandoned some of the automation and all that.

Johna (00:10:52) – Some of that was was customers doing stupid things, like they would go to their block file and block storage provider and say, I want object storage. And they’d say, Well, we can do object storage, but it’s a completely separate thing. They go, No, no, no. We don’t want to have two things to manage. We just want one. Can’t you just put the object on top of our block and file or put block and file on top of our object? And because that’s a stupid thing to ask, if you understood the technology, you would understand that that’s not how it works. You either you have to run two storage clusters, right? And so in the end, going to the going to off premise made sense because they could give you both and you didn’t have to and you weren’t allowed to do stupid things because the clouds just say, here’s our block and file, here’s S3 You know, you here’s my object store, here’s your file and block if you want it.

Johna (00:11:32) – But it’s just an emulated and if you’re stupid enough to use file and block will okay right and customers were prevented from being stupid unless they actively chose to be stupid whereas the legacy brand vendors would let customers be stupid or try and flex or be flexible or adapt or customer focused. And in the end, they just built solutions that were failures. Remember cars of old like we look at hybrid cars which had a petrol engine as well as batteries or, you know, hydrogen cars and they’re all dead ends technologically. They should never have come up. Well, hybrid, hybrid things are usually dead ends. That’s just my observation. I’m old enough to remember hybrid frame relay, which made no sense whatsoever. Keith, sorry I cut you off. Go ahead.

Keith (00:12:10) – No, but the the driver for stupid things. The cloud providers are old enough now that they’re realizing, to Jonah’s point earlier, that, you know, they not only are private clouds, in a sense, they offer on prem solutions to their public offerings, and they are realizing that there’s not a one size fits all.

Keith (00:12:38) – I can’t put everybody in a, uh, object store model of consumption because foul and block are still useful to the point that NetApp and made a big deal out of OnTap, which is, you know, the oldest thing in the world on AWS. And I’ve been told by AWS folks it’s massively successful this idea that we can move our infrastructures and our applications to a model that conforms itself to one delivery methodology is a broken idea. Fundamentally, we in the enterprise, the large enterprise, cannot pay down technical debt fast enough to make a transition from one platform to another. We’ve done a ton of research on this. The you have to rebuild every application for the public cloud. There’s a shortage of developers. And if I’m if you’re asking me as a CIO, which one would I rather dedicate my developers to? Replatforming legacy applications that add no additional business value or creating new applications or fixing the bugs in existing applications? You know, I can predict where the the priorities land for 99% of those CIOs.

Johna (00:14:07) – Yeah.

Johna (00:14:07) – And so, yeah, I would agree with pretty much everything that’s been said. I want to sort of circle back on a couple of key points. First one is everything you’ve said, you’ve carefully predicated for large enterprises. If you’re a smallish shop and you are turning to cloud for speed and agility based on all the drivers, Keith, that you just talked about, that may be the way to go. You’re not necessarily going to build up an enormous developer staff. You just want to get people out there, get it up and running. One of the things we found very early on, we did a very detailed research study where we looked at the costs of all of a private cloud versus public cloud on different flavors of public cloud. And we found that if you are going to public cloud without doing it right and by doing it right, that means really thinking through which applications go where, what kind of storage you want, what your networking costs are going to be. The average cost is going to be about 26% higher than if you kept it on prem.

Johna (00:15:05) – Same thing, same exact solution. We’re not talking about transition costs or anything like that, just the day to day operational costs. It’s going to run 26% higher. And some folks actually a cluster of folks who are spending as much as 40% more just to move to cloud. And that really comes down to the doing it right. Which brings me to my second point. A lot of our clients do not do it right. And one of the very, very basic things you need to do is consider is have a workload placement policy that says for this application, with its storage requirements, for its networking requirements, for its latency, security, privacy. Requirements. This is where it should live. And if you kind of go through that decision tree that says Application X should live in house, application Y can be colo application, Z can be PaaS or IaaS. Then you have an approach that says we’re going to optimize where we put stuff. Very few organizations have that WPP because they haven’t thought about it.

Johna (00:16:04) – In fact, one client is breaking my heart. And Keith, Keith and Greg, you guys are going to hate this because. Or love it. Because it’s proof that clients do really stupid things. They literally have a CTO who said, I don’t care what and I don’t care why everything’s going to cloud from on prem no matter what. And they have a ton of technical debt, a ton of legacy stuff, and everybody’s goals have been narrowed down to get it to cloud. And if you asked them why there’s no answer, what’s the best? I can tell you? I can tell you a story I was once involved with. I was working with a major bank and they decided that they were going to outsource their entire IT team to IBM. Oh. Yep. That’s it. That’s it. That’s exactly. Yeah. I remember having a casual conversation somewhere in a corridor or and drinks after work and I said like this, you know, this isn’t going to work, don’t you? Or words to that effect.

