Is Open Source Coming back to your Cloud?
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
Title |
| |
Title of Series | ||
Number of Parts | 542 | |
Author | ||
License | CC Attribution 2.0 Belgium: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
Identifiers | 10.5446/61660 (DOI) | |
Publisher | ||
Release Date | ||
Language |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
|
00:00
Cloud computing2 (number)Open sourcePerspective (visual)Computer animation
00:49
Open sourceStaff (military)Open sourceDecision theoryDatabaseMultiplication signSinc functionSpecial unitary groupQuicksortBuildingSpacetime
02:15
Open sourceAxiom of choiceOpen setEquals signOpen sourceAxiom of choice2 (number)Term (mathematics)Data managementComputer animation
03:02
Open sourceClosed setOpen sourceBitAreaSlide ruleOperating systemSoftwareCloud computingServer (computing)Pairwise comparisonComputer clusterArithmetic progressionMultiplication signSoftware developerReal numberBasis <Mathematik>32-bitTwitterComputer fileBefehlsprozessorState of matterStack (abstract data type)Software frameworkRight angleApplication service providerWindowPoint (geometry)Diagram
07:30
Open sourceCloud computingBackupOpen sourceOpen setVirtual machineComputer animation
08:14
Strategy gameSpacetimeDatabaseOpen sourceAdditionGraphical user interfaceComputer animation
09:11
UsabilityOpen sourceView (database)Software industrySoftware developerComputer animationDiagram
09:54
SoftwareSoftwareSoftware engineeringKernel (computing)Process (computing)Goodness of fit
11:12
Open sourceCloud computingOpen sourceComputer animation
12:11
Scaling (geometry)PlastikkarteOpen sourceCloud computingCloud computingScaling (geometry)Marginal distributionData managementDatabaseService (economics)TwitterInstance (computer science)Server (computing)Mathematical optimizationIntegrated development environmentModeling languageMathematicsBefehlsprozessorOpen sourceElectric generatorNeuroinformatikDifferential (mechanical device)Computer animation
15:25
1 (number)DatabaseCloud computingOpen sourceEnterprise architectureSpacetimeComputer animation
16:38
Open setFocus (optics)Open sourceCloud computingOpen sourceData managementBit
17:47
Cloud computingFreeware
18:32
Cloud computingInclusion mapMarginal distributionCloud computingCASE <Informatik>Vector potentialLink (knot theory)Enterprise architectureTwitter
20:18
Declarative programmingRevision controlCodeOpen sourceRevision controlCodeQuicksortDeclarative programmingGraphical user interfaceCloud computingPoint (geometry)
21:00
ArchitectureCloud computingSelf-organizationAxiom of choiceOpen sourceAddressing modeAnalogyUtility softwarePower (physics)Open setCloud computingCartesian coordinate systemFront and back endsData storage deviceInterface (computing)Game theoryMultiplicationAxiom of choiceBuildingProjective planePortable communications deviceGoodness of fitInternetworkingCASE <Informatik>Open sourceIntegrated development environmentSoftwareVisualization (computer graphics)NeuroinformatikMassServer (computing)Internet der DingeCategory of beingDistribution (mathematics)Graph (mathematics)ResultantService (economics)Variety (linguistics)DatabaseMomentumForcing (mathematics)Presentation of a groupIndependence (probability theory)Slide ruleType theoryComputer animation
30:00
Hill differential equationOperations researchCloud computingOpen sourceOpen setWordINTEGRALProjective planeVideo game consoleCartesian coordinate systemUsabilityGroup actionTerm (mathematics)System callDatabaseOpen sourceSoftwareComputer programMultiplication signGame theoryLibrary catalogCASE <Informatik>Expected valueCategory of beingSoftware maintenanceComputer hardwareRight angleProgram flowchart
33:57
Open sourceSlide rule
34:31
Open sourceCloud computingWordOpen setTerm (mathematics)Open sourceFocus (optics)SoftwareWordCloud computingOpen setTwitterTerm (mathematics)Variety (linguistics)Projective planeSelf-organizationCore dump
37:34
Open setCloud computingSoftwareSoftware testingSoftware developerDifferent (Kate Ryan album)CASE <Informatik>WorkloadStack (abstract data type)Portable communications deviceOpen sourceNumberFreewareSubsetExpert systemSet theoryLevel (video gaming)
42:05
Program flowchart
Transcript: English(auto-generated)
00:06
So welcome to the second talk of today in the Sovereign Cloud Dev Room. And I'm very happy that Peter, founder of Percona, is with us. And this is now your stage, Peter. Welcome. OK. Thank you.
