Computing power is a vital part of modern life. Should access to that power be more equitably distributed? Is there a role for a public-utility style cloud that could make computing more cost-effective and accessible to a broader number of constituencies? Is there a role for governments, whether federal or municipal, in providing compute to citizens?
These are the starting questions for today’s episode of Day Two Cloud. Our guest is Dwayne Monroe, a cloud architect, consultant, and author.
We discuss:
- A definition of cloud as public utility
- Constituencies for a public utility
- Building for citizens rather than consumers
- Whether computing is a right
- Funding schemes
- Governance models
- Whether means testing should be required
- More
Show Links:
@cloudquistador – Dwayne Monroe on Twitter
monroelab.net – Dwayne Monroe’s Web site
Azure Cost Management for Busy People: A Guide for Business and IT – Amazon
Ned’s Terraform Tuesdays – YouTube
Day Two Cloud 013: To Do Cloud Right, Leave Data Center Thinking Behind – Packet Pushers
Transcript:
[00:00:05.600] – Ethan
Welcome to Day Two Cloud and today, well, I don’t know if this is a heavy guitar sort of a topic or not, Ned, but we are going to be talking about the cloud as public utility. The big idea here is imagine you have a government agency of some sort that is providing cloud computing resources to the populace at large. And Dwayne Monroe is our guest here. He’s going to be joining us for this. I got to admit, Ned, that I came into this skeptically as someone who’s not very eager to see government doing all the things. [00:00:43.160] I don’t know, man. I don’t know. I don’t know. [00:00:45.260] My my opinion may be changed during the course of the podcast. Do you have any big thoughts here about the conversation I had with Dwayne? [00:00:51.440] – Ned
I had a similar reaction. In fact, when he pitched the idea initially on Twitter of a public cloud utility, I was like, no, no way. That’s a terrible idea. It’s going to suck. It’s going to be slow and it’s just going to waste people’s money. I am completely against this in every single way and now what? Over the course of the conversation. I started to come around and it was really the way that he framed it and the way that I started figuring out in my own mind some similarities to other utilities that exist out there and the way that technology has developed over time. [00:01:23.400] – Ethan
Yeah, I went through a similar kind of a process of strong skepticism. You just you were blunt. I was trying to be polite, but you just went for it, so fine. Yep, it was I was similarly like, no, just no. And also no. And then as he was talking through and I started thinking about it. No, wait, this does actually make some sense. And I drew some of my own context, having worked in government as as support for his logic. [00:01:51.110] And then actually early in the conversation, Ned, we talk about this, a big idea of is cloud computing or is computing a right, which was interesting, not that we came to any hard conclusions on that, but it was a very provocative sort of thing to roll around in our brains. Well, enjoy this conversation on Day Two Cloud with Dwayne Monroe. [00:02:12.320] Dwayne Monroe, welcome back to Day Two Cloud. Now, when we were getting prep for this show, I thought you hadn’t been here before. But you know what? There’s a lot of Day Two Cloud episodes that Ned recorded before. Ethan was along riding shotgun as the co-host. So, hey, maybe there are some folks, though, who had not heard you when you were on that show before with Day Two Cloud. So tell the nice folks out there briefly who you are and what do you do? [00:02:34.760] – Dwayne
Yeah, thank you. So I am a cloud architect. I am focused on the Azure platform. I have been in IT for a little over twenty years, started as everyone. And I guess who has my venerability, shall I say, in the field, started on premises working on big data centers and then pivoted my career to the cloud about seven years ago and did not look back. [00:02:59.990] – Ethan
All right. Well, we have you on the show today to discuss an interesting idea that caught several of our several of us that have read your blog and so on our attention. And this is the notion of cloud as public utility. [00:03:13.010] That is, if we think of cloud as a computing platform that everyone as a society could benefit from, should cloud be a public utility at least? I think that’s the big idea. And I want to hear it from you to make sure I’ve got that idea right. [00:03:29.030] – Dwayne
That is indeed the big idea. And the impetus for the idea is a recognition on my part and indeed others. There’s an initiative here in Europe, I should mention that I’m living in the Netherlands, a recognition that computing power is a vital part of not to put too fine a point on it, probably our survivability as a species. [00:03:55.580] That sounds dramatic, but we do indeed need computing power to solve many of our most pressing issues, such as modeling the climate and things like that, and that access to that power should be more evenly distributed and not dependent upon access to capital the way it is today. [00:04:16.670] – Ned
Right. So you’re indicating that rather than having Azure and AWS being the big players out there and they are pay for play, I mean, they have a free tier, but extremely limited, rather, you would have some sort of public cloud that is available to everyone at a fairly low or nonexistent fee. Is this is this the sort of thing that you would charge for or what would I guess what would the financial structure even be for this thing? [00:04:42.530] – Dwayne
