S2 E8: Bit: AWS

0:00Hey it's Tim here.

0:01Before the episode starts, first of all, just a sincere apology from me.

0:05This episode's about a month and a half late.

0:07I had some technical difficulties with the episode, so unfortunately I couldn't get it out on time.

0:11But here it is anyway.

0:12Enjoy the episode.

0:15Hello and welcome to season two episode eight of Datum.

0:18How are you, Ravi?

0:19How are you doing?

0:20I'm doing well.

0:21I'm doing well.

0:21We're we're over the heat wave that hit us a couple of weeks ago.

0:24Where we we all just completely melted, right?

0:27I'm gonna make my usual complaint that when it's happened six summers in a row, it's not a heat wave, it's called a season.

0:34Right, right, exactly.

0:35It's it's no it's no longer it's it's now a trend.

0:38Yes.

0:38It is consistently hot each summer.

0:41It's no surprise.

0:42Therefore, it is summer.

0:45But yes, it's actually getting hotter and hotter, which is in itself a concerning thing.

0:49So it's technically always a heat wave because every year it beats the previous year's record.

0:53So

0:54Technically a heat wave.

0:55Anyway, how have you been?

0:56We've been uh away from the podcast for some time, haven't we?

1:00We have, yeah.

1:00I think things have been going on.

1:02I think our last one was our quick hot take on on the

1:05Salesforce Tableau acquisition which is now completed.

1:08Yes.

1:09As of last week.

1:11So yeah, no, it's it's it's it's been good.

1:14What else have I been up to?

1:15I went to the Tableau Conference in Berlin.

1:17Oh wow, how was that

1:18Yeah, it was good.

1:19It was um I think it was the first one they've had away from Europe.

1:23Uh away from Europe, away from London in Europe.

1:26Um and it was a completely it was quite nice, like nice change of scenery, good vibe.

1:30I think a bit more

1:31uh quieter and um low low I wouldn't say low-key, but um it seemed a bit more spread out.

1:37Maybe maybe it was just the German efficient overtones that that did it

1:41But um a standard Tableau conference where you sort of go away and you you you meet people who have got the similar ideas and you come back energized.

1:49Um a couple of new features coming out, or ch

1:52um are worth checking out.

1:53But I think for me the biggest takeaway was the release of Tableau Blueprint.

1:57Yeah, I mean let's go into that a little bit.

1:59So like at a very high level, I'll probably talk about this at some other time.

2:02But at a high level, what what is Blueprint?

2:04It it's kind of like agile, I think, is the best way to explain it.

2:08It's sort of like this framework or strategy.

2:10Uh you can't it's not really a thing that you can hold and touch and grasp.

2:14It's more like a an idea or an approach.

2:16So um there used to be a document called Tableau Drive, which is about eight pages long.

2:20Um and it wasn't really fit for purpose for like companies who are doing things at scale.

2:25So what what's happened is Tableau's created this other thing called Blueprint, which is gives you a bit more detail about different aspects.

2:32So there's entire document or section about teams, about deployment, about measuring success, about

2:38um you know e education and learning.

2:41So it's all of these different aspects that some of it does seem like common sense, but it's quite nice to be able to point to something

2:47That gives you that answer.

2:49So um i i in this sort of p like what in the sort of job I do with with customers and customer success and helping them

2:56you know, be be the best they they can be.

2:58This is always really interesting for me.

3:00Um which then halked back to all things all the things I've been doing with Tableau Server as well.

3:04Good.

3:04Good good.

3:05So you you've just generally had a very busy um

3:08Busy summer, I think.

3:10Um it's been a lot going on now.

3:11You've also moving up in life, which is uh stifling again, did I hear you say?

3:16Yeah, exactly.

3:17I'm I'm back I'm back on the bike.

3:19I've done a couple of long-ish rides.

3:20Um and yeah, yeah, l lots going on on the pastal life side of things.

3:24So m moving moving on up and

3:26Yeah.

3:27So it's all good.

3:28How about you?

3:28How's your summer been?

3:29I have to say, I was about to say, uh next time you go on the cycle, why not invite me over so I can come with my new drone?

3:35Oh you finally got the drone.

3:37Yeah I did I did.

3:38So actually um the the backstory to this is so we were going to Ireland and um

3:43We were going to an area in Ireland called Kerry.

3:47So it's on the sort of southern, oh my my geography is all for southwestern sort of part of Ireland.

3:54So it's uh it's about as far as you could get if you were driving somewhere in Ireland.

3:58So that's exactly what I did because we were taking our uh Toby, our dog.

4:02Um so that was a long drive.

4:04Anyway, um I knew we're going to be doing some awesome hikes, good doing seeing some awesome scenery uh in awesome places.

4:11The weather was looking good and

4:13It was just one of those things where, okay, listen, if I'm gonna get a drone for like to fill something, and I'm gonna have time.

4:20Um I check the um, you know, Irish airspace uh

4:25uh restrictions and there were none in the area whatsoever other than the usual sort of safety um precautions you have to take when flying a drone

4:33And I was like, this is it.

4:35This is it.

4:35I'm gonna go in.

4:36So literally two days before we went to Ireland, this drone arrived.

4:39Um I got a pretty good deal on it, actually.

4:42Um it was this uh DJI Mavic air and oh my word, the the the man, I think the first three days I think I was just flying it aimlessly.

4:52I didn't really care about the quality of the fidget I was taking

4:55But once I got into the flow, I think I I I put down about four hours of flight time, which if you know the batteries are 20 minutes each and I only have three, that's quite an achievement over a few days.

5:07Like you have to b do some sort of midday charges.

5:09But anyway, no, I'll put a link to some of the um footage um which I shot in 4K, but of course I have no four K screen to look at it and nor do I have any laptop

5:20It's capable of editing 4K.

5:22But I have the raw footage, so I'll just upload it straight to YouTube and see what I'm talking about.

5:25Talk it onto YouTube for about 20 seconds.

5:27Yeah.

5:27Yeah.

5:28Very nice.

5:28Very nice.

5:29Yeah, very good.

5:29Very good.

5:30Much needed break in Ireland.

5:31Absolutely beautiful place.

5:32I think um highly encourage you uh anyone who has a chance to go.

5:36Drone aside, it's just a beautiful place, you know.