00:21
Thank you. Oh, well, that's fantastic. OK, well, so we're going to talk today about wonderful topic of open source and its relationship with the cloud. Now, I think most of you don't know me.
00:44
And I thought I would highlight my perspective first, right, so you can understand maybe where I'm coming from better. Myself, I was involved in open source since late 90s, which now seems like a long time ago.
01:04
I was at that time building my first startup back in Russia. And I took unconventional decision to use open source rather than stealing proprietary software, which was the most conventional thing
01:20
to do in Russia in late 90s. Afterwards, I joined MySQL AB before Oracle, before Sun. That was like a tiny company of about 40 people. And that's where I got a lot of exposure in building and sort of promoting open source.
01:43
And I also was their founder, and until recently, CEO of Percona, a company which specializes in open source databases. And we release all our software as open source. And since that time, I was also
02:01
involved as advisor, mentor, investor, in some case, in various companies around open source space. If you want to summarize all that kind of presentation, couple of words, is probably this. I believe what everything being equal,
02:22
open source is a better choice. Of course, you tell me it's never everything equal, but then I think this second thing applies. I think as a member of community, employee, maybe manager of a company, it is always better for you
02:42
long term to invest your time, invest your resources into the open source rather than just send that to some proprietary vendor which will provide more value to them than to you. Now, one thing to understand about open source
03:03
is that the open source takes time. This is a slide I shamelessly stole from a talk from Bruce with this last name nobody can pronounce. So I won't ever try it. But this is the Bruce in PostgreSQL community
03:21
who often talks about PostgreSQL and open source. And he uses this chart which says, hey, you know what? If you look at compared to the proprietary software, open source tend to take a longer time. Because if you think about saying, hey, you know what? I get one kind of get a small team together, fund them,
03:48
and say, hey, build this. They can do it faster compared to open source, which is kind of a lot more organic, fluid, a lot more discussions, a lot more people
04:00
operate on a voluntary basis. Each takes time. But then over time, it overtakes the proprietary software according to many criteria in many areas. And as example which I would use is this, like Linux and Solaris.
04:23
When I was getting started with open source, I have a lot of friends which are running a real Unix operation system. You know, Solaris, HPUX, AIX, anybody remember those? And they would think like, well, Linux, that is a toy.
04:44
That's like a 32-bit operating system. At that time, it doesn't handle multiple CPUs well. You could not even at the time have a file more than 2 gigabytes in size. I mean, it sounds like a joke and rightfully so.
05:03
But well, where are we here now, right? I mean, how many of you are guys still running HPUX? Anyone? Or like AIX, Solaris probably? Maybe there's only one or two people, no? No Solaris folks? Well, we can see what over time,
05:22
Linux with its kind of slow and messy at times development progress really overtook that their other operating systems, especially if you look at the server side, then Linux is well-dominating.
05:43
Now, we can also look, if you look at the cloud, we can see another comparison here. I remember also back in those early 90s where folks would be using ASP.NET with Windows stacks. And wow, it is so fantastic if your graphical user
06:02
interfaces, whatever, compared to us open-source folks who would be coding PHP or Perl in VI, or Emacs, if you are from the other camp in that time,
06:20
which was very different. But if you look at how things developed right now, now we have a fantastic development frameworks, very efficient for many open-source frameworks. That thing's evolved. Now, at the same time, if you look at the cloud standpoint,
06:43
we are kind of in a similar situation. You can, if you go to this kind of very polished, integrated, yada, yada, but very proprietary stacks which Amazon or many other proprietary, huge cloud vendors are ready to provide you,
07:02
or you can have open-source solutions which are not yet as polished, powerful, integrated, and so on and so forth. But I believe, well, they are going to get there. Now what I want at this point to talk a little bit more
07:21
about where we are now in kind of a state of ecosystem, what is happening, and what the trends are. One, if you think about relationship with clouds and open source, you can say, well, from one side, the cloud vendors have been making the open source a lot easier for other people.
