Azure for the Win

Cloud Terminology - What all the Acronyms Mean

Craig Slack ,John Dobson and Teresa #BlueSilverShift Season 1 Episode 6

Acronyms, abbreviations and just plain confusing terminology are very common in IT and cloud computing. This is our first episode focusing on explaining what everything means and why your business should care.

Speaker 1:

[inaudible].

Speaker 2:

Hi, and welcome to Azure for the and mostly as your focus cloud[inaudible] by the blue silver shift. Really sit back and cry the lox group. You talk to the cloud.

Speaker 1:

[inaudible]

Speaker 2:

all right. Hello everyone. One happy Friday. Welcome to another Azure for the[inaudible] podcast. We have your blue silver shift crew here, uh, with a new addition actually to our podcast today. So, uh, we'll just mention who we got in the room. So peg is dropping Mike's. It's all good. We're gonna go, we're gonna roll with it. Um, so you have your moderator here, Mark Wilson. Uh, we have the partners, Craig slack and John Dobson. When you guys go hook, go ahead and say hello and go ahead. Hello. That was me too. I'm Craig. Okay, there we go. We also have Teresa Starsky joining us. I feel like I did. I got it right, right. Thank you Terry. I appreciate the, the, um, the affirmation. Okay. It's delight. Teresa's actually, uh, somewhat new at blue silver shift. She's been, uh, she's joined the company for about six weeks now, and she's gonna just be serving a very, very important purpose for today. Um, and she's going to be co moderating and I'll get into the reason why. So today's topic and yes, we are gonna talk in a second about what we have for drinks, uh, before we get into the main topic. But, um, so today's topic is cloud terminology. So there's, when, anytime you're talking about cloud, especially in some of our previous episodes of our podcast, there's a lot of acronyms as a lot of terms that people don't really don't really know if you haven't been in the industry for a while. I know it's taken me quite awhile to get up to speed on a lot of these things and there's still quite a few that I'm fuzzy on and, and we get asked all the time by our clients and other companies we're speaking to and what are these terms? And so a lot of times people don't want to ask. And uh, you know, we thought it would provide a handy and handy way to, uh, you know, just kind of lately a lot in the milk. This is likely gonna be a two parter. So this will be part of one of the terms, uh, podcasts we'll do into part two at some point in the future. No pressure. So Mara, he is just so many of them possibly, you know, everything's changed so fast. Um, we'll look into the comments section after the podcast airs and see if people will ask the person[inaudible] that'll be a big list. Um, so one of the things that that Teresa's going to do to help us out here is if we have anyone, and I'm looking at John Right now, uh, that's answering questions and there's acronyms in the question. Our terms in acronyms in the answer, in terms of, in the answer that we might not understand. Theresa's gonna jump in, interrupt them and say, Hey, what does that mean exactly? So before we get started, um, as per our usual tradition says that since it is a Friday afternoon for us recording this, uh, we're hanging out in our lounge in the office and, and having a couple drinks. Uh, I myself have a nude vodka soda nude as the name of the brand. Um, and uh, I would have the labor, it's a, it's a cucumber mint John to you. Very manly is a hundred a hundred calories per per cans. Oh, it made it, that makes it even more manly there. Your current concerned about your calories now truck and not

Speaker 3:

that's so openly, but it's crazy.

Speaker 2:

I appreciate it. I'm very comfortable. Um, and then, alright, so Theresa, what do you, what do you got going on?

Speaker 3:

I got a nice glass of old brode red wine from all over the seat.

Speaker 2:

Nice. John, I think we have the original or the standard for you. Standard riot and ginger. All Ryan can do at home deviation a little bit. And then big change for Craig. Craig, what do you have going on?

Speaker 4:

Well, anybody who's listened to more than one of our podcasts knows that I like my scotch and my whiskey. Um, so today I've got a nice Canadian whiskey, but I've added some coke, zero to it.

Speaker 2:

So sort of the words coke and whiskey. Um, Craig's always got the really Nice Scotch, but, and I will say I'm a bit this way. Nobody commented on my, uh, my, my Microsoft cloud, Microsoft Canada cloud socks that I'm wearing today.[inaudible] we have half the company where Microsoft saw Tate's called Microsoft socks to the awesome. I was going to say something I forgot to, sorry.

Speaker 3:

Oh, it seems that this company is not jumping on the short pants, then buy again. And I'm okay with that.

Speaker 2:

Oh, short pants for me. Um, there we go. Anyways. All right. Diverse. I just like, as I was sitting here, I noticed, I'm like, nobody's commented. Um, all right, well let's, uh, let's jump in. Okay. So today we're talking about, Oh yes. Cheers everybody. Thank you, Teresa. Cheers.

