Cloud Economics Feedback Loop [Cloud 2030 Feb 4 2021]

So listen carefully, there’s a lot, there’s a lot going on, we’re gonna keep talking about this topic, this bridges into more conversations, including like open source and, and standard stack, that will, we will come back to enjoy the conversation, this was a great one, as always, we’ve been talking about this inflection point of moving more control back to the consumer.

01:02 - So in the last in the last couple of weeks, and especially since the summit, one of the things that, you know, we’ve sort of identified is that right now, the cloud companies have a lot of control over what gets built and how it gets built, and what those models are. And then we added that op x thinking is driving people more dependency in that model, where you’re saying, Hey, we’re, you know, our innovation relies on the fact that I can rent infrastructure and not own it.

01:31 - And what what I think we’ve seen more broadly is that that has translated to the cloud companies have a like having a lot of power in the customer relationship.

01:41 - And question to me for 2030 is, can that change? Will that change? And if so, what is the mechanisms that allow that to change? Right, is it? Is it something like having more edge technologies coming? Is it some other component? And so as a background, that sort of, since we’re, we’re since we’re episodic, that’s your recap, now I play the theme music, and you get it, we can go to the actual episode, right? From my perspective, I mean, I thought, I’ve liked this subject matter for a while now, and you and I’ve probably had several conversations about it over the course of the last two years.

The basic logic is that if you’re becoming digitized company more of a technology company with products, rather than a product company that occasionally uses technology, then the basic assumption is that the amount of technology you’ll be utilizing to make your company effective will grow, because it’ll expand in new areas and create new opportunities for you. And you’ll begin to look more like what PayPal is to a bank.

Or for that matter, even what a most modern, some of the most modern trading companies in the financial sector are, where they spend between 30 and 45% of their revenue on technology.

03:05 - So when you think about the average manufacturing company, do they need to spend 30 or 40%? Probably not. But if they spend one and a half to three, today, it’s not far fetched to think they might spend four to six after digital transformation. And so if all we did was add 40, to 100%, growth to the average IT organization over the course of the next five to 10 years, as a result of digital transformation, and we only applied it to half the companies in the world, we would still have an incredible increase in the amount of technology adoption.

What the other part of that story is that if my basic assumption in the comment that Rob just reminded us of is that if technology becomes more of what drives your business, then it’s more likely to that that technology is no longer an afterthought. It’s not just plumbing anymore. If it’s not just plumbing, and it’s key to your ability to do business and be successful, then you would think there’s a higher likelihood that some or all of that technology will stay within your grubby little hands.

04:19 - And again, is that a perfect analogy? No, it’s not because even some of the bigger internet oriented or digitally oriented companies today don’t own all of their own technology or even the majority of it, but many of them do. So again, going back to the percentages, the law of numbers, we look at 500,000 businesses around the world, which is a fairly easy number to pick because that’s like the number of customers that Salesforce has or something you know, that’s a pretty easy number.

And you say that they’re going to you know, add 50% to their it and then 25% of them will end up deciding to keep some or all or at least more of their it than before.

04:59 - It seems to open up a significant number of doors to, to how we adopt technology, how we sell technology, and how we consider some level of independence from others dictating our path forward relative to technology use and its cost. Right? So I mean, we could go on and on. I mean, I could spend the whole hour let’s rambling on this, but I’ll stop there.

05:23 - And you can explain. I mean, what what you’re and I love that I love the bottoms up analysis of the market, right? If there’s so much more potential money, right capital flowing into this.

05:43 - You know, it’s it, that’s I like the that approach, I think with edge, it’s going to be even more compelling.

05:50 - perspective, because I think real down for edges is going to dwarf what we saw for cloud plus, at least, at least initially, sorry, Rob to cut in. But at least initially, the variety of demand.

06:03 - And the the potential inability to effectively translate that to an open ready, perfect solution by an existing major cloud provider opens up a significant door for at least people to taste test in that environment. And then some, again, some of them will end up saying, it isn’t worth it, I’m not good at this, I’ll let somebody else do it. Even if it’s a, even if it’s a, you know, lesser of two evils type of decision, but many people will decide that this is an area where I can differentiate my business this and because I’m differentiating my business, I should do my best to leverage this technology to the best possible way, and maximize my return on investment.

06:47 - Go ahead, go ahead, click, the inner economist in me looks at this from a standpoint of scale effects. And so, you know, technology life cycles, tend to be kind of compressed, versus, you know, the traditional, you know, brick and mortar. So you’ve got this, you know, scarce power, that creates a whole lot of wealth, because a few people have it, and they deploy it, and they get incredible business advantages, and they pay a lot for it, then over time, other, you know, capabilities emerge, and then this other stuff becomes a lot cheaper.