07:44
If you look at deploying some open source technology, especially it applies, I would say, to a database, it's much easier to get, let's say, Postgres running in a cloud vendor compared to getting your kind of highly available,
08:03
monitored backup or PostgreSQL cluster on virtual machines. But on the other hand, we can also see all those cloud vendors use this kind of an old strategy, which actually they used even before the cloud, which
08:21
is called embrace, extend, extinguish. And what that means is, hey, why don't we take the open source, we make it fancier. Fancy may mean more features, like in a database space. Well, where I come from would be something like Amazon Aurora. It's wonderful. There are those kind of additional features, better
08:40
performance, or maybe more usable. Maybe there's like a fancy GUI so you can deploy it with a couple of clicks instead of have to put some elbow grease to do it the open source way. But then, yes, you go there, and you kind of check in in the heart of California,
09:01
which if you know the song, you can't really ever live again. Now, what I think in this case, it's important to understand and for all of us as open source developers to recognize what really the usability and ease of view is increasingly becoming that is a new frontier
09:24
of the competition. And I think that is especially hard maybe for people for my generation, which got used to that very kind of hairy open source. At that time, we remember compiling open source
09:41
from source, and kind of we have all that skills. But if you think what is happening right now in the software industry in general, as well in open source is this. One is a software just absolutely explodes in amount.
10:03
Everything is software those days. Amount of software which is being written, will need to be written, need to be maintained, extend, and so on and so forth is absolutely amazing and growing every year. The same happens with data. We're having more and more data and more and more ways
10:20
to process and wherever that data for good and for evil. But at the same time, we have to do that with less. And especially with skills. You can see as a software engineer and technology industry as well,
10:41
we have a lot of people getting in that stuff which are not going to really be able to debug Linux kernel code, try to have some just obvious basic skills, right? Well, for them, you need to build a software
11:05
which is easy to use. The fact I think what is very interesting is what we can see increasingly some expert, even in the related to open source technologies, which actually only have cloud-specific skills.
11:26
Well, especially in the US where I think adoption of a cloud is substantially forward alone, you may have somebody saying, I am post-greSQL expert, right? And you talk to them what that means. He is very capable of provisioning post-greSQL Aurora
11:45
on AWS with two clicks, right? But you say like, well, you know what? What is Patroni? Yes, I say, what, what? Well, that's like some high-ability solution for postgres, right? If you guys don't know, right? Which is very basic.
12:01
Well, I think another interesting thing is what, especially in the large corporations standpoint, a lot of them run to build their businesses by the book. So if a management made a choice, what we are adopting, let's say Amazon cloud
12:21
or Google cloud, Azure, doesn't matter. And they often then would come to them and say, okay, how do we do that? Give us the book, right? And what are those cloud vendors going to tell? Oh, they're not going to tell you, well, you know what? Use our basic services which you find in our cloud. They are going to tell you to use
12:41
where the most property, advanced managed solution, right? If you read with Amazon, they'll say, hey, you know, go use DynamoDB, right? It's awesome, right? Or it's Amazon Aurora, Redshift, all that kind of stuff which makes you the most sticky for that cloud
13:02
and which is also providing the most margin to those folks, right? So what is also interesting in this case, right? If you look for Amazon Aurora, for example, as they see their customers getting more and more locked in, the price differential between your computer sources,
13:24
right, you know, storage, right, and essentially like any other software, it becoming larger and larger, right? If every new generation of CPU can actually, you know, find that excuses, right? So find that changes.