5:38It's um

5:40It's amazing.

5:41Yeah, I mean uh exactly I'm back at it and now I've got into um sort of a role where I'm looking a lot at Tableau Embered.

5:47It's been quite an interesting journey.

5:49Okay, yeah, because embed's quite a new thing, right?

5:51So it's just like embed it's like taking the the chart part of Tableau and chucking it into

5:56Into like a HTML interface as it were.

5:59Yeah, I I'd say the practice of embedding tablet visualizations is not new.

6:03I think a lot of organizations have done it in internal portals and so on and so forth for quite some time

6:08The difference is that um Tableau's licensing now enables that to happen for um your customers who are external to your organization.

6:16So previously um you'd get a Tableau server, you'd

6:20Deploy it to everyone in your organization and then you'd be done with it.

6:24If you wanted external customers, maybe these are maybe your marketing company and you do

6:29um you want to give sort of data to your customers, what you'd have to do is create an internal email for that client and then they'd have to log in and you know look

6:37look at um the content that way.

6:39There's there there's a few security risks with that sort of IT teams will sort of look at you as if you've gone crazy and

6:46And it does it's not really feasible if you start um if you have something like four or five thousand uh plus sort of external users.

6:53And so um uh basically Tableau introduced a new license called the Embed license, which allows you to specify uh

7:00you know, certain data sources that are going to be used up front, um, sort of a sort of restricted domain of what your application's gonna do.

7:07Normally it's just being embedded in a portal.

7:09And then you can basically make your Tableau visualizations open technically to the world through your whatever embed solution you're going to be doing.

7:17So it might be a website, it might be

7:20sort of uh part of a more comprehensive dashboard that's web and sort of tableau based.

7:25Um but then also you get the benefit of doing things like authentication in your application.

7:29So you do this sort of

7:30security handshake with Tableau to verify that users are who they are.

7:34And then it means you can augment Tableau content with web content.

7:38And that's been sort of a really interesting thing I've seen in the last couple of weeks actually

7:42with a client.

7:43Um so that's cool.

7:44But it comes with its challenges.

7:46It's not a um it's it's not an easy thing to do.

7:49I think um we'll probably talk about this some other time, but yeah, it's uh it's one of those really sort of

7:54It's it sounds great but yeah it b there's so many moving parts and it actually means that unlike a normal tableau installation

8:03Tableau is not solely the focus in an embed setup actually.

8:07And it it it it's it's it's a mechanism for getting data from A to B.

8:13But it's not the maj sort of the predominant focus.

8:16And because of that, um it can c you can kind of find yourself in a situation where you haven't shown it enough love.

8:22So um yeah.

8:23Interesting

8:25Very cool.

8:25Very cool.

8:26But we've ranted on for a long time and we haven't even told people what our show today is about.

8:32Yeah, so we were talking about AWS and I think broadly the crook cloud, but the focus on is on Amazon web service today.

8:40Um mainly because we've both been playing with it a bit.

8:42I think we've you've done a few more exams on the other.

8:49No, no, no.

8:50Very true, very true.

8:51We we're getting skilled up.

8:52We're upskilling.

8:53We're trying to understand what the world of cloud is.

8:56So we're gonna talk a bit more about that today.

8:58Exactly, exactly.

8:58And I think I think when we say that uh it's one of these things that's um

9:02uh when you're covering technical things you always have to preface with the fact that we are not experts.

9:08We are at the very beginning of our journey um we don't have the decade year of of experience that you'll find

9:15uh recruiters asking for even though there's such a thing that doesn't exist.

9:19You know, it we're kind of coming at this uh from a uh from an analytics perspective because we we're not cloud engineers per se.

9:26Um

9:27I don't think we want to be.

9:28Oh at least I don't want to be.

9:29Oh yeah, we we.

9:30Who's we, Ravi?

9:33Full full st full stack developer here you come with.

9:35Yeah, exactly, exactly.

9:36Um so yeah, we're talking about AWS today.

9:39Um

9:39Uh for anyone who's wondering what the hell is AWS, uh it's an acronym short for Amazon Web Services.

9:47Um I've got a bit of an embarrassing thing to admit.

9:49I used to say

9:51ORS.

9:53Oh wow.

9:54Look at this guy.

9:56When I first came across it, I I I saw the I saw the acronym L A W.

10:01And you just said it as as it is.

10:03Amazon ORS.

10:04And y you know the funny thing is, um the person I said this to didn't correct me for a solid two months.

10:11Then I not even once said, you do know it's AWS.

10:15And and then I said it in a meeting and I think three people laughed at me and they were like, You mean AWS?

10:21And I was like, oh.

10:25Would have been nice to know beforehand.

10:27Exactly, exactly.

10:28Gotcha.

10:29Cool.

10:29So so Amazon Web Server sort of came out I don't think it came out of nowhere because the history of Amazon's pretty pretty wild, right?

10:36'Cause it started off as the online bookstore.

10:38Well um Yeah, yeah.

10:40I mean it came out of nowhere.

10:42It was you didn't expect it to come from a retail company.

10:45No.

10:45Right?

10:45Exactly.

10:46Yeah.

10:47Um but I think the the really interesting thing about Amaz Web Services and the fact basically what they're doing is

10:52We've they've got loads of computing power and lo loads of computing storage.

10:56Um and they're selling it as a service, right?

10:58And I think this was the one of the first of its kind where they were like, you know what, we've got computing power, let's monetize this

11:05Um and there's I actually th once once you think about it, there are only a few companies in the world that have the infrastructure and the

11:13uh redundancy so like if something fails something else picks it up that's what i mean by redundancy um to do that so uh amazon is one of them uh another one you probably say is google yeah

11:25Um Apple probably if they ever want to get into that space.

11:28Facebook even, you know, like Facebook must have m billions of servers

11:32Oh not billions, but a l a lot of server space at least that they might be able to monetize, but you know, if they want to.

11:38Yeah.

11:39And so that that's that's the interesting thing.

11:41Uh I think this is inherently because Amazon's a retail uh sort of company and

11:46If you think of when um AWS was launched, it was back in sort of 2003, 2004.

11:51So it's it's actually been with us for more than a decade.