Right. So it’s like the whole calculator issue, right? That those $300 calculators that are now free that sit on your, you know, computer. So I think we will have waves where a lot of wealth get generated, versus every new innovation, every new transformation, a lot of that initial costs are probably gonna be technical debt related, right, because you’ve got older systems that were kind of doing some of this. And then they’re tied into other systems that are business critical, and you can’t turn them off because the business has to run.

So you’ve got this juggling act, that to me slows digital transformation, because it makes it more complicated, more expensive, than, you know, just starting over for like a new company. So I think you’re going to see different ripple effects. One, smaller companies that are new and starting up are going to have lower adoption costs, you know, to trance where they won’t really transform, there’ll be part of the digital revolution, and then you can have these big enterprises that will have to spend more even to get less initially.

And then hopefully, they get caught up. And that’s probably this edge dynamic mark that you’re talking about where stuff they have direct control over, versus the things that are meted out to them by the different, you know, as a service providers.

08:37 - But there’s a downward pressure with adoption as as because of the scale effects of technology, right, especially around software and services, where the incremental cost of each additional person creates a downward pricing pressure, so things can get cheaper over time, and then you have new competitors and new innovations that have more value, and then so people have to cut their price.

09:00 - Further, I think this is a new era we’re entering that’s a lot different than when, you know, Salesforce and Oracle had huge amounts of advantages, because of the their technology and how deeply integrated it it became in the organization. And I think, you know, there’s going to be other types of SAS that are easier to adopt that do, you know, specific things, and there’ll be there’ll be cheaper and faster, in my opinion.

09:32 - I, the I agree with you about commoditization of what’s going on with cloud AI, and this is This to me is part of the the the not to untangle.

09:46 - Because it, it should be that this is why I was so concerned about the apex changed innovation cycles. It should be that the things that the cloud providers are doing become commoditized And people people replicate what what they’re doing because right, nobody needs all of the services that Amazon has been adding, although with the long tail, probably most of their customer, you know, a lot of their customers do need some of those services.

10:21 - But it’s, it seemed to me that it should be that, you know, eventually infrastructure and software become, you know, sort of a unit that you can use. On Tuesday, we were talking about the LAMP stack. And now the LAMP stack created this, you just huge spreading of, Oh, I know how to build a website, I have the standard architecture stack, I know what to do. And then that led this huge exposure explosion of Linux users.

10:52 - And I know I don’t I feel like that could happen in the markets that we’re talking about, but right now, I don’t see I don’t see where the wedge would come in.

11:04 - From a tech perspective, like we’re talking I mean, edge there, it doesn’t look like there’s going to be an edge stack emerge. I mean, I’m looking at mark in this case.

11:15 - I don’t, I don’t know that there’s going to be an edge stack. Right. And I don’t know that. In fact, what I do know is that nobody knows that right now. Yeah, um, no matter what anybody says nobody knows that. I think, again, this is probably a conversation that Rob and I have had, and maybe some other folks that are on the call today. But one of the interesting aspects of edge is that, you know, CloudFlare has some great solutions.

11:45 - And they’ve designed around a specific delivery and support model for what can run in their environment.

11:52 - terabyte has a great, easy to use easy to adopt cloud, like solution for the edge. But it’s not going to be what will be used for every analytics workload, heavy analytics workload where you’re looking at high density of compute and, and storage as close as possible to that compute is going to be a unique workload and probably unique design requirements, probably not, for the most part, just lambda or something like that.

So between the ability to use containers, microservices, wisdom, even even VMs, in a distributed architecture for just infrastructure, across a wide range of locations, there is no one way to deploy and what will happen first, at least in my opinion, what will happen first is the things that make people money right now. And some of those will take hold, like in everything we’ve ever done, some of those will take hold and grow and become more important.

Some of them will will adapt, and, or, or die. And that but the idea that the the I think this is one of the biggest questions is that we don’t know enough about the environment today. For anyone with confidence to be able to go to an enterprise and say, Look, if you use wavelength, we’ll eventually have you everywhere you need to go, and you’ll have all of the solution you need. Okay, you’ll have all the solution I need for 20% of my workload at a price that’s 40% higher than if I did it myself.

How is that a positive solution? I’m not saying there won’t be solutions that are worked for where the location is a perfect fit. And speed to value is more important than designing your own environment. It’ll probably work for many workloads. But there is there is no one right now who can say that, whether it’s anthos or or who always private stack, or Ali clouds, Ali cloud plans to do 5 million edge locations in China alone.