13:42
So that means it becomes more and more expensive over time, right? And also, you will see in this environment also really embracing scaling by the credit card, right? Because scaling is easy, right?
14:01
You don't have to figure out how to run that server you have because getting the next one, the bigger one, would take three months, right, even if you get a budget approved, right? Like we used to have like 10, 15 years ago. Now, you know, screw that optimization skills, right? You can just go to the larger instance size, right?
14:20
And that's easy. And again, like from what we see from a database standpoint, a lot of databases in the cloud are optimized very poorly exactly because of that trend. Now, what is interesting for me also is how that plays with their venture-funded open source,
14:43
which is a lot of open source those days, of course, right? And what we can see is what they are really, well, protesting against unfair competition from cloud vendors,
15:00
yada, yada, right? And many of them abandon open source licenses, right, and go to something else, right, which is, well, really shifting to a proprietary software
15:20
models. Now, this is, in my opinion, is very different from the community-driven open source, right? If you think about, well, compare something like PostgreSQL to Mongo, right, which I think in the database space are very opposite on that kind of open source embracement specter, right?
15:42
The PostgreSQL really has benefited a lot from a cloud from adoption standpoint, right, because cloud made PostgreSQL much easier for a lot of people, right?
16:02
But at the same time, of course, now there are different players which dominate the ecosystem. Yes, it used to be Enterprise DB was the biggest dog, right? Well, now it is probably Amazon, right? And other cloud vendors make more money,
16:20
but you know what? We are hiring a lot of PostgreSQL engineers and investing a lot still to make a PostgreSQL better. So at large extent, PostgreSQL community still benefits in this regard. What also has been interesting for me in the last decade
16:41
or so is the idea of growth at any cost, right? For a lot of companies, you can find, well, at least in the US stock market, you know, the more money company is losing, the better its valuation, right, as long as you can grow user base as well, right?
17:03
And in this case, you really gravitated saying, how can I grow faster? That means focus my resources. And cloud, which often allows you to, well, things simple, though expensive, was very much adopted, right?
17:24
And for those companies often didn't make sense to invest in efficiency, which I think open source often brings ease, right, even though it requires a little bit more investment and elbow grease kind of and management attention to make those things happen.
17:43
Even more, though, you can see the cloud has been quite good at giving you your first shot of heroin for free, right? You can find what every cloud out there saying, hey, oh, your startup, great, will give you some big chunk of money so you can spend on us, right,
18:05
or given the free tea, right, and whatever solution. But I guess with this analogy, you understand, right, well, if somebody giving you a first shot of heroin for free, they may not always have your best interest
18:24
in mind, right? Now, I think what is interesting is that there are some revelations which are coming, I think, in this case. And things are changing.
18:40
And I will provide a bunch of links to the detailed articles here, right, which you guys can check when you saw the cloud. This is interesting from this trillion dollar paradox, right, which talks about, well, you know what? Cloud is damn expensive. And for many cloud companies, the very large portion
19:03
of their cost of revenue is actually cloud spent, right? So if you are choosing to go with that cloud, especially most expensive versions of that, well, you know what? You have to be giving up on a lot of potential profit margin. Recently, you may also heard this 37 signals,
19:24
which talk about why they are leaving the cloud and they detailed that stuff. That is also a very influential company, right? It's not very big dollar amounts compared to enterprise here, but I think it's very nice what they kind of went public in this case and say, hey, guys, even in our scale,
19:44
even if you spend maybe less than $10 million a year on the cloud, it makes sense to actually do things differently, right? And $10 million may sound like a lot of money to some of you, but if you look at even kind of like a mid-size and, well,
20:04
especially more like larger corporations, they often spend hundreds of millions on dollars on the cloud spend. Now, another trend which I think is interesting, which is coming out, I think is GitOps
20:21
and generally the approach is where we have declarative version infrastructure as a code approach, right? Because when you do that, then often those kind of fancy GUI point and click I can do in the cloud
20:42
is not exactly how you do things anymore, right? That's where we kind of sort of do it more of, well, through API, command line, if you will, right? Things there. Open source tooling is already better.