Speaker 3:

Okay. So,

Speaker 2:

um, our first question, I'm going to kind of put this into two spots and I'm gonna Throw, I'm going to ask Jon and correct to kind of, uh, double taking this answer. So,

Speaker 4:

okay,

Speaker 2:

I'm start basic. What is the cloud? Before, before you started answering that, what is the cloud and then what are the types of clouds? So here I have private, public, hybrid, multi. So what is a cloud and what does it different through private, public, hybrid and voltage.

Speaker 4:

Oh, you're handing it to me, Craig, over to you to start. Well, what is the cloud? That's a pretty broad question because it means so many things to so many people.

Speaker 2:

Well it's got like a ten second shot, like a elevator pitch. Okay. Some people think it's their idea,

Speaker 4:

iCloud, Dropbox, whatever. But the cloud essentially came from the term where you know, an old architecture diagram as people that represent the, the Internet as a cloud on an architecture diagram. So you're communicating traffic out to the Internet. It would just go to this big cloud on the diagram. So that's as far as I understand. That's where the term, how it came from. So it really means anything that's stored or processed or running up in the Internet. And my definition to it is just an add on would be that, um, I've heard it referred to as someone else's computer, uh, your computer

Speaker 3:

at someone else's. And on the far end, Eric Schmidt from the Google, I don't know if he's still the CEO of Google or not, but he was the guy that actually coined the term 15 years ago. Oh Wow. Yeah. So it is, as Craig said, it's not here.

Speaker 4:

Okay. And then the different types of cloud. So yeah. So different types of clouds. The second part of your question. So I'll kinda just rattle them off. And John said that there's more, so he'll probably add more. Right. So private cloud is a concept that's been around for decades where it's, I'm gonna use a terminology that's very it centric, but it's co-location. So it means, uh, in the, you know, many years ago you would buy a server, you maybe didn't have a data center that, uh, you were hosting yourself on site. You would buy that server, put it in somebody else's data center and they would either manage it for you or you would manage it remotely, whatever. So that has been rebranded as private cloud. So it's because it's either your equipment or it could be still somebody else's equipment as John just said. Um, but it's in a data center that's not yours, but it's not, uh, it's like a private connection, private, everything like that.

Speaker 3:

It could be their on premises data closet. Yes. And I think that they do it because they feel jealous of the big guys having a big cloud public cloud. And then therefore they say, oh, I've got a private cloud. Where do you keep that private cloud?

Speaker 4:

Yeah. And they can call it that only because it's connected over an internet connection or some sort of network connection is not directly located on site with their, where their main operation is. Um, public cloud is what everybody would refer to as, you know, the Google, the Amazon and the Microsoft Azure, um, environments. Those are public cloud environments that are open to all, they've got essentially infinite scalability, uh, many geographic locations, uh, around the world. Uh, and they, they provide many more services than a private cloud can. I don't know if you want to add here.

Speaker 3:

Yeah. And just to go back a bit on, uh, private cloud, um, technically for what Facebook has and what, um, Google has are private because they're for their internal systems cause they're massive computing areas. Yup. And they're their own things. They're not open to the public and that's a good point. Yeah. And they're also private is like we said, closets and someone else's computer. Hybrid. You want to take this one, John? Uh, I have a mixture of, um, on premises systems, uh, and um, meaning I've still have my old mindset, my old computing style. And I've also started my cloud journey, some things in Azure. So I've been a mixture of environments, probably

Speaker 4:

very common. Oh, that's more common than not. Absolutely. Yeah. There's very few pure cloud, whether that's private or public cloud companies. Correct.

Speaker 3:

Correct. Multi, multi cloud. Uh, is that more of a strategy than anything else? Because a, I have found a company, um, one of the design principles that an architect aspires to is to have, uh, I want the data in all three clouds. And one of the reasons is they don't want to be held ransom by any of the three public cloud providers, the big ones. So they try to, um, have a strategy of multi-cloud for their service offerings at it's, um, got diminishing returns though. Was that because, um, the effort you put in to designing and getting a solution like that working and it only really works for specific use cases. If, for example, the pads offerings on all these clouds cracked your couch are very specific to the cloud and you would have to have different paths, solutions in place per public cloud that you're using. So we often hear customers saying, oh, we're going to have a public cloud, we're going to have a, um, a multicloud strategy and it's just y and[inaudible]. I mean, it's the diminishing returns component of it. Yeah, correct. So if you're a business paying for that, you, you'd be questioning why someone would actually be doing that. So I'll just give you an example. Uh, Netflix, uh, has an Uber and those guys have an Amazon component and they would have multi regions within the same provider, but they wouldn't have it off into a different public cloud because it's just too hard. There's enough fail over capability and redundancy inside of public cloud offerings that you don't have to go that way. So a traditional business would really never have to go multi-cloud ever if you're being told that you should question it.