14:09 - And yet, they’re not going to be the only one. Right, so I just I don’t know what the right answer your guys. I mean, these this is the whole point of these conversations is that I just see that there’s an enormous amount of opportunity here. And one of the things that that I’m trying to drive with some of the folks that are willing to partner with tiny ledge Ivana is how do we call market how do we help educate the market and even getting out there and making determinations on what can work for what solution and why is it worth approaching now versus waiting two years? Step back.

14:45 - Just one of them how I evaluate markets. Um, when I’m valuing markets I’m thinking of in terms of how many players are going to be can be dominant. So basically, can there be? Is, I mean oligopoly or not? So is it is it gonna be just dominated by AWS or a couple companies is gonna be dominated by just, it’s just to be a cloud market, there was a big cloud market or an edge market? Is it going to be big time by one industry? Or it’s gonna be defined by multiple industries? Is it going to be defined as? Are you gonna? Are there gonna be specialists in terms are focused on different use cases? Are there any specialists that focus on? And are those use cases could be based off of geography? I mean, basically, this is definitely competition based off geography.

15:49 - So basically, data locality is important. So in this in this case, is it? Is it gonna be based off of industry in terms of industry, vertical healthcare specialization? So enough? Is it based off of company size? So are there other gonna be industries that focus on SMEs on enterprise use cases? are large enterprises can handle these issues themselves? I mean, these are all issues that matter a lot in terms of how the competition is going to actually happen.

In the case of like cloud, are it? Are they going to run you or create their own data centers and buy their own infrastructure? Are they going to use a third party? These are all relevant in terms of how the whole market? Go ahead, Mike? And then I have a question, too. I think I think Lawrence Lawrence is on to something that’s gonna be really interesting, as we as we look, not only not only near term, but long term, which is, we’ve looked at, you know, we’ve seen the cloud, we’ve seen the cloud build up till now.

And you know, and we’ve and we just assume that it’s going to look a lot like the original, you know, the overall team on the one that we already know. And what I’m seeing, what I’m starting to see play out is closer to what Lawrence is talking about, which is, you’re going to have Amazon and Microsoft at that very, very top level, right? Sure, Google, Google, you know, they’re going to struggle to compete, they’re going to struggle to compete at that level.

But they don’t have the, the, you know, that entrenched enterprise base that though that those two have, right, they don’t have the same the same heft. So they straddle this, this sort of next level, which, which, you know, Google’s sort of in its own special case, but then you got a next level, sort of the purgatory of the cloud, which is IBM, Oracle, HP, SAP, all the existing players that don’t have anywhere near the haften aren’t going to catch up to Amazon and Microsoft, right, and don’t have the ability, agility to compete against the incumbents.

So their slice of the market is going to be much, much smaller, but like Lauren says, where they’re strong, is in some of these verticals, right? They own a lot of these industries, that an Amazon or Microsoft, it’s harder for them to break to break into there.

18:33 - And then you’ve got the alternative, you know, that alternative level, that, you know, with linode, do ob H, which hates what’s getting put into this category, which is why I do it, it’s just fun.

18:47 - That’s her rack space, you know, equinix, unpack it a little bit and upcloud.

18:53 - But they’re, you know, they’re never going to compete against any of them. nor should they, but they look really, really attractive when it comes down to things like multi cloud or right do and I don’t mean multi cloud shifting data all over the place. But when you look at edge, okay, we’re talking about cost, I don’t need to email an Amazon, you know, an Amazon level to host a little bit, you know, the host object data, I can I can actually look for a lower cost provider there.

That and then it gets into things like trust. I think this gets really interesting on what the future of our you know, of this of this industry looks like.

19:34 - And I think, Greg, your market are also right.

19:41 - Yeah, I mean, basically, in tight one on the internet. Yeah. In terms of markets, you always want to be number one or number two in your market. So you basically define the market down until until you’re number one number two. So it’s a Yeah, there’s no charter market. You Matt, Matt and Matt AC freight line. Right? And we’re in, we’re in this part right now.

20:07 - Bigger, bigger pie. Right? Would you would be needing you’d be wanting to identify the biggest slice of a common pain or need, you know, that’s unmet by the current mix that might just, you know, laid out, you would you would want, you would look, what is it the underserved part that could also be linked by API’s or whatever into other aspects, you know, the new stack, right, the new edge stack? And so what’s missing, to bring that about to make it easy, but unfortunately, we don’t have a Linux type player that I know of.

I mean, I guess you could kind of say Google, to some extent is, you know, their model. But you know, with the LAMP stack, you had Linux, right. And it was distributed and available. People could do a lot of things with it with a patchy, but I don’t know, do you have any type of altruistic player that’s in a slice that’s broadly underserved where that could get enough traction and, and growth? compares? There’s an interesting open source question in the in the middle of this that I would, I would pause for a second.