11:54And

11:55The the the approach generally within Amazon and something that Jay Jeff Bezos actually kind of instills in the Ethos uh Amazon is that

12:03When you build a solution, the first thing you have to do once you've built that solution is you have to make it in such a way that it can be packaged for not just your own sort of benefit and but so that other people in the organization can use it.

12:17So

12:18The team that built Amazon.

12:19com, they had to not only build Amazon.

12:22com, but they then had to modularize things like the shopping basket.

12:25They had to modularize things like the forum.

12:27They had to modularize things like the um server infrastructure that was driving that, you know, shopping carts and so on and so forth.

12:35So by forcing the organization to break the component parts of the solution down into its separate bits.

12:41it actually enabled them to turn each of those components into um products.

12:47And the main reason was if they were going to play in the retail space.

12:53Right.

12:53In a data-driven world, in a technology space, they needed not only just to sell goods and so on and so forth, but they actually needed to enable other organizations to do the same thing in the same space

13:05So that was actually his initial philosophy around AWS.

13:08Let's not just sell stuff.

13:10Let's not just enable people to sell stuff through our, you know, warehouses.

13:13Let's give the whole world the ability to sell at the kind of scale that we do through our infrastructure.

13:20Yeah, that's interesting, right?

13:21And and I don't think uh anyone else would would consider to get into that space 'cause almost it's almost like you've got the big, big players already now.

13:28I don't think there's th it's really hard it's gonna be really hard for anyone to break through into that.

13:32Like I think the original

13:34Um servers as a service, as it were, was like Rackspace who gave you a provisioned amount, but didn't give you the flexibility that Amazon Web Services offered when they first came into play.

13:44But then now we're in 2019 and we're talking about um the if if anyone is gonna break into this market, well it's it's pretty pretty difficult because there's such established players already.

13:54Exactly, exactly.

13:55And actually

13:56What's what's been interesting is to see the evolution of AWS as the pioneering sort of cloud uh product.

14:04And there's a if you if you get into uh sort of you know your first

14:08What is it?

14:09Your first hour of learning AWS, if you've used the AWS material, they give you this thing called a 10,000-foot view, right?

14:16Which is basically just showing the history and the overall view of you know what AWS offers.

14:21And actually

14:21You know, they started from one very basic service um, you know, back in um 2004.

14:27Um I think it was SQS, which is like a basically a messaging system.

14:31And then in 2006 they launched it properly.

14:35And over time, what they've done is they brought more and more features sort of modularize each of those things into sort of components to the point where

14:44This this what was previously a a problem in IT and technology infrastructure where if you wanted to test an idea

14:53Uh you had to go out provision your your infrastructure.

14:57It would literally take six months for it to arrive.

15:00Um and then you needed the skills to deploy it in a way that worked.

15:05And um the the the most important thing here is that

15:08Even if you were testing these things, this infrastructure would not scale at the same pace that you know your solution would.

15:16So you had to take a lot of upfront bets if you ever wanted to build anything that was at the kind of scale

15:22and a as Amazon or Google and so on and so forth because having that kind of hardware and infrastructure is was just a really big constraint.

15:29Mm-hmm.

15:30Exactly.

15:30And and and and that f that flexibility point is something that's so big, uh that from what I understand at least, right?

15:35So if I wanna add, um, I don't know

15:39a a bunch more of hard disk space or or or RAM or random like memory to my to my server.

15:45I can do that with the click of a button with all of these web-based services where I'm using the cloud.

15:50versus having to go down to my computer, open to pause it, open it up, actually put the physical actual hardware in.

15:58Um

15:59So yeah, I mean and and and then and I think the uh the other big thing that Amazon have w done well, I think we're into the the the sort of about about to teach onto the so what part of the podcast, but what they've really done well is the learning part, right?

16:10So we both have been using the Amazon materials.

16:13And I think what they what they did quite quite nicely was they got stuck in, right?

16:18They were like, right, if we want people to actually use our softwares and all the services we provide.

16:23Because there's a lot.

16:24Like, I think on on that training course if you have a look and and start l understanding the amount, the sheer volume.

16:32of services that Amazon Web Services offers with their their infrastructure.

16:37You know, you've got stuff like robotics and satellites and uh machine learning as well as the basic the the more simple stuff like databases and

16:45computers and servers.

16:48Alongside all of that comes a bunch of training content.

16:51And it's like, right, here's everything you need to know to understand how these products work.

16:56How these services work.

16:57We can teach you, we can educate you.

16:59And that's really interesting.

17:00And I think that's also a great USP, right?

17:03Because it's such a rich volume of content that people are using.

17:07It is.

17:07And it's it's interesting.

17:09You know

17:10They I think they did something really smart, which is uh very early on, they encouraged lots of developers to go on and build solutions on there.

17:19Um, if it wasn't for AWS, I don't think half the startup technology firms that we know today, whether it's Facebook, Twitter,

17:27Um Instagram, iCloud even runs on AWS.

17:31All of these things would simply not exist.

17:33Tableau Public, yes.

17:35All these things wouldn't exist because

17:37What what Amazon did was they took the economies of scale of all these ideas and gave them a platform.

17:45That's literally all they did, right?

17:47So everyone has great ideas, but the problem is you don't have the spare capacity just sitting around.

17:53You don't have like a

17:54What is it?

17:55Like you know, 64 core Dell Power Edge server just sitting around.

18:00Um, but if you could have that server for just two hours

18:04to prove out something, just to build something and test it, you know?

18:07Maybe it's a snowflake uh infrastructure, maybe it's iCloud or something like that.

18:11If you just have that for two, three hours, build it out, test it, show it to someone, show it to an investor.

18:17That opens up doors.

18:18And if you take the aggregate of opportunity that exists, where you know people's creative ideas in that same space

18:24You end up being able to run one of these Dell PowerEdge servers 24-7 and at marginal cost because it's always being utilized, right?

18:33And so AWS basically made this their business model.

18:37And they obviously charge you a premium on top of the actual cost of running these things.

18:41Um, but they've gone all the way to where we are today where they have a concept such as spot pricing, where you can basically say, hey

18:47I have this piece of work.

18:49I don't really care when it runs.

18:50I need it, you know, some point in the next month.

18:53When your servers are about, you know, thirty pence per hour, let me know and deploy this script.

18:58Turn my server on and run the script exactly.