So first, let’s talk about the the the big three in this field. This would have no impact upon them at all because there obviously would still be an audience for what they do. [00:04:52.970] And I mean, obviously, I work on Azure and I have nothing against Azure. Well, the only thing I have against it are the things that annoy you that are that are then solved like, you know, in some future iteration. This offering that I am I’m talking about would be something that would be targeted primarily towards to those organizations and indeed individuals who cannot take advantage of those platforms. So I think one of the examples that I had in mind was, let’s say, a community college that absolutely could benefit from access to powerful computing platform and wants to have the NIST definition of cloud. They want to have access to that. They don’t need to build their own data centers anymore. [00:05:37.890] They want that. But they don’t want and they and they should not be required to consume their resources for you to helping Amazon, Microsoft or Google’s bottom line, or indeed, you know, the individuals who are low income who could benefit from having access to computing power. I myself growing up, I mean, I experimented on computers and it was rather difficult, you know, getting access. [00:06:03.960] It was difficult because of the era in which I was growing up, but also because of income. I mean, I had to work an entire summer to get to to save up for a ZX81 from Timex Sinclair and I for years I had a letter that Dr. Sinclair sent me when I sent him a letter saying, hey, this isn’t working. And he’s like, well, here’s what you need to do. [00:06:24.750] But anyway, imagine that young man, you know, brought forward or young woman brought forward to the present day who who needs to do things. [00:06:33.720] Wouldn’t it be good if there was if there was a publicly available modern it doesn’t have to be as innovative or as or keep pace with AWS or Azure or Google, but modern in the sense that it is built along NIST guidelines. A platform that for a reasonable cost they could have access to, and that also that was answerable to the people who are using it because it’s it’s a a public a public utility. [00:07:00.930] – Ned
Right. That’s an interesting point, because what you’re talking about, I could see so far before you just said that was this could be easily solved with the existing public clouds that are out there by them being incentivized to offer a better or expanded free tier for specific segments of the population in the way that Microsoft 365 had an educational offering that was very cheap. And they actually have a nonprofit credit that you can get annually depending on your organization and stuff that could make large chunks of Azure free for you. [00:07:35.920] They have that, but they’re not beholden to those organizations in any way. So can you expand on what you mean by them having this organization, having a responsibility to the people at services [00:07:46.920] – Dwayne
So publicly, public goods, of course, are designed like a park, for example, are designed and implemented for the benefit of the citizenry, not consumers. And this is the difference. And it’s a distinction that has been that it’s increasingly lost in our current discourse. We talk an awful lot about consumers and we don’t talk enough about citizenry. And there are many reasons for that that we don’t have to get into here. But there’s a whole history of why that is. [00:08:18.450] And I think what I’m saying is that that trend has obviously led us in certain directions that are not good. So let’s actually that’s actually piece apart that or pick apart the idea of the free tiers or the low cost tiers of the the public cloud providers or any technology company for that matter. Not to just pick on those guys, those subsidies, or rather what makes that possible typically is not their largesse, but a transfer of funds. So in other words, the public is typically already paying for that, but they don’t know it. [00:08:51.240] So, for example, there are telecommunications companies, hydrocarbon companies, they’re all subsidized by people. But people don’t know that they’re subsidizing them because they keep telling you we’re awesome, we’re just making money because we’re so brilliant and like, oh, wait a minute. For example, we could get into the story of a certain space man who or what would be space man who, you know, receives tons of government subsidies and and but but nobody talks about that. They just talk about how brilliant this guy is. [00:09:18.870] So that’s really my point, is that rather cut out the middleman. So rather than say, you know, let’s just give you a tax break to do that, well, rather than giving you a tax break, you should be paying your taxes anyway. Why don’t we actually take that money and build something that people can have input into? [00:09:37.230] Because I trust people will want to learn the technology if they don’t already know it and to use it in intelligent ways, but they have to have the opportunity to do so outside of our field. There’s a lot of talent, there’s a lot of talent out there that is not being tapped because this knowledge is balkanized. And I think that that Balkanization is to the detriment. Well, of everybody, really. I mean, I. If there were like a million more Neds out there, a million more, Ethan’s out there. [00:10:07.070] – Ned
Oh, that’s terrifying, Dwayne. I don’t want that image in anybody’s head. [00:10:11.770] – Dwayne
Maybe it’s the wrong analogy, but but you get my meaning is that there are people who have ideas and not just ideas for technology, but but ideas that require computation for this language. [00:10:21.670] – Ethan
Well you’re coming at this from a few different angles and this is something I want a kind of a fundamental premise here that I want to get your take on. Are you saying that computing is a right? [00:10:32.470] – Dwayne