21:22 - Because I think there’s a different thing to dig into first, before we go the analogy of the LAMP stack, which I like.

21:30 - But there was something that Lawrence was saying, and, Mike, you you came back to this a little bit, which has me when we defined edge, we conflate two different types of edge in this in this conversation. And I think this has been an ongoing challenge. There’s an edge commonality question that we’re talking about where you have a hospital, like all the hospitals want to have their own infrastructure and Someone’s coming.

21:59 - They want to have like, we want to have a way to have an edge facility, that is a self contained edge facility, or edge self contained site, what we probably would have called private cloud back in the day, and, and have that repeated over and over and over again. So there’s this idea that we’re going to localized infrastructure with a repeatable pattern so that it operates consistently as a market.

22:25 - Something that VMware start did but didn’t do well enough. That’s right. And then we have this other idea of edge ubiquity.

22:33 - Where we’re saying I actually have an application that is in edge sites distributed, like SDN is today where the value of the edge infrastructure is the fact that I have edge that one application in a whole bunch of different locations, acting as a single location. SDN is the perfect example for that. But there’s, you know, if you ended up with, you know, more things that have very tight near when we get real AR VR integrations are real voice and camera stuff, and, you know, autonomous cars and things like that the application is going to be everywhere, but it’s going to be hyper local, from a from an operations person, this will be a ubiquitous app, ubiquitous vendor.

23:21 - And I think that we’re still, we’re still struggling with both of those things. But I’m thinking that they’re different. They’re actually different problems. Once one’s a bit once a business problem, the other one’s going to be a deployment management problem.

23:33 - And then in the deployment and management problem. Yeah, that’s sort of what you it’s pretty much right in your in your wheelhouse what you do, but that it gets a desk that’s going to get Messier, before it gets cleaner.

23:46 - I there. The funny thing is, I’m not sure that they’re really that that, you know, it occurs to me that we can we can fight them in conversations. And I actually think that the commonality problem is that we have to solve them both. And maybe maybe we can solve them together. I mean, I know RackN would love to, but the It, it, it strikes me when I step back at it, right, the lock in from a cloud, I still wonder right? locking sometimes I think is real.

And sometimes I think it’s not real, and it’s it’s real, it’s not, it’s not real until something’s happening in that your provider of choice that causes you to, you know, regret that decision. Like it’s not real until it is like a lot, you know, for elastic last week or, you know, Netflix competing with prime.

24:40 - Right? It wasn’t a problem until it was. But then we were not we’re not we’re not going to drift to standards, no matter what we do. And then that means that we’re going to be dealing with a heterogeneous, heterogeneous, heterogeneous environment and which Drink brings us back to lack lock in provides a benefit for that.

25:06 - wage that hampers innovation. Yeah, I always I had I had this like crystal clear idea in my head about this ubiquity and commonality, and then you have the power to get back into the consumers hands, which is where we started, we have to have a way that, you know, you can worry about what your your layer is on top of this, and, and then not have to be locked into a provider on that.

25:34 - What gets back to what Lawrence and I were were arguing about is people are going to end You know, people are going to end up where, you know, at the, at the strata of the ecosystem with a provider that they you know, that fits their business the best. For some of them, it’s going to be we need all the breadth of an Amazon all the security of an Amazon, and I’m okay being locked in maybe not technically locked in, but I put all my chips in and it’s made my bet and I’m too costly to, to fold at this point, right, some are going to be, you know, some are going to say, I have to be at an IBM I need all of their, you know, I need all of them, they might not say that, okay, maybe I’ll be at a sap, right, I need to be in an IBM because I need all of the that industry knowledge around retail to compete, because I also also because I’m not going to go to Amazon, because I don’t feel like feeding the hand that’s killing me at the same time, right.

So there’s a huge business trust challenge here.

26:35 - And then there are going to be the, you know, others that are gonna be like, I don’t care where my stuff is, I just want it cheap and easy and open and be able to just run my stuff, it goes to mark talk, or Greg talked about it earlier.

26:50 - of, you know, the companies that are, you know, that are, I don’t know, if I need to be all in don’t know where I need to be, I don’t need, you know, the construction company, I don’t need to spend $30 million on this, I need to spend $300,000 on this, to get it still seems like it’s a costly, complex problem. And that the infrastructure as a service problem got solved with a ton of cash that came in. And I think Rob and I were pinging on this a little bit before the call, but it came from other places, you know, and so it gave them this sheer amount of investment.

And even when you look at the lower end, you know, into the oracles of the world right there, there, they made investments to probably to protect revenue more than to become a new, you know, cloud player, but they had a huge amount of, you know, proprietary revenue, you know, they they had they had a source, but when you go out to the edge, who is it? And I thought if one person, one company could do this, I thought it would be potentially VMware, you know, because they weren’t there.