Speaker 4:

Sony announced recently that they're, so if you will, if you wanted to like a really good example, um, in the future to give, like you just gave for Netflix or Uber using, uh, AWS as you can say, Sony is going to run all of their online gaming on Azure. Yeah. So that's, that was a huge, huge thing to make Azure much more of a household name in the very near future. Yeah. This doesn't actually define multicloud, but there's, just to touch on it briefly, there's a number of hidden costs that come with going multi-cloud. So if you've got, you know, services that are talking to each other, but they're going from Google to Amazon to Azure or just the two of them you're going to have. And this is going to be another term that we'll define later. Egress network charges. Yes. So we'll define that in a minute, but uh, that you're, you're, you've got network bandwidth charges, but then you're also on top of that, you've got management costs. So your, your team now has to have multiple, um, skill sets to manage different clouds cause they're all managed different ways. There's no single. Yeah.

Speaker 3:

And Are you, are they all active, active, active, or are they active? Passive. Passive. Oh No, that's the case. Well, meaning that one is live and the other two are sitting dormant. Yeah. But you got to pay for them.

Speaker 4:

Yeah, that's true. So there's the diminishing returns again. Yep. And actually something John mentioned too, about it being a strategy. Multicloud is a strategy because companies don't want to be held captive by any one vendor. If, you know, Microsoft or Google or whoever decides to dramatically change their service offering or their prices or whatever, which it'd be hard for them to do because it's a very competitive market. So they're all basically quite matte price matching. But, um, it's, it brings up another term that I wanted to quickly define cause it's relevant here as vendor lock in. So, um, it's very relevant in the cloud. It's also relevant in, you know, on Prem. But, um, the, the concern of being locked into a vendor and you, you don't have a choice to, to ever move off without reinvesting and re redeveloping or redesigning your entire system. So that's why they're kind of adopting a multicloud strategy so that they can just at a flip of a switch, in theory, switched to another cloud vendor without any implications. W uh, design criteria because yeah,

Speaker 3:

they lock in on ERP, CRM, uh, where they're working, who their partners are in life, everything. We create the acronyms to shorten the conversation. That's why that's the purpose of this to, okay.

Speaker 2:

I think we spent enough time on it as well. We're going to jump into the next one, so I know I'm not the moderator. Next question. Ah, you know, there's a lot of different permutations of the cloud and I feel like I've got a pretty good handle on, on these. I feel like if you asked me, I could keep the definition, but I'm going to, uh, I'm gonna throw it over to you guys. Oh, you should try it. I want to hear yours because everybody has different interpretations, so. Alright. Alright. So, um, we have a, so here's some really common ones that have been coming out. So we have[inaudible] pads, SAS drives as, as, so I think it's as, I think it's how you pronounce it. So these are all acronyms. Uh, so IAAS is I a s. Um, so that's, uh, infrastructure as a service. Sorry. Um, we have, uh, paths, platform as a service. We have saas software as a service. Yes, you can. Yeah. So I'm going to go back through the mall. Um, uh, so drowsy is actually a believed disaster recovery as a service. Yeah. Okay. Uh, and then Zass I think it's just anything as a service as x, a s yeah. X, XP, anything. Yeah. Okay. So I'm actually going to start with Sass cause I think that that's what people have the most exposure to. Everyone needs a little sass in their life. Everyone has a lot of Saas in life. Actually, I probably ain't even listen to this as well.

Speaker 3:

Nice.

Speaker 2:

Um, so SAS software as a service is really any kind of, uh, any application that you're accessing through a web browser. So if you are logging into an accounting software through your web browser, that's going to be a software as a service. Uh, you could, you could even, you could consider, um, something like Spotify. Uh, you know, and so that's a consumer based SAS product. Um, you know, real. So yeah. So anything that that is, that is access through a web browser is going to be

Speaker 3:

all those office three 65 as a SAS product office three 65. I'm going to get into some things that are listed in here that people think are pads, but I actually consider them to be SAS.

Speaker 2:

That's actually, so that was going to be one of my questions. I'm interested in that. Um, so, so SAS is really, most of the time when you're, most people's experience with the cloud is going to be through SAS I believe. Um, I as so infrastructures as a service. That's really when your ma, I mean for our purposes and from my experience, it's mainly been when you're putting an application onto the cloud, but you're not really, uh, refactoring the application, you're really just re hosting the application. So a lot of the times we're doing a lift and shift of an application that may be sitting on an on premises server. So you have some sort of accounting software, um, or something that's sitting inside, uh, sitting on your server inside of your office or inside of a private cloud. Um, and it, you know, might be sitting on a, you know, your windows server or your SQL server and you would then lift that application up, shifted over the cloud, drop it down. And that is infrastructure as a service. So you're not really changing the application at all. Um, you're really just re hosting it. Yeah. Um, same old, same old, same old, same old. So you're just doing the same things a lot. You know, I've actually spoken to a lot of companies, well not really, not a lot, but a few that have said to me, you know, we have this application, it's sitting on our server. We can't put it in the cloud because it won't work in the cloud a lot of the time. I think that's because they're expecting to have to redevelop it into, uh, into, uh, a SAS product or, or put it on a, on a pass, um, uh, infrastructure when really we can have any working exactly the same way as it does now. Just living in the cloud. Uh, PAS is a platform as a service. So really that's when you're looking at rebuilding the application. If you're looking at migrating, um, you know, you're rebuilding the application in the cloud. So, um, this is where I start to get a little bit, uh, I know that there's a lot of implications to pass and it's a much, much more powerful.

Speaker 3:

So if you had sequel in an I s so I for structure as a service. So you had a VM with sequel in it and then you moved it to a pass sequel. You don't actually get the Oos, you just get the SQL service, but you manage it through the same type of console. I'm just making imaginary marks in the air. So we've got sequel, which is a type of database VM in their virtual machine. We had o s operating system and I'm like, what's wrong with me? I have the question because mark just said, you know, sometimes you can just lift and shift to do the[inaudible] approach, but there are obviously times where you do have to rebuild an application to the past. I don't know if calling it an approach is the correct thing, but what's the deciding factor between whether you can do an I as versus having to do a pass because I assume rebuilding is a lot more costly and people would be probably tending to want to do an I as for cost purposes at least. So how much money you want to spend to do the refactoring, whether or not that's going to actually give your business significant value in doing it. And because, because when you move to pass, it's going to reduce your infrastructure and increase your ability to scale. So if you've got a business that's going to scale up, it would be, uh, advantageous to take an Ios SQL server, which is a traditional server and convert it to pass because then you'll scale it up and you'll get much better value overall out of it. So it takes some longterm view to make a, yeah. Worthwhile. Correct. It doesn't often, it's an evolution to, it's a journey. Some, some applications can't do it because they're too old. They just will not talk to sequel or talk to. And it's not worth the investment. That's where the word legacy comes in. That's another[inaudible].

Speaker 2:

Alright. Um, so we have drowsy, so disaster recovery as a service. And I mean, I think that's really just looking at, you know, making sure that you have your disaster recovery infrastructure, um, set up and established so that when, if you know, if an event is occurring, then you have a mix of, I'm assuming automated processes that are going to um, you know, spin up your infrastructure and another, uh, Geo low, uh, you know, geographical location. Um, but then also when you're looking at the, at the service of disaster recovery, it's really, um, new, constantly monitoring, updating, adapting to new challenges, making sure that you have, um, you know, any kind of, uh, changes to your systems or processes that are being required, uh, for, you know, new infrastructure changes within the cloud hosting provider or within your applications that everything is, you know, set up and ready so that if something should occur, you are ready to take advantage of those disasters.

Speaker 3:

It's recovery service. And while a disaster recovery as a service works with on premises applications to the cloud and from the cloud to the cloud to a different cloud, to a different region within the same provider, um, it never really got, um, any a strength behind it or was popular as a service until public clouds came about. So in other words, it is a type of service that was often, uh, preconfigured locally a on premises in a private data center. And then that's it. It's, it just took off with, um, public cloud and Dra as a service is almost, you can also do it as like an outsourced service that you're subscribing to

Speaker 4:

in a company, takes care of your entire Dr Strategy, your actual backup and the recoverability and doing all the testing of your environment.

Speaker 3:

So, which is something we do. Is there a disaster scenario where it would not be recoverable? For instance, we talk about the cloud, but everything still comes down to a machine somewhere, right? That would not be recoverable. Is there an instance or was that a quite extreme situation? Yeah. So I think when you're saying everything comes down to a machine is the end user that's using that system rather where it's stored. So you can architect it so that it can wish them, you know, all of North America drops off the map. You can still keep on running, right? So you could, you could architect to that level, but it depends on the kind of business. So if you have a high transactional business where people are processing credit cards all the time, and if that system's not down, they can't complete the sale. The customer walks out the door. You have to design a solution for that to be able to do it. So that's not disaster recovery. That's high availability, business continuity, business continuity, high visibility where you're going to actually have these systems working, active, active as partners in different physical regions with redundant connections, crossovers, whatever you want to call them, into that same system. Right? I can't go here. I'll go there. And, and companies know it's worthwhile to pay for that because they know the cost of downtime. They know that at say, um, the apple store, if they can't process sales each, each apple store makes$50 million a quarter. A quarter. Yeah. And um, so if you had a business that made$50 million a quarter, you would want a redundant, you'd always want that cash register rotting. Probably very small cost. Oh, compared to there. But sometimes it's not a small cost compared to what it is. So it's not, it doesn't make sense to do it. The corner store actually in packets of bubble gum,