19:00And if it if it doesn't finish fine.

19:02That's incredible.

19:03That's just incredible.

19:04Like you're basically turned uh server infrastructure into an auction.

19:07You've you've done Google AdSense on IT infrastructure.

19:11It's crazy.

19:12It's it's super smart as well, right?

19:14It it just makes sure that people use it and and and as as once you get to the point of flexibility

19:19It's funny when people talk about Amma's web services, you talk about the price and you know it's really cheap.

19:25But it really isn't.

19:26Like if you're if you're going at scale, it's definitely not cheap.

19:29Or or it's cheap enough

19:31for you to get started and then you start thinking, hang on a second, this is pretty expensive as a recurring cost.

19:38Um I think a lot of that comes down to like not understanding how Amazon Web Services prices everything, right

19:45I think I think that's that's something that we've talked about in the past before.

19:48Um but it that spot pricing thing you're just creating a completely new market and and you're allowing people to just be like right

19:55I want to be using this at this time for this price.

19:58Yeah.

19:58So the most interesting example I can think of this is Dropbox.

20:01So Dropbox started off entirely powered by AWS.

20:05And they're probably the most um sort of forefront example of a company that moved to this concept of a multi-cloud setup.

20:12So they actually invested into their own um cloud infrastructure.

20:17And they still rely on AWS, um, but they have this sort of multi-cloud setup.

20:22So they've they've moved certain components of their, you know, um cloud infrastructure

20:27on uh onto their own internal setup, which they have more control over, sort of more visibility over.

20:33And we'll come onto this sort of concept of security and stalking the Amazon might do

20:38You know, you run their service on their platform.

20:39We'll come to that in a second.

20:41Um, but yeah, they they have more sort of flexibility and freedom, as it were, and then they're keeping the components that aren't as critical, that are cheaper to run.

20:49on things like Amazon and whether it's S3 or or whatever.

20:52And it's a fascinating sort of model.

20:55If you look at companies like Apple who also use AWS to power

20:58iCloud storage.

21:00Um, it's really fascinating that someone like Apple doesn't see its competitive advantage in setting up its own cloud infrastructure at all.

21:07Um so it delegates nearly all of its um iCloud storage to AWS, whereas it keeps uh things like um

21:15the servers that run iOS uh verification and you know iOS app delivery, those are all still maintained and run by Apple.

21:23So this whole multi-cloud setup is also a new concept that I think is possible for the large enterprise

21:29You can kind of do that long-term strategic investment.

21:32It's almost like a a test, right?

21:34If you haven't been able to

21:36sort of manage and grow your infrastructure to a certain scale on AWS, then you're probably not ready for the multi-cloud setup because you you probably don't have a a sort of various skills behind you.

21:47So here's a big question, right?

21:48How often would you even change that?

21:50So you've just migrated all your stuff from that computer sitting under IT's desk into the cloud.

21:55At what point do you say, you know what, this isn't working, let's move to something else?

21:58Do would you even bother doing that?

22:00So this is this is a very interesting concept because um up until maybe three years ago, I think it was uh if if anyone says it mentioned the cloud, they'd only talk about AWS.

22:11Now you have uh Microsoft uh Azure, which is Microsoft's offering, and then you have Google Cloud Platform, which is Google's offering.

22:18Uh and you have sort of the smaller players like Rackspace and a few others Oracle who are yeah I love the uh Gartner

22:26um magic quadrant on this because it calls them niche players which which is like a backhanded compliment in many ways

22:34It's like it's gonna take you a lot of like legacy stuff for you to want these guys.

22:38Yeah, exactly, exactly.

22:39These are niche players.

22:41Anyway, it's like Oracle, like a multi-billion dollar company.

22:44Niche player in the cloud.

22:46It's such a back-handed compliment.

22:47But anyway, um, you know, the these these companies have come to the fore now and they're bringing their USP.

22:54So I always think of Microsoft's USP

22:57Um people will disagree with me here, but they were kind of late to everything.

23:00So their USP is that so are their customers and so they're gonna bring them with them, right?

23:04Like, hey, uh you've spent so much on this um Microsoft platform.

23:09Let's move you to a cloud platform instead.

23:11And here are the savings we can offer you.

23:13Here are all the new opportunities that open up.

23:15And so they're being very successful actually, using that, you know, bringing all their existing customers with them on that journey.

23:22Google has come at this from their core competency, which has actually been machine learning and understanding.

23:27Um, basically, they were the first cloud company.

23:30I mean

23:30Google search was the first cloud product.

23:34Before even Amazon.

23:36com, that was a first cloud product because it required so much

23:40geographical and computing challenge.

23:43I mean, I can't I like I think it was only three years ago I stopped seeing uh Amazon turn up as a bookstore in Luxembourg on my invoice.

23:52Yeah, like and and Google figure that out sort of right on, right early on.

23:57I'm actually reading a biography about how uh the Google was set up as a company and you know it's good's going into sort of lots of amazing details.

24:04But

24:04That's Google's core competency.

24:06And so I think now in today's world, when you think about multi-map multi-cloud setup, how do you transition between them?

24:13Well, it's interesting because all the platforms are building setups that allow you to hop between them and they allow you to kind of um, you know, talk from one to the other.

24:23And uh the the only real restriction is actually

24:27Um the barrier which is data transfer.

24:30Um if if if the only the only thing that's actually stopping you from moving from one to the other is data transfer.

24:37Typically all And security to an extent I guess?

24:40Yeah, I I uh I always think scare security is like a scarity.

24:45Security is like one of those things that's um subjective because it depends on which side of the wall you're standing on

24:51Right.

24:51Um and I think depending on where you're standing, it it can look completely different.

24:55So I never objectively look at it and say, oh, that's the reason you do you choose one over the other, because it just requires you to look at security from the perspective of that company

25:04And if you don't, then it's not uh it's not something you need to go for.

25:07So it's not like you either like it or you don't with the with the respective platforms

25:12Fundamentally, all cloud platforms offer you to set up security in the way that you would want.

25:16So, you know, enterprise security is well understood.

25:20The concepts are well researched.

25:22Everyone's going to be offering the same thing

25:24If you go to servers, right?

25:26Everyone's running a Linux server, Windows server, and they're all basically just doing the same thing.