And they were all in, you know, between virtual desktops, desktops and virtualization, maybe there’s a critical mass that could be created by an ecosystem. But they went for right to the cloud to and didn’t do so well. So I do think you need to have some type of lock in for this to work, otherwise, you’ll never get the marketing momentum to invest in the product, the security, you know, all of those things that that need to scale.

They don’t scale by themselves. Sorry, Rocky go? Well, I was just gonna point out that, something we might want to look at a little bit more closely is this application we’re using right now, suddenly, zoom is the de facto default.

28:41 - And it took a crisis and then being in the right place at the right time. But at some point, some large company is most likely going to buy them, or they’re going to actually be able to fend them off. But the whole way, video conferencing suddenly became ubiquitous. So we’re talking, Mike’s ubiquity again, and we’re talking the technology of just keep zoom in mind, and what alternatives were there before and what alternatives might be there in the future and how this will continue to play out.

29:30 - That’s rocky Whoa, I, I didn’t so bases of communication, a chat protocol called matrix that’s being used in a chat app called riot, Pittsburgh to signal and it’s very popular to use is especially in open source community. And basically, yeah, shut down for two days by Google last week.

29:56 - And that is is a big deal because it’s the same thing.

30:02 - Same niches will take Google been shut down like telegram or everything like this. So it’s the same thing with when it bothers you.

30:11 - Why are they got shut down from a from a from a speech safety issue? And it’s they got put back up basically, they didn’t somewhere like Google didn’t know that they actually had an enforceable monitoring system. And they were doing everything right. But they just showed that Google is able to shut it down. Matrix. I mentioned this yesterday have a dis, we’re looking at looking at something that didn’t fall in for a while.

30:47 - And it’s very popular in the in the fire stem community, but them being this Belgium annual open source conference. It is laurencia. Ces, the matrix, that is the underpinning of the element. Yes.

31:07 - Because that’s one of the key distributed web technologies. Yeah. And it’s been used kind of as a as an alternative to slack and for both for individual messaging, but also for group messaging. And it, it comes with a good deal of baggage.

31:30 - Some of it not that all that savory. So that’s probably why I got that. I didn’t hear that it had been, it had been kind of known.

31:41 - So I didn’t know all that bad. I didn’t know about all the baggage.

31:47 - I had all these things that baggage. I Gosh, I but but that’s what are you supposed to do? Trying to have a an alternative to zoom? Well, I mean, it’s it’s not really that sticky. I mean, they’re good. But there are real alternatives. And having you know, we’ve we’re starting to see some ones that are coming in that have, you know, better dynamics for even this type of conversation, then then what zoom has zoom? Has he ever seen that team fill HQ? which one to go look at team flow HQ? Yeah.

It’s pretty awesome. So you go in, you design your office building. And whenever you’re in a build in a room with people, you can hear them and your video turns on, you can shed all forms of darks and stuff. It’s pretty cool screen.

32:51 - Gather does that too. I think I’m sure there’ll be a bunch to Canvas one I’ve heard I’ve come to cam. There’s another one.

32:59 - There’s there’s a couple of them coming. That seemed much more natural from like, for this lady perfectly reasonable for this to be a room in my office, my virtual office that people could come to.

33:11 - And, you know, one of the things that we’re looking at speaking of chat apps, like slack is one very expensive shot. I mean, it really is, yeah, there’s that they they’ve singularly failed to innovate around this whole thing, right. Yeah. And there’s so many ways that we will.

33:32 - Well, we want to integrate, right, we want to integrate chat, we want to integrate video, we want to you know, screen sharing documents, the works. And add guests. I mean, yeah, I think this drives me nuts. There’s guests, it’s super hard to have guests.

33:47 - Right. And when you have guests wandering around your room, they should be able to hear stuff, and whatever else. Yeah. Yeah, we’re complaining this for zoom, it cracked the nut. For years. Zoom was the best alternative to everything else on the market. That’s right. Yeah. I mean, remember when we complain about Skype? And WebEx and well, this is this is the how, with with innovation right there, they’re the the idea of reducing the barriers to getting delivery is a critical, critical, critical thing.

34:29 - That’s why I get nervous around this whole idea that Amazon can be displaced. I mean, frankly, this is because it’s so easy. Just try your billion row database. Just try it.

34:43 - You don’t have to I mean, that we don’t have to move a database. I mean, you can connect to it until it becomes a latency problem.

34:52 - You’re right. So what they’ll do is move up stack to reshare immersive service interfaces, which are totally sticky.