21:00
Hybrid cloud is another interesting trend, right? And what I want you to see from this graph is what if you look at the larger the corporation is, the more of them are embracing the hybrid cloud, right? That's from the CNCF poll, right?
21:22
And if you are using public, if you're public and private cloud together, then you probably do not want to rely too much on those property services, right? Because they tend not to be as portable.
21:44
In the end, I think what you are, what we see in the open source kind of catches up again, leaving us with kind of two choices here. If you look at the AWS, which is obviously the biggest dog, where you can replace that with Azure, GCP, right,
22:03
or some other major cloud, you can really lock in the approach of a cloud vendor, right? Or you can use their cloud native foundation stack, right, among others, right,
22:20
where you are really treating cloud as a commodity and building the value through their open source software. What I think is interesting from a cloud native environment is how big it is, right?
22:40
You guys probably cannot even see all the logos, right, on this slide, right? There is a lot out there. There is typically more than one solutions for even for a single problem, right? And some of you say, hey, that is too much, but that is also the open source way, right? In many cases, we need in open source multiple solutions
23:03
to the same problems, right, which will kind of evolve, right, and maybe some people will die off, I will join forces together and that is how great stuff happens, right? But I wanted to show that just to illustrate how much momentum there is behind their open source approach.
23:24
Now, I think in this case, what that allows us is to really, given the property cloud, it's originally intended role of commodity infrastructure. Why am I saying originally intended role?
23:41
Well, you can see this, this is actually the slide I took from some very old AWS presentation. Then they're just bringing the cloud to the masses and say, hey guys, you know what? You don't usually run your own generator at home. It's not convenient. Well, yes, in some cases it makes sense, but mostly you buy electricity.
24:02
So you should buy the same from a cloud when it comes to compute resources, storage, yada yada. It will make sense. But here is a key difference, right? If you buy the electricity, it is commodity. You can buy the electricity from a variety of members
24:22
or make your own, and you see all appliances going to work, right? You don't have to change your TV or your fridge if you're doing the other electricity, right? Well, and that is kind of what a lot of cloud vendors want us to do, right? Well, we'll provide you commodity thing, electricity,
24:41
but you know what? We also really want you to use those appliances which will only work with our kind of electricity. Well, that sounds like bullshit to me. So with this, I think we have really two choices, right? How we can approach the cloud,
25:01
which I will call as a cloud of software on account of freedom. One is to use a lot of property features of your cloud vendors, sell your soul to the devil, right, and all the other stuff, right? So let's vote. How many one is going to use that way? Any hands out there? No, no?
25:20
I would expect there's some people working for Amazon here on some. I would say, yes, yes, okay. And then we can use cloud of freedom, which is saying, well, cloud vendors are fine. It's like a good commodity infrastructure providers, like the internet providers, right? Just give us those commodity resources
25:42
and we will build the rest of the stuff using open source. How many of you think that a good idea? Okay, okay, that's better. Okay, now in this regard, what I think is a very good API in this regard?
26:03
Well, it is there, the Kubernetes, right? I would not say that Kubernetes is absolutely perfect. It's not, but I think what every successful open source project started as a piece of crap, right? I mean, it's just how things go, right?
26:21
Because you build successful open source projects, not having a solid technical foundation, despite what engineers want to believe, but building the movement, building the ecosystem, right? And then those ecosystems, right? Well, is able to redo the code, right?
26:45
Mistakes, whatever, and to build something, right? With Kubernetes example, I remember there was messes. Anybody remember messes? Well, that was a very cool project, right? I remember talking to those girls and they have, let's say five years ago,
27:02
a lot of very cool things which Kubernetes did not have, but you know what? They lost adoption game. And now you can see the things like I have next, oh, no, sorry, next. Their container storage interface and some other stuff. Hey, you know what?