Johna (00:16:55) – He went like, Yes, exactly. He said, But you don’t understand. What we’ve got now is broken beyond belief. So I’m going to go and give it all to IBM. They can take all my employees, all of my infrastructure. They’re going to manage it. And then for the next three years, we’re going to do that, get rid of all the stuff we don’t want, and then in three years time, we’re going to start bringing the rest back in-house and I can get rid of all the people I don’t want because they’re it’s a big company You can’t get rid of like 1000, 2000 employees. You said, I’m going to get rid of all the apps. I don’t want I’m going to leave them at IBM and their IBM’s and then I’m going to rebuild my infrastructure from the ground up. And that’s what they did. So this idea of outsourcing or off premise to reengineer, you know, a whole department isn’t new.

Greg (00:17:38) – We interrupt John and Greg to bring you this message. Too often it organizations don’t know what they don’t know.

Greg (00:17:45) – They may not have the proper insight into their environments because they’re lacking the right tools. At Nimrods, we’ve been helping I.T. organizations assess their environments for over 20 years. We have a proven methodology and detailed templates. We’ve helped clients ranging from start ups to Fortune 50 corporations, assess their IT environments and fortify them to support future endeavors. We’d love the opportunity to hear about your organization and determine together if we can help. Visit Nimrods. Click on the Contact Us form and we’ll set up a call. N As in Nimble M as in Mermaid e r t e. s.com. Let’s pick up with Greg and Jonah to.

Keith (00:18:33) – Support the thought processes. You know, I kind of see folks in three different phases. This legacy approach and where this cloud movement is, is is old enough that we can call it a legacy approach that we’re all in on public cloud. People who have started that journey and then people who in the third bucket have realized that it’s a broken model, that you they can’t keep pace with technological change.

Keith (00:19:06) – And Jonah, to further your color around application placement, workload placement, understanding where the workload should go, there is a opportunity to also decide and to selectively when to invest. Tying your point and Greg’s point together, once you’ve lost the muscle of running it and infrastructure is extremely difficult to regain that muscle memory. One The talent has moved on. Storage administrators have now become cloud administrators. Once we fired those people from their previous functions, they’re extremely resilient. They’ve gone on and develop new skill. They can now deploy a wide area network with a few button clicks as opposed to spending months and procurement months contracting with service groups to deploy and install and manage network gear. They can now do that in a few clicks. Who wants to go back to managing network devices in this legacy mode? So they find that kind of resetting their infrastructure to to manage it difficult.

Johna (00:20:24) – If you you don’t go back if you bring something back on prem, you don’t go back to managing it like you did before. You come back with a software defined operational model, you get management platforms, you come back.

Johna (00:20:36) – Keith just said, Yeah.

Keith (00:20:37) – And that is extremely difficult skill set to find no anecdotal experience. I interviewed for the manager of basically the public, I mean the private cloud for Comcast back, you know, maybe 8 or 9 years ago, that team that interviewed me, the wholesale team, you look at their LinkedIn, this is not anything that’s that’s private information. Look at that wholesale team. A few months at that interview was that Walmart deploying their private cloud. The whole team that skill set is in super high demand and extremely hard to find. The second point is that forces you into a situation. You have to optimize in place where it makes sense. Sometimes you optimize in the private cloud or the private data center. Sometimes you optimize in a public cloud. A born in the cloud, backup company, data protection company. I met with them last year during VMware Explorer. Talk to me about how they reduced their storage costs by two thirds in AWS because they looked at their storage patterns and they realize they almost never recover.

Keith (00:21:49) – So why not just go to something like a glacier, which is super expensive to recover from, but they only have like one restore request a month. They optimize the application in place versus going to, you know, the Dropbox scenario where they build a private cloud. There’s opportunities to do that in both in both environments.

Johna (00:22:13) – Let me ask you a weird question. Do you think that we’ve seen enough change and modernization from the brand vendors that already exist? So when I look at my heritage, it say from companies like Dell, EMC or Cisco or HP, do you think the innovation cycle that they’ve brought to change their their response to cloud like we are, what, 10 to 15 years into off prem cloud and only now are we starting to see Hyperconverged turn into managed services. So we have HP, Greenlake, Dell Apex, which is very early, very nascent. Is that the right response? Do you think they’re actually finally going to change or what?

Keith (00:22:52) – You know, it’s slow. There’s a lot of things to like in theory about the.