35:02 - And once you make the call, you’re there forever. It’s going to be that way. I mean, once you start using redshift, and relying on the features of redshift, you are there forever. That’s it. Even the name scares the shit out of me. And, and, and, and the price you end up paying for that is just.

35:26 - Yeah. Well, well, the prices I think there will be I mean, if they ever get ready to get out of control, then government won’t get called in or something. But yeah, you’re there. Yeah. No. And I think Nikki Yeah, so not doing that that is insisting on doing your own thing. You’re probably screwed up.

35:48 - You won’t scale as well. And it will be more expensive in the long run.

35:54 - Bradley’s may maybe this that that the question, the thing that I think is an industry falling on its own sword problem, Simon, is the assumption that you can’t do it as well. And you won’t scale as well. That, to me is where it’s so it’s so broken, there is no domain there is no or get runs databases that can do a billion row database. Yeah. But why it why I mean, alright, so answer me this. Why is it that running a billion row database requires PhDs and operations? I mean, does that song that’s just inherent in the fact that this is the tech or is it inherent that we’re not spending the time operationalized? So in my mind, there are a couple of things going on, which we haven’t mentioned.

One is there is a major D skilling of it. Okay? And if you go my kids, your your kids and say, you know, how’d you build such and such a thing, even if they’ve been to computers, know that, besides how they were born, when the cloud was up and running, and for them? This is harder stuff, right? So you know, it makes sense that it as a consumable, I need a function, right? I need database or whatever becomes a thing for them.

And it’s a way to deliver and then having a few of which are really good at that makes sense, right? Whether it’s Oracle, Amazon. So, you know, the idea that you are an expert in running databases, is a core right is a core to your business, it probably isn’t. So, I think once you make that bet, it is the case, we won’t have standards from the cloud guys. So whichever one you choose for a function like a database you’re in, or at least the cost of moving will be high.

37:52 - I saw many edge apps, I think, will be the ones that see lots of data. Right. And if they seeing lots of data, they’re gonna use traditional mom in the cloud services that are provided by the big guys. And so once you start to do that, it gets hard that is you can see how Amazon offering an edge offering or Azure or whatever, you know, that becomes more and more tempting or, but then I mean, this this is this is where it’s interesting, because what you just described is sort of how we got legacy infrastructure already, you get locked in, because you don’t have the skills or because you’ve you’ve invested.

And then you get somebody like what swim is doing. It’s like, Hey, we have a database that doesn’t require all that stuff. We have a way to, to not have, right. So there’s a there’s a instead of redshift being the thing that gets pulled back into on premises software. Could be that that’s, that’s already there. And then the next group of nimble companies is going to show up and saying, you know, what we didn’t we figured out an easier way to move this data.

We didn’t need it. Well, the thing about I think we’re moving from point where you had legacy software, and you didn’t move it up in functionality, right, because it ran forever. And now well, what Amazon does, or whoever the car guy is modified functionality continuously under your feet.

39:30 - So redshift might go from a million rows to a billion rows, capability wise, you didn’t necessarily see interface change. You just use it. And they are better delivering backups and snapshots and availability guarantees and all those things, right. So those become more those will improve all the time.

39:57 - So is there an off ramp, I mean, from that perspective, why do you Why do you own one? I mean, it seems to me that we’re still trying to get people onto the cloud. And here we are worrying about off ramps, we have a pricing issue we have kinda, but we know where the barriers are. And you know, where we’re all foaming at the mouth of bad luck in but probably would be better security wise than bunch of other ways. If we could just get people on the cloud.

40:32 - Trying to get on the cloud, a bunch of crap, they’re still running on prem.

40:39 - To be honest, there’s a bunch of stuff in Australia on prem that people are fulminating about security wise, we’re not gonna let this run in the cloud. It’s all nonsense. It’s just people who are afraid of losing control.

40:55 - Those are Are those the last 30%? The I’m sure there’s a bunch of apps that are very hard to move, by the way. And so I’m not trying to make it easy. Easy, but I mean, and by the way, is your cloud, I mean, somebody else’s infrastructure. Right? Right. And then you pick the level of infrastructure service that will represent your needs. But, you know, people running around in data centers tripping over Ethernet cable.

41:29 - Well, you’re, you’re describing cloud in a generalized dynamic infrastructure sense. And I’m hearing it is one of three providers.

41:39 - And so so the question that you’re asking, I think, is the way that that’s for an app? I think the cloud I mean, we remember three or four years ago, we went through sing, which was, I’m going to choose Azure for my cloud for my whole my whole company. Well, we know that idea that ideas is a non starter, you’re going to choose it per app, you will choose the right cloud region.