27:20
Kubernetes is getting a lot better for data intensive applications, right? Not just for this narrow use case of a stateless applications it was originally designed. We also now have either the data on Kubernetes community, right, which is a special community which is focused
27:42
on running data intensive applications on Kubernetes, right? And what is interesting in this case from also the CNCF's survey, we can actually see what from 2001 to 2002
28:00
where the databases, right? So that something which is technically was not something you would ever think to run on Kubernetes five years ago is one of the fastest growing application type which is being run on Kubernetes those days.
28:20
Now, I would say even more, what we see right now is if you look at a lot of modern independent database as a service providers, right? All of those folks, they are actually using Kubernetes on the back end to build their database as a service solution.
28:41
You may not know that, you may not care but they are doing that and that means for them, well, that is both efficient way to build it, operate it and that can be run stable enough like at least with a good skill. Another interesting result what's going on with Kubernetes
29:05
is what we are having, well, despite having Kubernetes, already quite ubiquitous, right? Hey, you can run it on pretty much any cloud. If you look at many, every distribution,
29:20
a little distribution now have their own Kubernetes destroyed, at least in the major one, right? You steal the Kubernetes from VMware, right? Then, and many other solutions, right? But we are also having a Kubernetes coming to other places. Like I like this K3S solution which is lightweight,
29:44
a Kubernetes which you can deploy on the edge IoT devices, right? Or if you just have couple of servers in your basement that you want to run Kubernetes on. We also are getting more solutions for visual kids.
30:05
Because if you look at like a raw Kubernetes, you often would be seeing that console and YAML files, right, and everything. But now we are also have a pretty good dashboard which comes with Kubernetes project.
30:21
And as well, integrations with applications. Like there are solutions like Kube apps or Rancher application catalog, right? So you can say, hey, you know, if I'm running Kubernetes, I can deploy open source and the appropriate applications
30:42
on my Kubernetes solution in a couple of clicks. Kind of similar to what we see now in their major clouds, right? Where we have their marketplaces. Okay, so with that, where do I think things
31:06
need to be going and will be going next as we are getting more and more open source software in a cloud? One, we still need to continue working on the better integration. Because I think that is the name of the game right now.
31:24
We expect really all the software and the software and hardware to be very well integrated with each other, right? And that is something where I think a lot of property vendors put a lot of folks in, right? Because if it's not easy to make two things work together
31:44
then for majority of the people it would not be accessible. Because remember what we spoke about? A lot of people do not have very advanced skills those days. Related thing of course is their better usability, right?
32:02
I mean, we have expectations of usability very high. Attention span, very low, right? I mean, I'm not sure. Any of you have like teenage kids? Or am I just too old here? What do you think about the attention span those days? Okay, well, and remember, right?
32:23
Those teenagers are probably already doing some work with computers, right? It's often taught as early as middle school those days, some programming skills, right? And they will be going in the industry the next few years. And the last thing I would mention, and that's again maybe comes from my database background,
32:45
is making sure we also have a good experience if a day two operations, not just day one. In very many case I see there is a lot of application catalogs and where I say, hey, look at that. You can develop that application very quickly, right?
33:01
But then if you really want to maintain that, it's like with database you can't just tear it apart and deploy a new one. Well, that is something which can become like instantly quite complicated, right, if it's not handled well. Well, with that I will end with my call to action.
33:24
It's pretty much same what I started, right? I would encourage you to embrace the open source in the cloud, also in desks, both in terms of your personal time and efforts and as well more like in your company to invest in making it better
33:43
because open source very much depends on our collective effort, right, for its success. And of course, spread the world. That is how the movement grows. Well, with that, it's all I had and I would be happy to answer some questions.
34:24
Thank you very much for your talk. You've said in your final slide that people should be investing in open source and invest in making it better. What are your thoughts on the new kind of trend you see of organizations like the Sovereign Tech Fund
34:40
in Germany allocating millions to open source projects and things like that? Where do you think the money should come from rather than just, or is there a structured way it should actually be allocated? Well, so where should money come from, right, from open source? Well, look, I think that is a variety of things, right?