You know, that’s a very interesting point. I haven’t made haven’t stated it that way, but I will say it this way. That is something that should be debated. I think that is a perfectly valid and debatable point that reasonable people could could sit down and say perhaps it is because it’s so critical to a complex society’s chances of doing well, of surviving that, rather than thinking of it as something that is a nice to have. [00:10:59.770] It’s no longer that. It was a nice to have fifty years ago, but in the twenty first in the very, very difficult twenty first century, I don’t think it’s any longer a a nice to have. And I’ll give you I’ll give you one really precise example of this artificial intelligence. [00:11:16.630] One of the things that you hear from the majors from Silicon Valley is that we are democratizing A.I. And now the question is, are they? Because the what’s being developed facial recognition, self-driving cars, which don’t even work properly, things like that. [00:11:34.270] These things are targeted towards the interests of venture capitalists in Silicon Valley. What if machine learning, for example, was actually democratized? What if people who had ideas, who could learn the mathematics, learn linear regression, learn statistics and say there’s a ton of data that I need to analyze and machine learning is perfect for that, that linear analysis of data. [00:11:59.560] But I can’t get access to a platform, even if I do, to learn the math, because I know you have to be here or be there or have a certain amount of capital. [00:12:07.990] There’s so much that we need to do and we need all hands on deck and all those hands could really benefit from having access to computational power. [00:12:18.760] – Ethan
OK, so it’s a point to be debated. Is computing a right? And I think you walk through it about as safely and sanely as you could because it is it’s an interesting question. There is some precedent for it. And that communications, were frequently subsidized by government because it became apparent that a communications infrastructure, whether it was telegraph or later, that the telephone companies, et cetera. That was something that everyone should have access to for the benefit, for the greater good of society, and perhaps perhaps we can make those arguments for computing now. [00:12:55.120] But I like the way you put it. It’s something that should be debated. I don’t have an opinion here either. It was just a genuinely open curiosity that you’ve really got my head spinning on this. Let’s take it a step further, though. Let’s say we don’t know if computing’s are right one way or the other. But but OK, let’s still say there is some kind of an existent cloud as public utility kind of thing. Now, you’ve given many examples of who could use this, who would benefit from this universities and people with low income, where they would be unable to access that computing power in a different way. [00:13:30.300] So are we saying it should be means tested or qualified? In other words, it wouldn’t be universal access. Hop on the you know, the big central cloud provided by your local government agency. Instead, there’d be some qualification of who can use it and who can’t. [00:13:45.750] – Dwayne
You know, means testing is something that Americans are very, very fixated on. And I understand why. Because you’re because there’s there’s the concern of abuse. And however, I I have a firm belief because I how can I put this? [00:14:04.860] I no longer have the cynicism about human nature that I did years ago, having met lots of lots of different people. And and despite recent events, I still I still have a great deal of faith in the common sense of the majority of people and also the fact that they will put things to good use if given the tools and the training. [00:14:28.740] So I think that if we want to do means testing, the means testing would be imagine a teacher in a classroom talking to a group of kids and saying, OK, OK, you know, Jane. OK, John, here’s what’s possible. Here’s what here’s what you can do. [00:14:44.040] I have seen children when when I have taught classes. Years ago, I did a “What is Computers?” class in Philadelphia, which is where I’m from. And it was a fascinating thing to do because. The kids that I was teaching had no idea of the infrastructure that had such control over their lives, like banking, what happens with computers and banking? How are they tracking you now? A cynical person would say, oh, well, now these kids are going to be hackers and destroy the world or they’re all criminals. [00:15:14.980] No, no, no. You need to understand, like, the the the infrastructure that determines the substrate of your life, so to speak. [00:15:23.980] And so I think that if people are given those tools and given that knowledge, that would be the means test. Right. Here’s how you learn how to drive. Now, goest thou and drive. [00:15:35.870] – Ned
Hmm. Right. I want to go back and make a point that I think is important about the idea of computing as a right, because I saw it firsthand when the first lockdown’s happened with the global pandemic. And so all these kids had to start doing schooling remotely. [00:15:55.480] And I’m a very fortunate person. I’m in a position where I have steady Internet. The children have devices that they could use at the time. But there were many families who were not in that fortunate position, who either did not have a device or in some cases didn’t have an Internet connection at home. They couldn’t go to the library or something to use it. And so the district had to hustle up and get hotspots out to them and and give and give them devices. [00:16:20.770] And then eventually they rolled out devices to everyone to have consistency. But it did highlight how necessary those resources are to educate our children, so if we are really all about public schools and public education, then it’s almost like we have to be at least a little bit in favor of providing computing to those children as part of them growing and becoming citizens. [00:16:46.710] – Dwayne
Yeah, absolutely. And this is something that I think we’ve seen around the world that people suddenly realized, oh, especially parents who, as you said, were called upon to to suddenly be home schoolers, ad hoc homeschoolers, you know, because of because of the emergency. You know, if they didn’t have the resources, then they were in a bad spot. [00:17:07.800] And that should not be so, particularly since, you know, you need your children to be educated. You need your children to understand what’s going on in order to to to hope for a peaceful and and well, well maintained society. That education. And I don’t just mean OK, now I know algebra. I mean, you know, just just process of learning about, you know, how the world works and so forth is key and but also creativity. [00:17:36.900] And I think that that’s really what I’m driving at, is that to unlock greater human creativity, we need to to expand who has access. And the best way of doing that. And I know