42:07 - Right. What so would your driveway? Yeah, that’s fabulous. I love the disagreement. Because I see some oaks with that made that choice.

42:17 - We’re in Azure Cloud, but I see others where it’s correct. Yeah, I mean, but I, we, we kind of jumped by some stuff. And I guess we can’t focus on every item that’s brought up.

42:28 - But there was a lot to unpack with what you said before Simon, and well, I agree with her. No, no, it’s good. I mean, you bring up a lot of you want, you’re worthy of a whole nother hour on this platform. But there there are a lot of assumptions and truths about where infrastructure is one of the assumptions or not assumptions, one of the truths that a lot of people don’t understand because of the visibility of cloud and the size of the cloud operators, is how much of what’s on cloud now is net new workloads.

And how much of what was originally on premises is actually in public cloud now. And that number is still below 20%. Right? I mean, probably closer to 15%. So what that says is that SAS has been the predominant net to write well, SAS, potentially, but no, I mean, even net new apps that are custom built by, by their owners on a cloud platforms, right, right. Because, you know, it’s like anything.

43:26 - I mean, we’re talking about the Jevons paradox effect, right? You make it easier to solve for a problem, and people solve for more problems. Yeah. But the other part is that it’s, well, I agree with you and have made that same argument publicly that many folks use excuses for why they don’t have to adopt cloud for an application rather than reality, like it’s not good security, or they, they can’t get the compliance they want or Yeah, whatever, whatever, whatever.

And that’s all true. But it’s also true that on the other side of that coin, there are some apps that from a money perspective, it just doesn’t make sense to move.

44:05 - Right, there is no value from the cloud. And so operating it on premises until it decides to die is the right thing to do. And there are also a significant number, and it’s not and I’m not trying to say that, you know, oh, this is evidence that cloud is dying. Quite the contrary, as with every, every solution, there’s a maturity curve. And as customers get to a point where they realize that they can offer a quality of service at a price point that makes sense for them from a scale perspective, then they will do it themselves.

44:41 - And, you know, that’s been evidenced by any number of large and small companies that have taken work and put it back on the equivalent of on premises. Yeah, they have to, but they have to have a super modern skill set and application. omus here, right. Yeah, no, absolutely.

45:00 - They do today. But I would argue that the natural trend here would be that that should that should become less of a of an of a specialized thing. And that there’s there really shouldn’t be anything that keeps us from from the standard processes and practices that are refined in the cloud moving back into self managed infrastructure.

45:22 - Right. Right. Except for the redshift argument, right. Which isn’t even the redshift argument becomes like an Oracle argument, there should be somebody who’s willing to come in snowflake, right? But stuff like is cloud only this is this is this is where, but because when I look at this, this is where it gets really interesting, right? snowflake is not a on premises self management thing. They chose it to be a SAS, in part, because that eliminated them having to deal with their customers operational capabilities.

45:54 - If you’re somebody could take what they’re doing and make it operational on premises.

45:59 - And then it’s by hardware, fixed investment by software, right? Run it yourself. And it’s certain scales, even potentially medium scales that shouldn’t it shouldn’t be overwhelmingly bad. Oh, well, no, it shouldn’t. But we’ve either SAS is really easy, right? So I mean, consuming a SAS base database. So something is our will be much, much easier. And so why would you, then this is this is the trend line. And my my point from a couple weeks ago was that we’ve built our, our research and development designs around that assumption.

Yeah, you’re right, right. So at this point, we’re saying I wouldn’t even innovate new technology in a way that that I had to do things like that myself. restate the assumption, Rob, that you’re talking about when we were talking the optics capex conversation, and we said, hey, we’ve moved everything to this optics thinking, what where I went beyond that is our r&d designs are, are integrated into our op x thinking also.

So when somebody builds a new service, or offer something new, that design from the first design is why don’t have to build this, I can rent that I can I can op x that. And so when you build that innovation, or you build that new product, it’s already tied into exactly what Simon’s talking about, like, why red shift right from the start? And that that innovation is now tied into the service provider models? The idea that I also think, I mean, from the, from the capabilities provider perspective, I mean, how many companies are there still an Oracle level sell you a database of brands on prem? Very few.

And that’s because the cost of sale is high and PS dominates and whatever else, right. So from the vendor side, it’s far easier in a way to offer everything SAS. And I think many of the infrastructure startups you’ve seen are going to die because infrastructures just moving up, right? So, for example, security for Kubernetes, there are a bazillion of them out there. It’s just gonna happen. It’s gonna happen from the majors, right? It needs to be it needs to happen as part of the the SOT This is where Kubernetes is at risk, frankly, because the majors have no incentive to make Kubernetes itself include those features they have, should they do? Should they do they want to offer a secure Scooby service? They do.