25:33Where you get competencies are when people like Google have this

25:37TPU Tensor Processing Unit, which is a custom processor that processes specific types of computations, much more than things like TensorFlow and all that machine.

25:47Exactly, exactly.

25:48Exactly.

25:49And so And I'm sure I'm I'm sure I'm sure Amazon have the same sort of thing with their their machine learning capabilities and then you also have the uh you know the hardware game changing because

25:59Of course Intel's no longer a dominant player in the CPU market.

26:03You know, people like AMD and even Qualcom uh not Qualcomm.

26:07Um AMD uh

26:10And uh ARM, ARM, ARM.

26:12ARM, I see you're thinking.

26:13Yeah, ARM ARM process, which isn't actually one company, it's a consortium of companies that are, you know, invested in a in a particular way of making it shared interest, yeah.

26:21Yeah, low power usage, uh high efficient

26:24um highly multi-core setups um that are now starting to be considered for things like server usage.

26:31Well you know Apple as well, of course, in the process of marketing.

26:34Absolutely, absolutely.

26:35They're, you know, the Apple just acquired, I don't know, Intel's um

26:39LTE 5G, whatever uh acronym we're on now, number, um um, you know, chip unit.

26:46So

26:47The the the USBs are revolving around, you know, what is your core competency?

26:53What is your hardware infrastructure like?

26:55Is there anything unique to you that you can offer?

26:58And where those things are unique, those platforms are pretty good at talking to other platforms.

27:03They're not sort of siloed.

27:05So I could have a process running on TensorFlow.

27:09and then dumping the results in Amazon S3.

27:12You know, I don't I'm not sort of wedded to one place.

27:15And I think this is gonna make

27:17The cloud wars very interesting because I think they've only just begun.

27:20The cloud wars.

27:22They've only just begun.

27:24And wow.

27:25Amazon has sort of first strike advantage because it has the biggest player base, but I also think it's got the most of the brand

27:31Uh I uh well I'm not sure about the brand.

27:36Uh yeah, of course.

27:36Are they the big evil empire?

27:38Exactly because you can't disassociate AWS from Amazon and the the company that was the first to put a microphone in your living room

27:45and listen to everything you've ever said, even And transcribe it and then send you ads based on.

27:50Exactly.

27:50Yeah, exactly.

27:51So it's hard as an IT director or a CIO to sit in your home and

27:55worry about that kind of stuff when at home and also know that that's the same company that's powering your your infrastructure back in the office.

28:03And so it's Yeah, yeah, that you're you're financing that.

28:06Yeah, exactly.

28:07Your findings in the economies of scale that makes LX only twenty to fifty pounds.

28:10Yeah.

28:11Um just for Amazon to listen to you.

28:13Yeah, and and it's it's

28:15It's fas it's fascinating because that becomes a real problem.

28:18There are companies now who uh wish to compete with Amazon and inherently therefore can't use AWS, yet it's the best option for them, if that makes sense.

28:28And they're finding it very, very hard to do business.

28:31This is a a large sort of um discussion in America at the moment called antitrust, where

28:36And if we take Amazon uh marketplace as an example, Amazon sell lots of products.

28:42They of course have analytics on what's selling and what's not.

28:45They also have their own brand goods and it just so happens that their own grand own brand goods highly correlate with some of the most sold products on the Amazon Web Store.

28:54Okay.

28:55So then go for it.

28:57Yeah, no and and this this also comes down to the fact that Amazon aren't exactly you know, the reviews that Amazon had about being the company to work for, or or or rather being the worst one of the worst companies in the world to work for, uh, given the long hours and sort of

29:11Uh what's the word?

29:12Like just a a lack of compassion for a company.

29:20Basically everyone's told as soon as you go in.

29:22Uh that that doesn't do great for their image either.

29:24Exactly.

29:25Um and a especially given that, you know, th their market cap is so big and Jeff Bezos's personal wealth is just growing and growing.

29:33Um a and as as we said before, like it's not like people

29:37generally would change their their cloud platform too frequently so you're just stuck on the subscription model where you're spending twenty, thirty, forty thousand pounds or dollars a year.

29:46Exactly.

29:46And you d you don't see this friction with uh

29:49with Microsoft because guess what?

29:51Microsoft has been your enterprise um partner for the last I don't know ten, fifteen years.

29:57And, you know, they've they've yet to prove that they look at your server logs and start upselling your next.

30:03They've yet to

30:04produce a product that sits in your living room and stalks ye.

30:07They've yet to have a bad sort of uh story in the press about

30:12privacy or or anything.

30:13So what what you mentioned earlier is quite funny given you know that Microsoft Wars late to the game.

30:18I mean th in this case it's a great thing that they're late to the game.

30:21Right.

30:21They've not they've got not got a brown to sort to

30:24are such a negative brand about them, they're like, oh, we're the Windows guys.

30:27You know, we we you know, uh Excel, that's us.

30:29Yeah.

30:29Like that's that's their USB.

30:31That's their that's their like

30:32platform for selling and because of that they get leverage on these massive corporate deals they do.

30:38It's like well you're buying Office 365 from us.

30:41You may as well just push up all your stuff onto Azure

30:43Because then you get all this discount.

30:45We'll actually give you discounts on all the other stuff you're buying because you're just doing all your services through Microsoft.

30:50And and and whilst you're at it, what's what well chuck in SharePoint lists for you

30:54you and uh yeah by the way oh I see you're using Slack let's chuck in teams as well no that that'll cost nothing and uh what's your what's your database tool let's let's give you Power BI.

31:04Yeah exactly you've already got Excel here's Power BI you know

31:06They have this immense capability to bundle and actually they've been very good at it for years.

31:12I mean, everyone's always said Microsoft was gonna die, but guess what?

31:15They're still here.

31:16Still one of the top three or four companies in market cap in the world, and they're still making money.

31:21I think they got rid of certain individuals um in leadership at the right time.

31:26Uh Bill Gates at the right time, Steve Barmer, you know, he had his ups and downs, but

31:30Uh uh people criticize Steve Barmer.

31:32I'm gonna defend him, I'm gonna defend him a little bit here, because he's the hype man, but guess who started Azure?

31:38Three years before Satcha Nadella uh came to the fore, it was Steve Barmer because he saw this coming and he turned the ship in the right way.