that in many quarters in the States, that’s not so much true here in Europe, but in many quarters in the States, there’s a belief that private entities are more nimble and better at doing these things than public entities. However, many of the private entities that we admire would not exist were it not for the innovations. [00:18:10.530] And that’s the word I will use created by public entities. One of the things that’s that annoys me to no end is when someone and we all love, for example, to see the space X rockets land themselves. It’s very cool. It’s like watching a 1950s science fiction movie. [00:18:26.490] But what annoys me is when someone on Twitter says, oh my God, NASA never could have done this. And it was like NASA totally could have done that if they had been funded for 30 years. Right? [00:18:36.150] And how many NASA engineers are probably involved in that, but you don’t know it. [00:18:39.810] So ARPANET, the National Science Foundation Network, all of these things that are the foundation of the technologies that that we believe could only have been done by, you know, by the Internet providers or by by Apple or by Google, these things started in government labs. [00:18:58.200] And in some cases, they are still being developed via the public purse, but being privatized. And so I’m saying, OK, that’s fine if you want to do that. If that’s the kind of societal contract that you’ve decided upon, I have no problem with that. But you cannot also strangle the public option. Like, it’s absolutely fine to have, you know, the private. But but there should also be a public option. And the public option will probably not be less than because you will find that talented, driven people will be attracted to it. [00:19:30.690] They will say this was very cool. If I can produce something for my municipality, for my state or for the for the nation as a whole, as part of like a network of public of quote unquote public public cloud, a lot of people would be quite excited by that. And I think that there would not be a talent, a shortage of talent. [00:19:51.130] – Ethan
I’ll counter that by saying, as a former government employee, not federal in America, but a state government employee, the challenge wasn’t talent so much. [00:20:00.960] I worked with some amazing engineers up and down the I.T. stack in back in my days of working with a state government supporting many thousands of people and millions of citizens. It was other constraints that made it difficult oftentimes to for for a government as an entity to produce the thing, whatever the artifact was that we were meant to produce. And it mostly had to do with funding or funding and bureaucracy, I’ll say or, you know, the executive portion of that job where you’re bounded by law to spend money in a particular way and you need to account for it. [00:20:39.420] And then the bureaucracy adds a burden that makes it difficult to create an I.T. artifact the way that you would when you have a little more of a free hand as a business that’s just trying to get it done in the most efficient way possible. Again, I agree with you on the talent piece, but you’re shackled. You are constrained in ways working for government that you aren’t in the private sector. [00:21:02.380] – Dwayne
You are absolutely right. Some of those shackles are necessary because you don’t want you’re a friend of mine used to say, I don’t know if I want every organ of government to be super efficient because sometimes efficiently leads efficiency leads to rapid decisions that hurt people. For example, here in the in the Netherlands, there’s a right now there’s a big scandal because an algorithm was used to determine who. Could and could not who was or was not eligible for certain social benefits that people receive, and as a consequence of this highly efficient algorithm, people were fined, in some cases over one hundred thousand euros that ruined them if that if that particular workflow hadn’t been made less efficient. [00:21:42.400] In other words, if it had been are you defrauding the government? Let me call this person and find out ninety nine point nine percent of that would not have happened. And so there’s a certain degree of inefficiency that actually is necessary, I think, for things to remain human. [00:21:57.280] However, you are absolutely right about bureaucracy being a problem in many government agencies, particularly at the state level and also at the federal level. [00:22:05.830] I mean, an organization like NASA, I mean, they have like a mission. I mean, you know, like if I need to get the space probe to Mars or to to Jupiter, Saturn and JPL gets it done. But the the bureaucracy issue is something I think, that can be addressed by sunlight. In other words, many of the problems that we have with our with our governments today are the result of obscurity, opaqueness. [00:22:29.350] People and people don’t understand how these things function. And so they don’t know what to demand. They don’t know. Well, wait, why does it have to be in triplicate? Why can’t it why? Why can’t there be a simpler process and indeed there can be. Funding? The funding issue is often also a function of deliberate strangulation. It’s often a political decision. You know, that some things are well funded and some things are not well funded. So not to minimize because these are indeed like real issues that have to be sorted. But I think the answer to that is, is democracy. It’s to say, listen, let’s get this done. How do we get it done and how do we engage? [00:23:12.530] – Ned
OK, so let’s assume we’ve moved past the funding question and people have gotten on board with this idea. Now we move into the you’ve got to design this thing. And we’ve all worked in data centers. We know that those are not simple things to design or install or build or run. And especially when you move to anything that approaches like a hyperscaler level and it becomes that much more difficult because you start testing the limits of what’s possible. How would that design process go down? Do you think it would be an international consortium that would develop a blueprint for this or would it be more regional? [00:23:52.750] – Dwayne