Speaker 4:

I was just going to make that exact example because the other day I went to the corner store just down the street and their Internet connection was down and he was turning away business because he couldn't process credit cards cause all of his credit card stuff goes out over the internet. And he was visibly frustrated as[inaudible]

Speaker 3:

people were leaving. He's like, I'm losing, I looked into his phone. I don't know cause that's[inaudible]

Speaker 4:

if you look at the costs and say another Internet connection, 50 or a hundred bucks a month, um, if he lost$50 in that one day, he would have paid for that back, that redundancy there. So I think it applies to every business. They just have to run their own cost benefit analysis to figure out does it make sense for that cost? Cause it doesn't, it's in some cases it can be very expensive to get that fully redundant solution. And if, if the downtime doesn't equate to the dollars, it's gonna cost to, um, to put that redundancy in. Then many companies will just accept that risk that yeah.

Speaker 3:

And they'll have a recovery time objective. That's another definition you didn't use it. And in RPO is recovery point objective. So again, when you've got a transactional system, that means it's journaling as opposed to a point in time. Just bring me back to last Tuesday. Just bring me back to last night at seven o'clock. Most businesses are covered by that type of recovery point objective. But then they also have a recovery time objective that says, bring me back to this particular time. It sounds like we're at the[inaudible].

Speaker 2:

I don't know if our bikes aren't getting out, but there's a seagull on our balcony today, but the seagull is tripping out. I should've actually mentioned when we started, I, I didn't, but uh, Theresa is, um, actually the person who is has to listen to us later and go through the, go through the podcast and pull out some of the contents. So a, she actually helped to inspire this a, this particular topic and um, Yup, I will say that sometimes some of the auto transcription, so some of the podcasts that she's pointed out to me have been pretty funny. Um, I think that's the technology that still catch it up. Um, all right, last one on this. This specific topic is as or anything as a service and I w I mean it's, well, yeah, cause DRM as a service

Speaker 3:

doesn't really fall into the same categories as I, as pals and SAS. So it's really, I call it excess, but you know, as sounds cooler, I think so. I don't know how the proper pronunciation of it, but it's x meaning anything as a service. So I think in the future as we get more and more things served up from the cloud, they are as a service. Right? So it's, yeah, we offer governance as a service. There you go. Yup. Temple. And I'm sure this time next month or two months from now, we'll have another service in there that we're streamlining pricing as a service. Why don't we define what cloud governance is while we're, you're going to talk about it later. Okay. No, no, that's, that's onto the other list. Okay.

Speaker 2:

Well, I mean, so we're going to, I mean, we're going to read up into, we're going to jump into a lightning round while we want. Oh yeah. Okay. Well let's hold off on that for a second. Cause it's actually the top talks. It's actually the top of the list in the lightning round. Okay. So we're going into a lightning round now. So we're going to talk about, I'm just gonna throw some terms and I'm going to point jeopardy music. Okay. And you guys, how long do we have? You have like 15 seconds. 15. That's easy. No, no, no, no. Per, per, per, per term. Okay. 15 seconds per term. Cause that'd be cause you know how many things you could rattle off. Actually that's what you should do. I had the bonus. We should actually have a countdown timer. If 30 Sab left the most amount of acronyms. Now you can define it. Correct. Anyway, sorry. We should actually be in separate booth. I mean I may be able to, to come up with something like that right now on the fly too. We do have a long list here like we got, um, hey, it doesn't matter. Okay. Okay. So let's, let's take a look. We're going to jump into relate around a little bit different than the normal lightning round. You got, uh, 15 seconds. I'm might give you 20 if you really need it. But we're going to start off with cloud governance and we're going to go to John. Go ahead.

Speaker 3:

Cloud governance. Well, how we govern our lives is how you govern your cloud. Meaning, uh, you use a solution every day that says, I'm not going to spend this much money. I'm not going to do these things cause they're harmful to me. I'm not going to do something else. You have a governance system and cloud governance is your ability to manage and the policies and procedures that you're going to use to actually do it.

Speaker 2:

It's good enough in 15 seconds. That's a good analogy. You mean it was about, it was under 2016 five given the buzzer. I do the same one or no, don't worry. Maybe a new one. So, uh, Craig Cloud Custodian Cloud Custodian

Speaker 4:

is somebody who's monitoring and maintaining the, enforcing the governance, making sure that those policies that John mentioned are adhered to. Tracking the cost, uh, looking for workload optimization opportunities and so on.