31:45And he obviously had to leave to not be associated with it.

31:49That's essentially what happened.

31:50But anyway, yeah.

31:51Um

31:52Uh Microsoft has a lot to gain in this place.

31:54Now Google, on the other hand, you know, they suffer from the same sort of problem as Amazon.

31:59If you put something on your Google server, can you be sure that they're not going to look at your logs and see what you're doing

32:04I I remember I remember a previous company I worked for and we we sort of um we there was a massive overhaul so I used to use have you have ever heard of the IBM's email client?

32:14Uh I don't know what it was called, but yeah.

32:16You know the one you know the one I mean.

32:18It's it's something horrible.

32:19Is it lotus lotus?

32:21Lotus notes, exactly.

32:22Yes, exactly.

32:23Oh wow, yeah journalists still use that stuff, man.

32:26There hasn't got no way.

32:28God.

32:28Yeah, I know.

32:29So Lotus Notes was what what was used um at the company I worked for and then they they were like a judging and a couple of other vendors like so Google came in, Microsoft came in, they let IBM pitch again.

32:40And apparently the security message that Google with them is, ah, you know, you can trust us.

32:44We're Google.

32:45You can I was like, well, that's like I don't know, you you going into a bar and an Eastern European being like, hey, let me hi i let me take your phone, we're not gonna steal your data or anything like that.

32:54Like

32:54You know, it's the stereotypes you uh you've done to it.

32:57If you say you just trust me, you're net inevitably gonna be like, why shouldn't I trust you?

33:02Um Exactly.

33:03And actually it shows

33:05Even even when a company says, look, just trust us, we're not going to do anything crazy, that does not negate the fact that people make mistakes.

33:12And I I hate that sort of

33:15um complacency that it's okay, we won't screw up, it's fine, you can trust us.

33:20You know, it's exactly the shady guy in the bar saying, like, hey, just trust me.

33:24Yeah, that's not the problem.

33:25The problem is is that mistakes happen and actually if you don't have

33:29um due diligence up front, you don't give yourself enough chance to see where the mistakes can happen.

33:34And that's where these companies kind of fall over.

33:37in in the space at the moment.

33:39So uh we've kind of gone off tangent a little bit there, but if we go back to sort of the cloud and and AWS.

33:46I think it's an interesting time.

33:48They are going to increasingly um sell more and more products.

33:52I think they're going to relaunch more and more products.

33:54I think the cloud market's now starting to get

33:58Saturated in as much as the big players have all turned up and they're all basically selling the same thing.

34:03So it's really the same thing.

34:04It's really hard to innovate.

34:06in terms of new solutions without a competitive advantage that is unique to yourself, right?

34:12Otherwise everyone's just basically buying the same processing unit from Intel.

34:17Shoving it in the same Dell PowerAge server and putting it in the same rack that everyone can buy and put in their own data center in generally the same locations in the UK that's slow.

34:27You know?

34:28Yeah.

34:30I think we said earlier the the cloud wars will ultimately be decided on who can give you the best bang for your buck.

34:35Exactly.

34:37Always come down to cost.

34:38Like you work with any IT team, always comes down to cost.

34:42How much can we save?

34:42How much can we get a discount on this product?

34:45So and this is where I start to think

34:47How long before cloud infrastructure, funnily enough, starts to become about hardware again, basically?

34:53So you know, this thing where we kind of try to

34:57uh uh obstruce ourselves from the hardware goes full circle and actually s companies like Amazon and Google and Microsoft start investing in their own hardware divisions because the competitive advantage

35:10moves away from being the same components everyone can buy to being specifically how you design that.

35:15The kind of thing that guess what what kind of company is good at that kind of stuff.

35:21Uh tableau?

35:23No.

35:23I wish.

35:24I wish.

35:25Begins with A and ends in E.

35:28Uh as you?

35:29No way.

35:30Apple.

35:30Apple.

35:30Apple Apple.

35:31Apple.

35:31Apple.

35:32And no Apple has no desires right now to get into that game.

35:35But you can absolutely see how that model

35:39absolutely suits them, right?

35:40You know we're just gonna make you the best product possible.

35:43Exactly.

35:43We're gonna make it as easy as possible to deploy and you just didn't even have to think.

35:47We're just gonna do it for you.

35:48Exactly.

35:49Snap your fingers and it will turn up

35:50Which is funnily enough where AWS is going with all its latest offerings.

35:54Here's LightSell, it just figures it out for you.

35:56And here's um I think one USB of Google actually over Amazon is that it automatically adjusts the hardware, so you're automatically paying the lowest price.

36:04I think that's one of the features that it has over AWS, whereas AWS has this cost estimator that shows you how you can save money and then you have to take the action

36:13yourself.

36:14See see I I'm I'm I'm in two minds about that.

36:16I don't I don't I'm not sure I agree that's a good thing.

36:18I think it's just it it depends.

36:25It depends

36:26Um but but like if you're scaling, for example, if we if we we look at the world we live in with Tableau and Ultics, you d you kinda want that throughput that you kinda want that high ceiling.

36:36Because you like what happens when

36:3810,000 people come in on, you know, if we're you're at university and everyone wants to look at the student survey.

36:43Why's when everyone clicks that link on Monday morning at nine a.

36:46m.

36:46Right.

36:47And then you're going to get a coffee and if someone tells you the server's gone down, you're like, oh no.

36:51What on earth happened?

36:52And you look at the calendar.

36:53But that um but that's that's you want that ceiling rather than a flexible cloud deployment that'll sort of just eke it up slowly on it on its way.

37:01But the thing is, is that it's it's interesting because again, I I actually trust Google to do that better than Amazon, right?

37:08Because when has Google search ever been down?

37:10And how many times has Amazon been down

37:13Right?

37:14When was the last time Google search went down?

37:17Never.

37:18Exactly.

37:19And like Facebook goes down, WhatsApp goes down, Instagram was down.

37:22AWS goes down, you know, when and the funny thing when AWS goes down

37:27Everything goes down.

37:28Everything goes down.

37:31Your phone doesn't know what's going on for a couple of days.

37:34Uh apps stop working.

37:36You know, it's just it's just a nightmare.