You know, international consortium would be ideal, but we actually already have standards that we can leverage as to what the outcome should be like. How do you define you know, that it’s it’s scalable, it’s elastic. There are there are already design principles that we know. And so if someone says, you know, let’s say it probably be California, for example, because they’re huge and they have cash. Well, I don’t know if they have cash now, but but anyway, let’s just say it’s California and they say, hey, you know what, Dwayne, this is fantastic. [00:24:23.590] Why don’t you return from the Netherlands and lead this up? And your budget is is, you know, a budget is twenty million to get started. Well, I there will be many unanswered questions, but that’s fine because there were many unanswered questions before the Saturn five, you know, launched, right? [00:24:39.310] However, you know what your goal was and you knew that you know that the engineering principles that were required to accomplish your goal, even as you along the way solve the problems. So if the hyperscale leaders can do it, you know, they’re they’re not space aliens. They’re human beings. So other human beings can also do it. [00:24:59.050] And also, to be honest, I would recruit I would recruit away from Microsoft, Google and Amazon, maybe not Oracle, but I but I would I would definitely recruit. Ah, no no no I, I kid I kid there’s fine engineers there but no no no. I would recruit away because there are people, as I said, there are people who they enjoy their salaries. But, but, but if you were to say to them, listen, you can use your skills to do something that will have a direct impact upon your community, I guarantee you, because there are there are good people working in these organizations, of course, and the people like ourselves who you know, who are and don’t work in the organizations but understand their technology stacks fairly well. [00:25:41.080] Some of these people would would say, I would love to do that. I would absolutely love to do that. I would love to lend my skills to this. And I helped to build this Azure data center, you know, in the US East. And so I can absolutely help you to understand how that how that’s done and we can use different principles and so forth. So there are definitely unanswered questions. But we we know what the principles are. [00:26:04.000] We know what cloud is. We know how cloud should behave. So those and those questions are already answered. The the the technology stack part of it. It’s not simple, but it’s been done and so it can be done again. [00:26:17.530] – Ethan
Well, let me let me be cynical then and put on the hat of someone who goes. [00:26:22.120] Yeah, but Dwayne, look at AWS and how fast, they’re innovating, I guess we can say, so-called innovation because of some of the announcements, but anyway, they’re really quickly, they’re always bringing new products to market and so on. Who’s going to be interested in dealing with that slower moving, presumably public cloud that’s kind of boring and and so on. [00:26:43.030] – Dwayne
So let’s talk about the innovations. You know, they’re interesting. They’re good. We all love when, you know, when there’s an announcement, although sometimes the naming conventions are a little odd, especially on the Amazon side, as Corey Quinn says, I don’t know who’s doing the naming of some of these some of these products, but many of these innovations. [00:27:02.530] What you are doing is you are presenting, you are finding emergent opportunities from your unified API. So, for example, when you when you said, you know what, let’s let’s build Sagemaker. Well, Sagemaker is was latent within the infrastructure. You had to you had to figure out how to build it, but you had the compute. You have the storage. [00:27:26.350] You had the database, you had the you have the serverless architecture. You have the ability to create a UI. So the ability to create a machine learning platform was latent within the within the infrastructure. [00:27:44.320] You had to assemble it well in the same way within this, within any infrastructure with a unified API, you have the exact same opportunities to build value added services on top of that API based upon the the primitives of compute storage and database. [00:28:00.430] – Ned
Right. One thing that I always go back to when it comes to AWS is their most popular service, or at least the highest consumption is EC2. And yes. And it continues to be that because. That’s right. That’s one of their core building blocks for all the products that they create. I mean, when you look at something like take a relatively simple example, RDS is using EC2 and EBS under the covers and that’s very obvious. When you spin it up, it’s using the same core constructs. [00:28:31.600] It’s just that they’ve built a higher level of abstraction for you to consume and they’re managing a portion of the process for you. So I guess that gets down a little bit to the question of what list of services would you expect a public cloud utility to offer? Would it have those more managed platform services or would it be more of a here’s IaaS and go forth and make new things? [00:28:59.650] – Dwayne
That certainly would be part of it. And actually, to return to a minute to one of the earlier questions that Ethan asked, or perhaps it was you Ned, I forgot. You would start small, obviously, you know, in a modular approach. [00:29:13.870] And and you would offer, you know, imagine Azure stack. I mean, you know, Azure stack starts small, then you have you have your scale units. I think that I think that is a logical architectural approach to follow in which you would offer compute storage database serverless. Absolutely would be an option. [00:29:30.490] Like, you know why if people will need to learn, you know, virtual machine management or need that full stack, you know, that full abstraction of a pretend virtual machine, then absolutely. That should be available. Linux and Windows, although Windows might be might be. I’m leaning towards Linux because, of course, you know, so we can save money on on licensing. But yeah, you would offer that. [00:29:55.360] But but also why why do that when maybe you can skip ahead. Maybe you can say perhaps it should be what Heroku was. Right. I mean, perhaps it could be like a complete a complete serverless stack with storage and database, with a limited virtual machine offering at least to begin with, because, of course, that that would consume a great deal of resources. And again, those are operational decisions that you would have to work out as you build and as people start to start to use it. [00:30:25.360] And then you say, hey, you know what? This this isn’t working. That isn’t working. So obviously, you’d have to be nimble and say what do the people who are using this want? What do they need in order to be successful? [00:30:36.520] – Ethan