Keith (00:22:59) – About Hpe’s Greenlake. I like that they’ve taken the integration, the partnership. Integration and put it on single paper. So as enterprise it large enterprises buy stuff. I can get my SAP licenses, I can get my Equinix hosting cost, I can get my hardware. I can get my VMware all delivered to me. OpEx or CapEx under one piece of paper. In theory, it’s one engineered solution in theory, and I just buy that all from HPE Dell Apex a few years behind in my opinion. But I think HP particularly is answering the bell from a marketing perspective and a market texture perspective Now where the stuff is physically, you know, the they announced their private cloud at HPE Discover. I have one, I bought an HP Hyperconverged It’s not hyper converged, but let’s just call a hyper converged solution. Over a year ago, I invested in that in my own data center. So I have access to their private cloud solution. It is nascent, it is the most basic of things. I still gets me maybe 60 to 70% there and I have to put staff and smarts on top of that to build a true private cloud.

Johna (00:24:20) – Yeah. And I kind of feel like as a customer, we’ve been let down by our by the brand vendors who’ve done so little to change. Am I, am I that’s, that’s, that’s, that’s basic innovator’s dilemma, right? The established players who made their mark in the era of on prem are going to deliberately be slow because that they can’t afford to be nimble and innovative. So I would agree with your point. Let’s let’s put a bow on that and say, yes, I agree that the vendors are doing what vendors do and dragging down the pace of innovation because they have to for their own for their own security. I want to also, since we’re coming to the end of this conversation, I want to just toss out there that we’ve all sort of we started this by saying there was this wave of innovation that we’re kind of at the end of the move to to off Prem Cloud happened ten, 15 years ago. I want to highlight that even though the pendulum swings, it’s already gearing up to swing back in a certain sense with the advent of edge computing.

Johna (00:25:22) – And Keith, when you were talking about how Amazon is thinking about new ways to do things, I love the fact that Amazon has thought through what it would take to make edge computing actually work and is selling edge computing solutions, which are starting to become more and more interesting for folks doing things like Iot, where latency is a huge issue. So coming back to that workload placement policy that we’re always recommending, one of the things we encourage people to think think about is, okay, your choices are not just on prem. Colo Public Cloud and various flavors of public cloud, but also edge computing when and where it makes sense. And by edge computing, we’re not talking about old school. Put a bunch of servers under somebody’s desk. We’re talking about something that is standardized, that is consistent, that can be installed or or gotten rid of very, very quickly. I mean, literally, you’re talking about a data center in a shipping container with all of its self-contained power and cooling built in. So you know exactly what you’re getting so that you don’t have to worry about all the different, you know, all the different personalization and customizations that need to get made.

Johna (00:26:27) – And that model, I think, is just beginning and it’s going to be very, very interesting.

Keith (00:26:32) – John, I think you’re hitting the main point of my, you know, pulpit message. If you have more than ten years left in your IT career, some of us don’t. Some of us are closer to retire retirement and others. This is where you roll up your sleeves and say this is where the opportunity is at between edge on prem, off prem. The management challenge. The integration challenge is massive. There is a lot of opportunity from vendors creating solutions to these problems. It’s not going to be we’re not going to buy a single solution from a single vendor. And this is where it leaders and practitioners have an opportunity to have really great change. This is really complex. Just managing on prem and off prem cloud is difficult enough. Throwing in this third method of delivery edge into the mix where you now have to deal with OT and manufacturing and field restaurants and all of that. This this true business facing set of it and functions.

Keith (00:27:45) – This is this is my my head hurts thinking and.

Johna (00:27:48) – What you’re effectively.

Keith (00:27:49) – Saying is support.

Johna (00:27:50) – That so what you’re saying is that off prem cloud made it worse. Because, well, now we’ve got a whole bunch of proprietary external, uncontrolled vertical services which the client has to take responsibility for integrating in addition to all the stuff that they do on prem. So cloud data prove anything that made it worse. No, that’s not true. Greg, you’re a curmudgeon. Yeah, I know.

Keith (00:28:12) – I was about to say the same thing.

Johna (00:28:14) – It’s improved. It improved a lot of things, but at a cost of additional complexity. Yeah. And I think, you know, I’ll take put words in Keith’s mouth, but I think we all can agree that, you know, there are a lot of benefits to doing cloud for smaller companies, even for large enterprises. The problem is that the blind cloud will solve all my problems. Has never made any sense, ever.

Keith (00:28:37) – And to, you know, I don’t know if it was Andy Jassy, who as of last year or even year before, said that they’re still building these services with the objective that everyone will eventually go all cloud.

Keith (00:28:51) – That is not helping.