37:38And that's the thing

37:39Yeah, I actually trust a company like Google to to nail that bit of the game a little bit better because I look at their experience in the running their own stuff and

37:48Cloud infrastructures basically make available what you use internally to the mass market.

37:52Uh uh I I also I also I also think the scale comes into this as well again.

37:57Like we talked about Azure a bit, we talked about well, you know, if you've got a redshift database

38:01Go for Amazon.

38:02If you've got something in Google Bible query, go for go for Google.

38:06The only thing I will say for Google is the interface isn't great.

38:10To use.

38:11So this is this is an interesting thing.

38:13I think you're absolutely right.

38:14Um because I think we've we've um

38:17It's funny.

38:18Uh when I use AWS, when I'm learning AWS, uh more to the point, and I've used uh some aspects of the AWS.

38:26It feels like something that was built with jQuery and back in I don't know when.

38:31And you know, it's all very, here's a button, here's a select box, type in what you want here, here's a drop-down, and it feels like it could have been built with a like a very

38:40sort of straightforward UI.

38:42Google, on the other hand, have maybe gone a bit too far on the right hand side of the dial where they've they've tried to purposefully design a really slick interface.

38:52And it's it's

38:53It uh it feels like it's it got as much love in terms of UX design as it did cloud design, right?

38:59And in a in a product like that, that's not reassuring in in in one sense and actually it gets in the way of using the product.

39:06Which I think Amazon gets absolutely right.

39:08Um an example I'll give is learning AWS versus learning Google Cloud Platform.

39:13Now

39:13The Google Cloud platform way of doing this is to literally create you a temporary Gmail account that is active for 45 minutes to cover the 10-minute lesson you've started.

39:23And then by the end of the lesson it deactivates yourself and gets rid of the credits that you had associated with that account.

39:28I mean that's just that's just incredible, okay?

39:30That's genius.

39:32Amazon solution.

39:33Let's just give you a whole year's worth of free tier, off you go.

39:37Like it's just much simpler.

39:38It works works exactly the same way.

39:40Figure it out, have some fun, and then you'll go from there.

39:43Exactly.

39:43And we just need to make sure your credit card's on file.

39:46Like you know, yeah.

39:47You know, Google doesn't actually ask you for a credit card up front.

39:49And so it's an interesting model.

39:51It's uh the Google approach is over-designed, but it took me 15 minutes just to figure out

39:57How that sort of learning platform works, if that makes sense.

40:01And that carries through to an extent to the to the cloud platform, but

40:06I think the thing to bear in mind is that we are judging it as new users and therefore our experience of using these things is mostly through interfaces, right?

40:16Uh and in reality, most companies and most enterprise solutions don't use the interfaces at all.

40:22They use SDKs or command line um uh arguments to basically control these platforms.

40:29So

40:29In one respect, I think we're qualified to talk about sort of how you use these things on face value

40:37I'm not I I don't think we're we're quite there yet in terms of understanding how differentiated the SDKs and

40:45the command line sort of integrations are and you know how do they support a wide range of languages so that you can have developers skilled up

40:53in a few languages but still find the thing you need in any cloud platform.

40:56I think that's always a valid concern for sort of IT teams who, you know

41:01Always recruiting for full stack developers, but just want someone good at one thing.

41:05Yeah, I I love the term full stack developer.

41:07I think we've been over this before.

41:08It's like you don't want that.

41:09You just want this tiny part to someone to do that and then someone available to you the rest if you ever need it.

41:14Yeah, exactly.

41:15Exactly.

41:16So it's a it's a fascinating it's a fascinating uh sort of area.

41:20I think the the the now what um

41:23Uh we've uh w you know, we're talking about AWS today.

41:27It's been around for over a decade, it's nothing new.

41:29Um there's no big revelation, I think, in in our discussion.

41:32I think it's

41:33It's an interesting journey for you and me because um it's worth it w worth keeping an eye on, I think.

41:38I think for absolute general consumption it's it's sort of like

41:41Everyone's gonna start everyone's been talking about cloud and this all disappeared, but now everyone's starting to do cloud.

41:46Yeah.

41:49for ages and ages when the whole the first big data phenomenon came along cloud is the future but then that that future is now so no you know this is the time to start

41:57thinking about what what is Amazon Web Services and and Google Cloud Platform and Microsoft Azure and what what does this brave new world mean for me?

42:05Exactly.

42:06I I saw an interesting chart last week shared by a colleague.

42:09which was talk about the hype cycle, right?

42:11And it you know initially a product shoots to astronomical peaks and the you know the peak of its hype cycle, it plummets to well below its anticipated sort of actual settling point.

42:22and then rises back to healthy levels of hype.

42:26And I feel like with the cloud, we are now emerging from the trough and coming back up to this healthy level of

42:32of sort of of of of understanding of what it actually is and where its core competencies are because a sort of a few of the early innovative companies have used it well enough.

42:42and s and trained up people well enough that those people have gone out to other organizations, started their own things and now disseminated that knowledge more broadly.

42:50I always feel like that's what has to happen before these things are industry-wide, right?

42:54You know

42:55Um the engineers that worked at Facebook, Google, Amazon ten years ago, they're the ones now running, you know, head of IT or head of infrastructure at all these other new companies, you know, that are in the FTSE 100s and stuff.

43:07And now that stuff is becoming mainstream.

43:11Exactly.

43:12Exactly.

43:15This is when you you set up and start paying attention.

43:17Not the early days.

43:18Exactly.

43:19And yeah, and you know, the market's gonna continue to grow.

43:22I think cloud infrastructure is continuing to grow at a healthy rate year on year.

43:25So that only means that you know

43:28I think more solutions will end up on the cloud.

43:30I'm not convinced we'll ever go to a cloud-only world.

43:33I think No, definitely not.

43:36Definitely not.

43:36The hybrid infrastructure is too valuable

43:39uh for companies exist and and yeah I mean banks exist but even they're threatened by new technologies like Bitcoin and you know the emerging sort of Challenger banks who are you know have app only sort of infrastructures right

43:51True, true.

43:51So which are all driven on the cloud.

43:53Like uh Monzo doesn't have like a data center sitting somewhere.

43:57Uh 100% has that entirely on some sort of cloud platform.

44:00I'd actually be fascinated to know which.

44:02I might have to do some uh

44:04IP tracking.