Well, you’re saying you, but in fact, it isn’t to you. It’s a conglomeration of bureaucrats and executioner, executioners [laughter] [00:30:45.510] – Ned
Operators, maybe? [00:30:48.670] – Ethan
So what would the governance model look like for this? Dwayne, who would be making the decisions of we should maybe start with this more of a service kind of offering rather than you go over that ground that’s already been been trodden for a decade now? [00:31:01.240] – Dwayne
Yeah, and here’s where, of course, things that here’s where you’d have to have a lot of a lot of arguments. [00:31:05.740] I think that because no doubt there would be some some crotchety legislator who would say, well, I think you should use, you know, Windows 3.11. [00:31:17.110] It was good enough for me, but it should be good enough for you. I think that the structure of this would have to be built in such a way from a bureaucratic point of view that the teams empowered to do this would have the liberty to make those kinds of decisions, and the requirement would be that your decisions should be based upon feedback from your user community. In other words, you are responsible to the user community. It is not merely a matter of you saying, I know better. [00:31:49.480] You can start someplace. But once your user audience grows and once they become more educated about what’s possible. So, for example, education would be a big part of this. You know, you’re teaching people what this is. And then they go from I don’t know what the hell you’re talking about to. Yeah, you know what? I think we should have more virtual machine instances. And I’d like it to be CentOS. People can go from from not knowing what in the world that is to having strong opinions about that, as we know in several steps based upon their exposure to it. [00:32:21.820] So it would be a light bureaucratic touch, although certainly there would be auditing and oversight to prevent abuse and to make sure that finances are good. And that and that, in other words, that the mission is being fulfilled and the mission being providing the service to the community. As long as you’re doing that, then I and I think that would be that should be that would be written in law that we can’t be interfered with. Just as you know, I’ll use NASA again, just as NASA had a remit and there were plenty of congresspeople who wanted to interfere. [00:32:52.420] But NASA had a remit and they had the authority to to do what they needed to do. Obviously, that was based upon the space race with the Soviets at the time. [00:33:01.300] So whenever some congressperson could say, well, why are we spending one hundred million on this or actually a hundred billion, they could say, well, because the Russians might get their first, OK, that’ll shut me up for about five years. [00:33:14.890] – Ethan
Well you could have. I’m still pondering how all this goes down because you mentioned input from the community. Got to listen to the people that are consuming this product and listen to them and give them what they need and so on. [00:33:25.370] Yes, you also have a component there of people who are, let’s call them the Bureau of Cloud Services that made up of people that understand the law and the mandate behind this Bureau of cloud services and then the folks that are technologists working within this bureau to deliver the product. [00:33:47.620] But then maybe, you know something I’m thinking that’s maybe more formal, where you’d have a an advisory body from maybe other cloud organizations like AWS that could provide technical input and insight and recommendations and history. [00:34:05.350] – Dwayne
That would be fine. Yeah, that’d be fine. [00:34:08.080] – Ethan
And users that would represent that group of folks and, you know, and so on. That would help drive some of this as well. Do you see a role there? [00:34:16.000] – Dwayne
Yeah, absolutely. In fact, I like that model a lot in which, you know, you have a voice of the user. Only in this case it’s the voice of the citizenry that’s using using the platform. Who can, you know, who would have their own meetings and their own user groups and and then they would present what they feel they need for their community. So let’s just say in Amsterdam, there would be like the Amsterdam, you know, cloud, you know, public cloud user group or public utility cloud user group. [00:34:46.540] And then they would say, hey, this is what we need to do. And I could see this working, for example, in the Netherlands quite well because of the, you know, the cultural fit and also the technology drive here. And then, yes, that that that the the levels of oversight. And that might mean that to some extent things would be a little slower, but ultra speed would not be the goal now. Is that that that that isn’t because you’re not you’re not it’s not speed to market. [00:35:10.870] It’s service that you’re trying to provide. So, for example, just like if your power company has an idea for an upgrade that’s required, it’s more important that they they do the upgrade in such a way that you don’t lose power for like for for, for seventy two hours. Then then that they rush that thing in. [00:35:27.820] – Ned
Now I’m curious on two facets, because we’re comparing this to other utilities that are out there. For instance, power you still have to pay for it might be subsidized in some way, especially for a lower income households, but you do have to pay for that. So would there still be a cost associated with consuming this public cloud as a utility? [00:35:49.390] – Dwayne