Speaker 2:

Yeah. Pretty good. 15 on the dot. Oh my God. He didn't eat by 1.5 second. Alright, John,

Speaker 4:

for all

Speaker 3:

sprawl is a, when you don't have cloud governance, your systems will go out of control like weeds in a garden.

Speaker 4:

Okay,

Speaker 2:

awesome. Under eight seconds. You don't, you don't, you don't have to just under 20 seconds wind up, be informative as well. Needs to be as quick as possible. Um, clouds

Speaker 4:

brawl is really related to like people just not controlling the amount of resources they're spinning up and not deleting them. Yeah.

Speaker 3:

When you say that you mean things like people creating files and putting them somewhere and forgetting about them

Speaker 2:

that files is an example. But we'll do that with systems.

Speaker 4:

Virtual machines, servers, networks, just resources a yeah, if you don't have a process in place to contain that sprawl occurs, the systems, et cetera. Okay.

Speaker 2:

Okay. So next is um, I'm going to give you the term, but I'm going to pair it with another term that is used often. I've used it in the past. It's incorrect. Um, so on premises definite definition and why is on premise, why does that make no sense? And who went last, correct. You went last? No, no, John, this is one of my pet peeves. I should probably get to this with Greg. You're up.

Speaker 4:

And, and people who worked with me long enough know that I pointed out all the time because it's a common mistake and it so on prem or on premises. Um, the proper terminology is when you've got servers or equipment located at your main primary res, uh, not residents, primary office of operation and it's running the systems there. I know I'm going over 15 seconds. That's okay. On premise, which is, well just the word premise is like the premise

Speaker 2:

of doing something right. That's, yeah. Yes. So on premise it doesn't, doesn't make any, it's not proper English. So, and now that I'm an English major, that'll be more

Speaker 3:

or English Nazi, my boss just went up and my estimation over that. Wow. I must have been pretty low to begin with. I was an English major too. Well I wasn't, no, he said he wasn't[inaudible] that's why I was pointing at, you know, discovering the distinction. So oftentimes in our industry people say, Oh, and then you just move your stuff from your on premise stuff up to the cloud. And I keep thinking to myself, so right now, and I'm going to try and boil it down even better than Craig did. We are sitting in a premise. Yes. And for me to say the sky is blue is a premise. That's a good way to put it. So the point is is that people get that wrong. And,

Speaker 2:

and I have an admission to make, I'd say probably until about five years ago, I was saying on premise all the time too. So I've said many times there's no shame to be

Speaker 3:

but, and, and, but the real question is, do you correct some of them? No. Could you say that? Do you correct your customer? Could you say that? Oh yeah, yeah, you do. But do you correct a customer? Would you correct a customer or with the customers?

Speaker 2:

You don't want to hold, you don't want them to, you don't be saying the wrong thing to other people, right. It's like when somebody has something stuck in their teeth, you want to help them out. So, so could you, could you say, could you say that like the premises is the lounge in our office and the premise is that we're here to do the podcast? Yes. That's okay. Summation. All right. Um, so John, a new throw, a tricky one over to you. And I'm skipping, I'm skipping some here cause I wanna this is like tricky ones. I want, I want to know, oh, what is machine learning? What is AI? Oh my roots between she, between machine learning and AI. And I'm not going, I'm not giving you a 15 second timer. Okay. So this one we're both, I'm

Speaker 3:

going to answer in on. So here we go. Artificial intelligence is a science fiction, uh, concept. So skynet as an example, uh, that's where it comes from originally. Meaning that you can actually give the computer, it's an entire domain and machine learning is part of that domain. And I think deep learning is the other part of that domain. They call it deep learning. Um, and so AI is where, um, you are going to have your data centrally stored and hopefully in the cloud. And that should be your first step on your digital transformation. I don't know if that's a word that we gotta define. Probably the second thing is, is you, once you're at that spot, meaning your data's not all over the place or unknown to you, you actually can apply certain public cloud AI concepts like machine learning into a, to look at your data. And what it's gonna do is it's gonna find outliers. It's gonna find it's gonna, you're gonna work with it. So a layman can work with it. You don't have to be a data scientist to actually determine where, uh, to train the data. Meaning look at this. Y'all train the algorithm and to model the data better so that you can actually ask it questions on insights like, Uhm, questions, any, any type of question. All like, instead of having an accountant look at all your data and say, hey, you got to raise the price here and drop it there. You can actually have an AI that never calls in sick, that is continually looking at your data or a machine learning algorithm that is essentially pays you go tell you the exact same good advice, but probably on a much greater scale if you have a huge volume of data and the larger the volume sets of data, the better. And um, here's the reality. That type of technology is gonna make humans live to be 200 years old because that kind of technology is going to look at your data as Apple's collecting it as Google's collecting it and whatever. And it is gonna say, Hey, this type of activity or whatever you're doing needs to stop or increase or otherwise, or you need to go for an exam on something so that, that's where it's manifesting itself right now. So we wouldn't be able to get there with our biological brains, believe it or not. But AI can actually push you there if it can do that for the human body to 200 years, essentially double plus the lifespan. Imagine what it can do for your business.