44:06Just put my phone on a specific IP address and put a sniffer on and let's have a look.

44:13Might be able to find out

44:15But yes, I think it's a really good time to be interested in the cloud.

44:19More importantly, there is now enough, I think

44:23you know, tangible and good good practice examples um compared to where we were maybe five, ten years ago where it was quite an innovative tool.

44:32And in terms of analytics

44:34I'm actually interested to see what possibilities this open up because to me this is like a first tier thing, right?

44:40What before before the car was invented, first you needed the uh idea of a combustion engine.

44:46Combustion.

44:46Then you needed the idea of a tire.

44:48Then you needed the idea of a wheel.

44:51You know, all of these things had to happen um, you know, in a certain order.

44:55And the final thing was the industrialization.

44:58So you could have made all these things bespoke, but actually to make thousands of them you needed

45:03the industrialization to take place.

45:05Likewise with the cloud, I think the cloud is sort of an enabling sort of thing.

45:10And with analytics, now that it's here

45:13Now we have all sorts of computational power available to us, whether it's RAM, hard drive, compute, databases that respond nightling fast.

45:21What new solutions can we mold together to answer really tough analytical problems, right?

45:27That can that need scale, need agility, sort of

45:31You know, need solutions that need to be built as quickly as the people who think of them need them, if that makes sense.

45:38So if your data scientist diamond scientist thinks

45:41of a question tomorrow, how quickly can you stand up an environment that lets him or her answer that question?

45:47Right, right.

45:48And and and that that that's that speed to insight is exact exactly the thing that that's interesting.

45:54Mm-hmm.

45:55It's like how how is this gonna enable more people to ask better questions and also scale those questions better across the board as well?

46:01Right, right.

46:03Absolutely.

46:04Right.

46:04I think that's been a that's been a uh really interesting topic.

46:08We we kind of talked freely there today.

46:10Um it was this very interesting chat.

46:12I think it's still still further structure to an extent though.

46:14Yeah, we'll like it's fine.

46:15We could easily do a whole season on AWS and you could easily do a whole season on any cloud platform.

46:21That's, you know, that's that's maybe not us.

46:23There's lots of good uh AWS and cloud platform podcasts out there.

46:29We'll probably find a couple and link them in the show notes.

46:31Um if you're at all interested.

46:33Um if you see us in person, by all means come and chat to us.

46:35I think this is a very interesting discussion.

46:36We'd love to hear people talk to us about.

46:39People's ideas, yeah.

46:40Exactly.

46:41Um we have uh

46:43We changed our name in case you've missed that.

46:45Um see a nice new logo as well.

46:47So we are datumpodcast.

46:49com and also datum pod on Twitter.

46:52So you can you can come find us there and um give us some feedback.

46:56Let us know what you think.

46:57of um of the show and also give us ideas.

47:00Um right in terms of future shows Raves we have something kind of slightly different planned don't we?

47:06Yeah, we've got something brewing.

47:07We've got something brewing.

47:09We've got a couple more episodes in this format with the bits and the bytes.

47:14And then we're going to try something new called analog.

47:16Exactly.

47:17And then Tim, do you want to talk a bit more about analog or do you want to leave it as a teaser?

47:23So we've had the concept of bits, you know, which are small sort of bytes about technology.

47:28uh bytes which are softer discussions about concepts AWS being one um you know learning being another and analog is is a slightly new area where we're going to start

47:40talking to the personalities behind the technologies basically.

47:44So whether it's a developer, a technical evangelist, we're going to be talking directly to those people and

47:50Find out what really motivates them and drives them and sort of inspires them to work in the space that we work in.

47:56And um the k the key reason we're doing this actually is because it suits the

48:00the the timeline of the show where we go into October, November, and December and we have this very compressed period where we we get together as a community.

48:09uh v in in a sort of very short amount of time.

48:11So we thought we'd try and capitalize on that and try and sort of um flip the focus of the show a little bit.

48:18So I think going forward next year you'll probably have

48:20you know, 10 bytes, 10 bits, and then towards the end of the year you'll have analog uh sort of shows.

48:26And that that will sort of see us off nicely for the for for the end of the year.

48:30That that sounds good.

48:31Yeah, I'm really excited for analogs.

48:32I think we're we're we are as you as you say we're entering September, October, or Tektober, as as Marcus as Marcus Brownley would say, right?

48:42Three weeks from the iPhone launch.

48:44Exactly, exactly.

48:45And if if you live your life between iPhone iPhone and MacBook launches, um or Tableau Conference to Tableau Conference or whatever it might be, uh this is this is the time when you start gearing up and cracking your knuckles to be like, hey.

48:57Let's let's get something cooking here.

48:58Absolutely.

48:59Um so yeah, it's very very exciting times, very exciting times ahead.

49:02Um and yeah, I'm looking forward to the um analogue, but I think before that we've got a couple of bits and bites to come.

49:09But yeah, like keep keep tuned and yeah, keep keep sending us feedback.

49:12Absolutely.

49:13Um it would we would love if you could take a couple of minutes to leave us.

49:17uh and a nice little star um on on iTunes if you're listening on iTunes.

49:23Um uh the reason being is that uh in in the last couple of months a lot has changed in the podcast world hasn't it Ravi and

49:30Um reviews now matter a little bit more than I think they used to in the past.

49:34So in order for us to kind of stay relevant, uh we'd love to know what you think of the show on places like iTunes.

49:41If you like the show, you know what to do.

49:43Um, you know, give us

49:44As many, many likes as uh you can.

49:47Um and if not, then give us give us feedback.

49:50We love feedback.

49:51Um, we're going to be opening up more ways that you can give us direct feedback about the shows and topics that you want to hear about.

49:57So

49:57Um sorry to do a shameless plug there on reviews, but yes, um if you have a moment, just just just hit the start button.

50:05Yeah, exactly.

50:07At least at least four stars.

50:08Um but yeah, no, uh I think I'm gonna leave it with a um

50:11something from the Peter Cratch podcast where he's like pass the pod, you know find someone to pass the pod onto.

50:16Pass the pod.

50:17Right, right.

50:18And then and then and then sort of goes from there.

50:20I never thought I'd be seeking advice from Peter

50:22crash but here we are right exactly take it easy mate nice one take care see late guys bye