There would absolutely be a cost. So, for example, if you know, if I’m doing reasonably well financially, then this is available to me. Absolutely. As a person living in the community. But, you know, there would be a cost. [00:36:02.170] However, it would be the model would have to be capped in some way. In other words, it wouldn’t. And that might also cap my usage. [00:36:10.510] But as you scale out so as I imagine this, there would be kind of kind of a build out that as the service expanded, that the capacity would expand. And as the network grows, the capacity would expand because you. You start with just the Amsterdam or the Philadelphia or the New York one, right, and then you have more. And then as you have more, you know, you have greater capacity and then you have economies of scale as well. [00:36:36.460] So subsidies at first because the target audience would be, you know, underserved communities and that could include educational institutions as well, not not just individuals. That would really be the first to market the foundation. And then you can build out from there as you build additional services. But I wouldn’t see a situation. You would have to be very diligent to make sure that people who can pay more don’t consume more resources simply because they can pay more, because that would that would wreck the entire model. [00:37:10.150] And and I’m not saying that wouldn’t be difficult. You know, there would have to be a lot of a lot of monitoring, a lot of metering. And I say monitoring. I don’t mean, you know, what are you doing with your data? I mean, OK, where’s the usage coming from and what usage profile? So it could be a lot of work, but it’s probably no more work than what we’re already doing now for for profit organizations. [00:37:32.410] – Ned
And I think that raises my second question. We’ve mostly been talking about this in an educational context. We’re providing access to this utility so that people can educate themselves about public cloud, but people are also going to build things on this public cloud. Would you envision that it would be limited to not for profit projects or would commercial projects be allowed on this public cloud utility? [00:37:55.240] – Dwayne
I think absolutely. I imagine, you know, I’m imagining small, medium sized firms like large firms, enterprises would no doubt still want to continue to use the majors so that there be plenty of business for the big three. Plenty of business, many billions to be had. However, I’ve opened up a flower shop and maybe I’m OK. Maybe I have a few. I’m doing OK. I’m not rich, but I’m doing OK and I have a few stores. Wouldn’t it be nice if I could? [00:38:26.440] If I could, in addition to the other things that are going on, benefit from usage of this, of this platform in whatever way makes sense. You know that my computing stack could live there all or some portion of it could live there because I’m providing I’m providing employment and providing a service. And I don’t have the the ear because I’m a small business person. I don’t have the ear of of Congress people or state legislators, I, I have to get together. [00:38:54.490] And even so, in my small business consortium, I still don’t have the ear necessarily in a unified way, and I certainly don’t. And although all the all the cloud providers do give good service to everybody, regardless of what they’re paying, maybe there is a ceiling that I that I reach or it’s just very confusing to me. And also, I don’t have a voice there because larger customers do have a bigger voice. [00:39:16.390] It’s just just how it is. You know, if you’re an enterprise customer, you’re you know, there’s there’s more movement for your concerns than it might be for a smaller scale customer. So, no, not no. Yes. Yes, absolutely. Commercial interests, particularly those that, you know, provide services to directly provide services to a community and employ people directly in a community would be more than welcome. [00:39:43.000] – Ned
It sounds like, to a degree, these public cloud offerings would be multi and varied across the globe. And each individual one would be more of a municipal project or a regional project. [00:39:55.990] – Dwayne
Yeah, yeah. But but there’s nothing to stop them from linking up as Colossus linked up with Guardian and Colossus in the Forbin Project, a movie everyone should watch if you haven’t seen it, two supercomputers who rule the world. Anyway. Maybe that’s a bad analogy. But yeah, like the National Science Foundation network was in which the systems themselves could be disparate. But there was of course an API for communication and as long as. But I think that also it would be simpler to get back to one of your earlier points if there was indeed a template, you know, a build template that the people agreed upon. [00:40:34.210] And that’s what we humans tend to do. I mean, the internal combustion engine is fairly standard. And that happened over time, right? I mean, things tend to get standardized unless there’s absurd reasons not not to do it, because it’s always absurd. It’s always absurd reasons. [00:40:49.570] You know, it’s never like, well, the reason why this phone is so different from this phone is because there’s a good there’s no good reason. So, yeah, that standardization certainly would be, I think, helpful, because if if the Topeka, Kansas Municipal Service and the Cincinnati Municipal Service were built along the same guidelines, same principles, then linking them up and extending services would be very, very simple. That would be a powerful incentive. [00:41:17.170] – Ned
Yeah, I know there’s there’s been several municipal broadband projects in the US that have been fairly successful when. Let’s say lobbyists have not been successful in squashing them. [00:41:29.340] – Ned
Well, Ned you and I both from the Philadelphia area, there’s there’s there’s a story that stories that we know about that that we don’t need to get into. [00:41:36.800] – Ned
That’s about different and longer podcast. Yes. [00:41:39.540] – Dwayne
Yes. About why the the city’s actually quite brilliant, brilliantly conceived municipal broadband project of many years ago did not succeed. But yeah. [00:41:50.330] – Ethan