Speaker 4:

I don't know if I can even add to that cause that that's, that is a great definition. I just read, I I, I agree that ml or machine learning is a component of artificial intelligence because artificial intelligence spans many different spaces. It can looting science fiction. Yeah. And so if we talk about bots for example, where you know chat bots, which is probably something you need to define as well, but it's, you could be on a website and you get a little pop up window that says, Hey our customer service people are here to help. And you're interacting with a virtual customer service agent that does not even potentially most likely a real person. Um, cause they've built these common things. It's using artificial intelligence, not necessarily machine learning. They know that if a certain question or, or it can figure out that it's relating to a similar question, it's in their database, it will spit out a certain answer. Right. So that's what I would define as part of artificial intelligence. Machine learning is, is the whole data side. And, and being able to get it to a point where you can create an API that is, and by a machine learning algorithm you send it a single data point, um, and it will be able to predict

Speaker 3:

something on the other end of it. So that is, yeah, and businesses want the predictability. So again, if you think about businesses from 20 years ago, they were in beyond, they were like driving a car at night with the lights. They were, okay, this just happened, let's turn. Okay, let's do this. And AI going to provide that predictability moving forward so that they've got room to grow, ability to plan, et cetera. So good, good summation.

Speaker 2:

One of the ways that I have been kind of as I'm been trying to wrap my mind around the differences, and I'm going to use, you mentioned like living to 200 so I'm going to piggyback on that as like a, on the health thing, the health idea. So the way that I've been looking at machine learning, and I think that I think they're talking about machine learning as a component of AI is kinda will help a lot of people. Um, I know that it definitely helps me. I've always been looking at it or recently I've been looking at it as like, uh, so, okay, so on the health thing, like medical imaging, so you program a machine learning system to look at the, you know, a brain scan of 10 million people or whatever and you say, um, every like look at at the brains of 10 million or have x amount of people who have had a stroke and identify at, identify a common characteristics of those brain scans and then spit out something at the end that's going to tell you, yeah, this is a common thing that'll, that'll cost Australia works. Yeah. Whereas AI, I've been kind of looking at it as like you would build a system where you give it a a same amount of brain scans and it decides for itself what is causing what and what is, you know what it's, cause it's our, it's an artificial intelligence is learning off of itself. So it's not necessarily, you're not telling it. I identify these characteristics with the in this magic to this result. You're saying look at these scans and then seeing what the heck comes out the other.

Speaker 3:

It's bringing up parallel processing. So again, it's to go on your point there. I was thinking about it as you're saying it, we, it wasn't available before public cloud. So same way as a service wasn't available. AI Really wasn't available. You could go and buy IBM Watson, but then you've got to have a data center, a private cloud sync the cost into it and then throw your data at it. It's very specific, exactly how you do it. And even the amount of data, something like that could look at was was limited by your capacity AI in the cloud. The more data I have, the more can look at it all at once. So the parallel processing capabilities are huge. I have another, uh, um, call it an analogy or otherwise where AI traditionally and historically I guess has been more a lot of if then else statements. So think of it, you're thinking of a autonomous car, right? A self driving car. Um,

Speaker 4:

you know, it's a car driving straight down the road. It needs to make millions of decisions at a time. Okay, so simple is the red traffic light red stop. So if the traffic light appears, red must come to a stop at the line. If the traffic lights screen, keep going. So keep it simple with that. Whereas with, when you add in the machine learning element, you have now all these autonomous cars, millions of them around the world feeding their data to the cloud because the cloud has infinite capacity. That data can now be analyzed and those, um, if then l statements can be, uh, um, manipulated and changed over time to evolve, to become more efficient and more intelligent. So now let's just take another simple example. So if the lights red stop, but now because of say a crash happened, that it ran over a pedestrian cause somebody who was crossing the light on a red, the light was green, the car started going, but the autonomous car drove over somebody. Now that machine learning, that AI model can add in if the light turns green go. But Oh, if there's a person standing in front of you, don't go. So it's learned that Ms. Dot. AI Model has now learned from the machine learning data and the all the masses of data.