Johna (00:28:53) – Yeah, that doesn’t help. Yeah. And I think that’s I think that we can all agree on that. If you’re listening to this and your organization has decided to go all public cloud, then one of two things are happening. One is they don’t know what they’re doing and hasn’t thought through this. And the other one is they’re doing what Keith said and just basically using it as an excuse to get rid of all y’all, which in which case either one. Yeah, you probably want to be putting your resume together because either way, it’s not going to be fun. When the CEO of company says you should be all in on company and the products that we sell all in on Cisco, all in, on IBM, all in, on HP, all in on AWS. What do you expect him to say? It is self-serving and facile to the extremist degree. And what off prem clouds have managed to do, in my opinion is they’ve managed to create craft and amazing marketing message around off prem cloud as a lifestyle choice.

Johna (00:29:43) – I mean a professional lifestyle. If only you came to use all of our proprietary products, only our proprietary products, then somehow your professional life will be vastly improved. That is obviously patently false, but that is a marketing message that has been created and fabricated by Azure and Google and Oracle and especially AWS, right? Because they’ve got one of everything in and that is. But it is a compelling message. If you’re an IT executive who’s floundering with insufficient funding and insufficient headcount or you’re just incompetent, you lack skills or lack capabilities or the people above, you just don’t value it. That message becomes compelling because you can just sit there and go, That’s not my phone at all. That’s AWS problem. It’s not me, right? It’s just a very lazy way of working.

Keith (00:30:32) – So and it’s playing it safe. I won’t get fired if I say I’m all in on cloud. I’ve seen some massive mistakes. I’ve seen I’ve gotten to the argument with people way smarter than me. On the business side, you should never, ever, ever put legacy SAP or sap Hana in an infrastructure service platform, ever.

Keith (00:30:55) – I’m going to definitively say that I don’t like making these sweeping statements, but SAP is one of the worst applications to put in the private in the public cloud. It is a monolith. You can’t it just doesn’t fit the public cloud model. I’ve seen.

Johna (00:31:11) – Way no amount of no amount of wishing or praying is going to make it cloud.

Keith (00:31:17) – Ready. I’ve seen CIOs spend millions of dollars a year more on SAP infrastructure than when it was on private cloud, and I forget my point, but that’s just a example of what of the nonsense I’m seeing when it when we’re all in in public cloud. That’s my point. All in in public cloud just doesn’t make sense for some applications.

Johna (00:31:39) – You can be all in on SAS. So SAS is not a public cloud is actually a public cloud and you and you can be all in on it, but at some point you may be spending too much. Most of our clients have found that for core applications like payroll and HR abso freaking lutely all in in the sense that you only ever buy off the shelf SaaS applications again, only works if you’re a very small company.

Johna (00:32:06) – No real requirements for customers. And that’s why and that’s another reason why SAS is so expensive is because you’re usually hosted on an off prem cloud and they’re paying way over the odds for the infrastructure and they pass there, but they don’t care because they just passed that on to you as a customer. Exactly, exactly. And there are good reasons to go SAS, but, you know, and we do consider SaaS, IaaS and PaaS all as as components of public cloud. That’s all part of your workload placement policy which says, okay, which things do we just buy off the shelf, assess which things do we build ourselves? And if we build it ourselves, where do we put it? Do we use I asked, Do we use pass? Do we use colo? Do we use edge? Do we use on prem? That’s how you should think about things. All right. We’ve got to wrap up the discussion here. I think we’ve hit the point at where the inflection starts. Keith, where can people find you on the Internet?

Keith (00:32:51) – I am still on I don’t know if they call it Twitter.

Keith (00:32:54) – Or whatever it is, but I’m still there.

Johna (00:32:56) – Formerly known as. So you can say the exe, the platform formerly known as Twitter.

Keith (00:33:01) – So I’m there unless name changes again at CTO advisor and the website is the CTO advisor.com for floor slash projects. If you want to see kind of the cool things we’ve done to validate some of these messages from the vendors.

Johna (00:33:17) – Joanna what about you from CEOs at numbers in like Nancy M like Mary ertz.com join us on the community there’s a tab up top and you can continue the conversation with folks like Keith and Greg and myself on our community. And thanks very much for listening to Heavy Strategy. It’s been a pleasure to bring this to you. If you’re a sponsor and you would like to put some ads against this or get your message out and amongst this, please do contact us over at sponsors at Packet Pushers Dotnet. I’m Greg Ferro. You can find me on Twitter as at Theory of Mind, and I’m also hanging around on LinkedIn. I’ve started to discover the dark side that is LinkedIn and try and work out how to make it work.

Johna (00:33:55) – And remember that heavy strategy will be back in a couple of weeks. Thanks so much for listening to us. We’ll see you then.