35:00
I think if you look in this case, like from the government standpoint, right, it's classically have been supporting public good with investment in research, right, the medicine, right, wherever. I think that open source is one of such ways how you can create a lot of public good, right,
35:22
which is wonderful. I think, though, in many cases, and probably the majority can come from, right, these are the public corporations, or not public, private corporations, right, acting out of their own interest, right? I think in this case, we all need to also work
35:41
with that kind of like mentality change, right, because you see like for years, right, even before like Microsoft changed it and became like all, you know, we love open source kind of company, right? It was all about like, oh, open source is evil. It's like, you know, communist run open source, right?
36:00
That's how you, you know, lose all your, you know, property, right? Open source is insecure, yada, yada, yada, right? And I think that is something there that's kind of spread the word is important, right, is to make sure, well, that is actually none of those things, right, and really embrace an open source
36:21
and the focus on the real open source can be very good long term ROI for a corporation. So maybe not the short term, right, but the long term. Well, in this case, I probably want to mention one other thing, which I think there is still challenge those days, right? I think what open source term is being bastardized, right?
36:44
We can see many companies, they try to say, oh, our software is kind of almost as good as open source, right, but then they put you like, well, some very specially restrictive licenses, right, in this case, right, and we need, right,
37:01
or have like an open core when you have like a crippled open source, right, and real property software, and we need to make sure the company understand there is, what is open source, right, and focus on growing that compared to those, well.
37:20
Others. Make sense? Thank you. Hello, thank you for the presentation. In the beginning, we started with a good willingness to make it very portable to other cloud and that kind of stuff, having the deployment that can be multicloud, and we missed something because at some point,
37:44
the experts that were there to help us were pushing for some proprietary or non-compatible deployment, and we are totally stuck now in one cloud provider. How do you manage such a risk, and what would you put as a whole set at the beginning
38:02
so that you would not fall in that pitfall? Yeah, well, I think that's a very question. I can say, hey, we made a mistake, and now we are stuck, right? Well, I think in this case, right, if you're looking at portability, right, I think it's important what that portability is is being tested, right?
38:22
So in your case, right, if you say, hey, we are mostly around that on that one cloud vendor, and we want to be able to move to another cloud vendor if we need, well, that means what you need to regularly test your software that actually works on that cloud vendor, so if incompatibility and over-reliance is introduced,
38:42
you can instantly squash that, because if you don't do it, right, then like three years later, right, then you finally want to do that, you can say, oh my gosh, this actually turns out so much things what we did not do, right? And I think in many cases what is also important is what developers do not read docs, right?
39:04
They tinker, right, and often then you think like, oh, I'm using this kind of feature, right, in this case, I assume it works everywhere, but maybe they are actually using something which only works in this particular place, right?
39:20
That is why testing is important, you cannot rely on the developers, right, or infrastructure engineers just using the subset they are supposed to be using, okay? Well, thank you.
39:41
Oh, one more, okay. Thank you. When we have some companies using proprietary cloud provider, it is difficult for us to say that they need
40:02
to move all their infrastructure, as you just said. So is it a problem of offer, it is a problem of the free software cloud provider offer to provide some easy tools to use,
40:21
or it is really a problem of politics from the companies to say that, yeah, we should go to the difficult, but free software, how do we manage with that? Well, I mean, you're asking like, is that like a technical question, right, or is it politics, right?
40:41
And the question is both, right, in this case. Like, yes, I mean, if you look from a technical standpoint, I think they're easier is if open source software, where more is going to be, you know, easy to convince, adopt, right, and so on and so forth. But in many cases, it is possible.
41:03
It is also like a politics, right, because there can be some entrenched interest for whatever reasons, right, or some belief, right? I have seen a number of companies on my watch, right, where people say, oh my gosh, we are running on, let's say that like a Microsoft stack.
41:21
There is no freaking way it's possible we move to open source, yada, yada, yada, yada, right? And then we have a new CTO, right, and boom, right? You can say in a few years, right, like more than half of workloads are moved to the, you know, to different open source solution, right?
41:40
So I think in this case, like if you look at what is a primary thing, right, it is politics and desire of management, right? That is always primary. Thank you, Peter. Okay. Next on stage will be Fabio.