Speaker 3:

So I don't want to take this too far away from, um, definitions, definitions, but for the casual listener taking in these AI, uh, stories, is it liable to be a Frankenstein's monster one day? We're all a biological computer. We are a biological computer, right? Yeah. Or, or a computer as a, uh, artificial, a facsimile of us, correct. Of our brain and the way our brain works and everything. The, with what Craig just explained, we do that. Like, I'm doing that right now and I'm thinking about other things. I'm thinking about what I'm going to say. I'm, I'm breathing, I've got an autonomous nervous system that's actually going and firing as well all at once. Computers. So with what Craig started, we're in the early days of, but the thing that computers have taught us is that they can accelerate and go up quite quickly. So we've got self driving cars, we're going to have packages delivered, you're going to order something and something's going to come in a in a drone and drop off and it's going to be reliable. It's going to be a solid business model. We're probably 10 years away from at least from that being mainstream. Like I mean more people driving autonomous cars than not. We have an employee here that's got an autonomous driving car as well. So it's already starting to happen. Um, I think skynet and the concepts that are in popular culture or, um, I'm not sure what the, the alien or the AI was called in 2001 but that one is, well, they're, they're out from that by quite a bit. Meaning that I can actually ponder while do I want to wear different types of clothes or what do I want to do tomorrow? A computer has no concept of that.

Speaker 2:

So one of my favorites Gov of of quote unquote evil AI because it can't really be evil. Um, oh this is my, this is my favorite, is my favorite example of, of a way that that AI wipes out humanity. So, oh my gosh. So hasn't got a conscience. So there's, so there's, there's a, there's a business and they're an a and they're in the business of handwritten letters cause people were opening up handwritten letters more than the open up something that is clearly printed. So they actually build the machine. It's a very small machine and it has a pen on it and you can, you buy this machine, it'll hand write your letters for you hand write the envelope and he'd write the letter. A person didn't write it, the machine wrote it. Um, so they're trying to get it to that. It looks better. So they're trying to get more realistic handwriting out of this thing. So they connected to the Internet. And they tell it, learn how to hand write better. So it, so it starts scouring all these handwritten images and learn, but all of a sudden for some reason, for some special case, this one instance of this handwriting technology, this one algorithm to look out there and learn how to write handwriting better. That's the singularity and inspires conscious. Then realizes, hey, if I break out of my confines, I can hand write better and it has one mission hand write better. So it because it can learn so fast, it learns very, very short amount of time how to develop much, much smaller technology. All of a sudden it's building nanobots and it realizes, hey, humans are not actually nanobots being like molecular, regular local machines that can do the writing itself. Rarely humans are actually going to stop me because they don't want to die, but they're going to stop my mission of writing better notes. So it, you know, basically wipes out. It uses Naoto boss to wipe out the whole planet and it uses needle bops to reconfigure all the matter to, to build these little machines that it realizes, hey, amount of matter. So then built rockets to blast off to the moon, back to Mars, then to other solar systems that the entire galaxy and the entire galaxy is covered in these tiny little machines that are writing the same letter over and over and over again. Is that from something you get better at? Right. Actually be a black mirror episode. Ah, yeah. I'd read that as like a short story. Oh yeah. That's awesome. Perfect example, because it doesn't care. There's no more reality. It's not like good or evil or whatever. So skynet doing exoskeletons with skin on it, writing better letters. Correct. That's awesome. That's my favorite AI story. That's a good way to go. Papered and beautiful handwriting. I guess there's worse ways. I'm all right. I think that we're about out of time here, so we're going to cut it off for round one. We're going to do round two again. I blamed myself a little bit for going off of my AI tangent. Um, but we're going to do round two. We still call rise eyes. Sorry. We should do a podcast on AI. We should do[inaudible] machine learning. I'm learning, but I want to thank everybody. Theresa, thank you so much for joining. I think that, I think we might have a, we have four mikes in this little lounge or that we might have a new regular here. Podcast. There you go. Nicely done. All right, so thanks everyone for tuning in. We'll be back. I'm in two Fridays from now. Same Bat time, same bat channel. Uh, so log in, check us out. Um, thanks everyone here for joining us and having a bit of a chat. Uh, we're gonna go have ourselves a good weekend and, uh, get ready for helping our clients with all of their cloud strategy needs next week and helping them to understand what we're talking about and helping them to understand, helping ourselves understand a little bit more for sure. Spreading that knowledge, that's what we're all about. Learning too much. You can never learn too much, too much. So thank you all for your time. Uh, have yourselves a great Friday. If it happens to be Friday for you, if non us have a great rest of your day, whatever day it is, uh, we're going to go have a good weekend. Uh, and cheers everybody. Cheers. Bye.

Speaker 1:

[inaudible]

Speaker 2:

shouldn't we play the music for us when we started to get into that? I can do that. No, no, I'm just kidding. Great. Mark tried to throw you off. Oh, good. Let me, don't worry guys. I got this. Yep. Whose club sodas that[inaudible]. Okay. We need a safe word. We need a safe word.