Well, one final question I have for you, Dwayne, is about interconnecting all of these public clouds, the way you kind of envisioned here, the way I’ve understood the vision here. [00:42:00.180] You start with underserved communities to kind of build it up as it goes not we’re going to buy a lot of land in Las Vegas and build IaaS, but rather, you know, little pockets of cloud services around. So do you ultimately interconnect all of these resources? I asked that, I guess, in the context of what the big cloud providers have done. They built global networks that interconnect points. [00:42:22.460] – Dwayne
The network effects are always more powerful, as we know. Right. [00:42:27.000] So, yes, in order for this to truly, truly be useful, it would not be. Well, the one the one in Topeka is fantastic, but the one in Indianapolis is horrible now. But it would have to be, as I said, standardized and yes, linking them up over the Internet to to to get economies of scale and also so that you’d have regions. [00:42:50.850] Imagine the number of regions you could have, the number of regions you could ultimately have over the course of decades of work would make the current a public cloud providers look minuscule because you would see a map that would you know that the lights would be across the globe from Lagos, Nigeria, to, I don’t know, someplace in Siberia. [00:43:12.360] I mean, so and this, I think, is a 21st century, proper twenty first century design concept. You know what? You’re using this technology in a way that all the tools that we currently have, but using it in a way that expands our capabilities globally, ultimately like it wouldn’t be that initially, but that would be the goal. So by but by the time I am long dead, you know, there should be you know, it would be nice. Well, I would I would know it, but it would be nice. [00:43:43.080] There would be like this vast network of computation available to people across the across the globe. [00:43:49.650] – Ethan
Well, Dwayne, this has been a fantastic conversation. This has been, I got to say, interesting, because my gut reaction is maybe more typical American than European. I look askance at government as a provider of things and go and I would be more typical of that person you cited in the beginning. Private sector. That’s where it’s at, baby. The government can’t do anything right. I do have a bit of a leaning that way, but you made me stop and think you’ve made me stop and think. [00:44:17.040] – Dwayne
And yeah, when you when you examine the history and really what what convinced me, it’s just it’s just looking at the history, you know, I mean, just if you wanna examine the history carefully because you say, well, wait a minute, it’s not as straightforward as I thought. The reason this bright line. [00:44:34.110] – Ethan
NASA has my very highest respect and esteem. I listen to some of their podcasts. I have marveled at the products, the artifacts that they’re creating from all the space exploration that’s happening. And they put man on the moon. And I want not to take anything away from the other aeronautics achievements that have happened in agencies around the world. So that’s a prime example and many others as well. And again, you really made me sit back and rethink that. [00:45:02.070] I don’t know, reflexive, negative American sort of attitude about these things and about it more more diplomatically, which is also a reflection of my age. You know, when you’re younger, you kind of think more gung ho and independent. And then as you get older and observe society and understand more clearly how these things where you taxes go and the needs of different areas in your community and so on. Yeah, all those government and what’s appropriate becomes much more clear and much more. [00:45:33.440] – Dwayne
Yeah, exactly. Yeah. [00:45:35.400] – Ethan
Thanks for thanks for this conversation today. [00:45:37.380] – Dwayne
The thanks for having me on. I really appreciate it. [00:45:40.080] – Ethan
Made us all think I think Dwayne. So, so if folks want to follow you Dwayne do you have a blog, Twitter handle a book. Anything you’d like to share with folks. [00:45:47.730] – Dwayne
I do on Twitter. You can find me at @cloudquistador. And I also have on Twitter a Twitter account dedicated to Azure FinOps, which I just recently started, and that’s @AFinOps. And I have written a book on Azure FinOps called Azure Cost Management for Busy People. And I have a blog about the impact of A.I. on on the world called Computational Impacts, which I’ll share the URL. I think you can do that in the show notes. [00:46:21.990] – Ethan
Lovely. Yes, exactly. We will share all of. Those things in the show notes here, and you can find those show notes at daytwocloud.io or at packetpushers.net under the Day Two Cloud section. Again, Dwayne, thank you for being on the show today. And our thanks to you out there listening. Virtual high fives. You are wonderful human beings for tuning in, and if you have suggestions for future shows, we would love to hear them. You could hit either of us up on Twitter We’re @daytwocloudshow or fill out the form of Neds fancy website, nedinthecloud.com. And hey, if you didn’t know Ned, our very own Ned Bellavance is a prolific Pluralsight course author. [00:46:59.060] Now, of course, he’s not giving these things away because Ned’s got a family to feed. But Ned, I thought this would be a good chance to tell the nice folks which have been creating up on Pluralsight lately. [00:47:07.770] – Ned
Well, thank you so much, Ethan. And yes, I recently published a series of courses that are all around attaining the Azure Security Engineer Certification AZ-500. So if you want to check out those courses, they’re specifically targeted at just getting that certification. There’s no fluff. There’s it’s just very streamlined and really get that information into your head. And also just want to drop that I do a weekly Terraform Tuesday episode on my YouTube channel and you can find that its youtube.com/c/NedintheCloud [00:47:39.350] – Ethan
YouTube, Ned in the Cloud. Very good. And until then, just remember, cloud is what happens while IT is making other plans.
Podcast: Play in new window | Download