But they’re not going to put that back into Kubernetes. So Kubernetes itself has a platform. Yeah, we’ll have security hooks, but the security services will be built into the Maybe, yeah, this is where I was talking to mark about this other day.

49:01 - There’s still has to be on site services being being dealt with in terms of the hardware. And the services are just going to be given to is being given to service providers, small the cloud service providers, the people managing the Kubernetes service providers, the managed service providers, not necessarily the the small enterprises that are using those services. Well, let’s get ready to be damaging machines.

So good. I know we guys and gals. I know Mark’s going to lose.

49:47 - Mark here, you’re gonna get to wrap us up again. All right, I appreciate that. And I apologize for stealing the last minute um, we have yet to really discuss where Potential breaking points are in some of the big assumptions we’re all making, even if we don’t recognize and say those assumptions as assumptions, we’re making assumptions that because it’s hard to do X, our only option is to leverage public cloud, we’re not looking at what the potential opportunity is of x, what the value associated with x will be.

And what that might mean to breaking assumptions.

50:29 - Because people want to get there, people can’t afford to get there using the traditional methods. So they’re going to have to break assumptions and find new ways to get there. And those companies that do that first, especially as we talk about adoption of edge and digital transformation, in general, are the ones that are likely to win those that attempt to maximize what they already know and stretch it out over New Horizons are likely going to lose in the long term.

But is an innovation in software? No, it’s innovation. And in software, it’s innovation in hardware handling, it’s an innovation and deployment, it’s innovation and how to adopt and manage low code and no code applications, natural corporate architecture in the back end and and aligned with security. I mean, there are so many places where the opportunity to break the existing paradigm about what is assumed for a successful deployment.

There is just too much money out there right now that isn’t being addressed effectively, that we’re not going to wait for the cloud providers to put their hand handed fists all over it and take it all and have it cost us 50 to 100% more than it really needs to. Yeah, even more than that.

51:44 - They’re ham handed fists all have lenses that point into their own infrastructure and service. That’s right, which it’s gotten to a point now that it they’re they’re actually now have their own inertia in how those things work. And so yeah, right. And I think this is this is the thing every time you ask me why? Yeah, we got we got to wrap this up.

52:11 - The the idea that these giant clouds are going to stay nimble and stay customer No, they’re not, they won’t, because those giant clouds are legacy already.

52:23 - And, and this is why 2030 is so important that we need to keep looking at what does what were the crashes, I’m I’m the reason that they will become slow is really availability. I mean, you see, whenever somebody does something wrong, right, the whole thing falls over. So they’re gonna get very cautious about rolling out updates and new features.

52:48 - And that’s good. That’s the challenge with any of these big events. What slack was down first Tuesday, first or first day of the year? Simon? Are you saying that the SLA is there? Is? Is there in some senses their downfall there in? Yes.

53:05 - interest? Yeah. So so they helped me become very slow. What was surprising to me about the last big Amazon outage was that they were, internally massively depend on kinesis.

53:20 - And calorie, fantastically, and they, they paid a very big price for that huge. Yeah. Yeah.

53:31 - And, and as they grow, those dependencies become bigger and larger and larger. Yeah, right.

53:38 - The good thing is they are using formal methods to try and do checking, but still, they have to go very slowly. Yeah, well, you know, the, the complexity factor just explodes with, you know, the approach that they’ve taken to, to architecting, all of this. So they have no choice but to go side by necessity. And the thing that then gets fascinating from that is, by using these massive, massive platforms, you are inheriting their complexity, yeah.

For what could have actually been a very simple, or at least you have to be aware of the potential failure modes of their platform, which is going to expose to you without and they’re not even aware of their potential failure mode. So it’s like, they can’t wait to write their eyes and bugs.

54:33 - And this is a really good place to pick it up for next time. I think we’re right, that the question of why and how things are going to fail, or drive, you know, we if we don’t understand that, then we can’t build the next system. No, I agree.

54:49 - All right, everybody. I love these talks. I want to thank you guys. Good to be back.

54:57 - Thank you for listening to the call. 2030 As always, is a fantastic place where we vet these really important ideas about how things are going to change our industry. And we’re going to keep teasing this apart and looking for the motivations people have for changing the status quo, which means just giving everything to major cloud providers.

55:19 - So if you want to join, please do that 2030 Cloud, RSVP get notified about these meetings, their Thursday mornings, very consistently. And we also have a DevOps luncheon, learn where you can hear more about what’s going on in day to day technology.

55:36 - We’re talking about Infrastructure as Code prior to starting talking about Git ops and serverless and how those things shaped our day to day work environment. Thanks.