How Convenience Is Killing Open Standards
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 | 49 | |
| Author | ||
| License | CC Attribution 4.0 International: 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 | 10.5446/51762 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
FrOSCon 2020 Cloud-Edition18 / 49
1
2
5
8
9
10
13
19
20
21
22
23
25
26
27
29
31
32
33
34
35
42
44
46
48
00:00
Cloud computingStandard deviationOpen setInternationalization and localizationCore dumpOpen setStandard deviationCellular automatonBitGoodness of fitOpen sourceProduct (business)Event horizonTouch typingTwitterBasis <Mathematik>Squeeze theoremNumberWave packetTrailCommitment schemeComputer animationXML
01:09
OracleOpen sourceBit1 (number)Point (geometry)Wave packetOpen sourceOraclePerspective (visual)Process (computing)Information technology consultingSystem administrator
02:33
WeightSemiconductor memoryCurveInheritance (object-oriented programming)WaveTerm (mathematics)Panel painting
03:24
Business modelLevel (video gaming)Different (Kate Ryan album)Disk read-and-write headProjective planeNP-hardSoftwareSpecial unitary groupRadical (chemistry)Physical systemCASE <Informatik>Multiplication signWindowGame theoryVirtual machineProcess (computing)Transport Layer SecurityPoint (geometry)Network topologyComputer animationMeeting/Interview
05:07
Physical systemWeb 2.0Projective planeMultiplication signComputer architectureService (economics)Identity managementComputer fileInformationNational Institute of Standards and TechnologyCentralizer and normalizerFilesharing-SystemSoftwareNetwork topologyExtension (kinesiology)DatabaseScaling (geometry)Proxy serverLattice (order)Server (computing)
05:55
Lattice (order)Physical systemSoftwareSuite (music)Web 2.0DatabaseMultiplication signProxy serverComputer architectureLimit (category theory)OracleSpecial unitary group
06:38
BitStreaming mediaSoftwarePatch (Unix)Open sourceRevision controlMultiplication signOperating systemNumber
07:39
Standard deviationGame controllerComputer hardwareSoftwarePoint (geometry)Standard deviationOperating systemMultiplication signComputer animation
08:13
Physical systemStandard deviationOpen setMultiplication signOperating systemMathematicsOrbitPoint (geometry)Impulse responseSymbol tableComputer animation
08:48
Multiplication signSoftwareComputer configurationComputing platformPoint (geometry)Open sourceFeedbackSource codeWeb 2.0QuicksortServer (computing)1 (number)Physical systemSpacetimeProduct (business)CuboidEndliche ModelltheorieDenial-of-service attackKernel (computing)Module (mathematics)Computer animationDrawing
10:29
Limit (category theory)Multiplication signOpen sourceMeeting/InterviewComputer animation
11:05
Multiplication signOpen sourceComplete metric spaceOcean currentResultantCASE <Informatik>Mathematics
11:43
Open setCASE <Informatik>Game theoryInformationOpen sourceBasis <Mathematik>Physical systemWhiteboard
12:34
Cloud computingLogicExplosionNumberCloud computingPhysical systemStandard deviationInternet service providerBitOpen sourceGoogolTwitterTraffic reportingOrder (biology)Service (economics)Suite (music)Hybrid computerComputer animation
13:32
Standard deviationOpen setCloud computingOpen setStandard deviationOpen sourceInternet service providerChainRouter (computing)Interface (computing)Gateway (telecommunications)Set (mathematics)Integrated development environmentSelf-organizationPhase transition
15:01
Right angleBit rateData storage devicePhysical systemInternetworkingBitComa BerenicesLogic gateService (economics)Enterprise architectureGateway (telecommunications)Meeting/Interview
16:00
OracleCloud computingExplosionRight angleScale (map)State of matterBit rateCloud computingConnected spaceEnterprise architecturePerspective (visual)State of matterComputer animation
16:46
Standard deviationGame controllerReal numberInstance (computer science)Internet service providerOpen setCloud computingEndliche ModelltheorieLimit (category theory)Data managementDecision theoryBlogForm (programming)Internationalization and localizationFormal languageIntegrated development environmentMultiplicationConfiguration spaceStrategy gameLatent heatEngineering drawing
18:33
Open sourceThomas BayesCloud computingInstance (computer science)Open sourceOpen setElasticity (physics)Stack (abstract data type)Design by contractGoogolTrailInterface (computing)Software
19:24
Internationalization and localizationStreaming mediaSoftware as a serviceLatent heatModule (mathematics)Enterprise architectureGene clusterOpen sourceSoftwareBusiness modelOpen setMathematicsStandard deviationInstance (computer science)System callSocial classEndliche ModelltheorieGoodness of fit
20:58
Open sourceStandard deviationBusiness modelCloud computingOpen setStandard deviationProjective planeMathematicsOpen sourceStreaming mediaGame theorySoftwareLecture/Conference
22:19
Server (computing)Computing platformService (economics)Enterprise architectureFunctional (mathematics)Internet service providerExterior algebraOpen sourceCartesian coordinate systemData storage devicePosition operatorVirtual machineMathematicsQueue (abstract data type)Video gameData conversionParameter (computer programming)Real numberComputer architectureSelf-organizationRow (database)
24:34
Data conversionParameter (computer programming)Real numberVideo gameData centerComputer animation
25:14
LoginService (economics)Computing platformFunctional (mathematics)Office suiteProcess (computing)Decision theoryPerspective (visual)Parameter (computer programming)
25:59
Parameter (computer programming)Cartesian coordinate systemSpacetimePhysical systemOpen setCloud computingOpen sourceArmWater vaporCode
26:39
Mathematical optimizationState of matterCloud computingVirtual machineNumberBasis <Mathematik>Business modelEnterprise architecture
27:12
BitAreaDifferent (Kate Ryan album)Product (business)WindowPhysical lawInternetworkingCommunications protocolCloud computingRight angleGame controllerOvalIncidence algebra
29:06
WordWater vaporData structureComputer programmingPower (physics)Limit (category theory)
30:04
Limit (category theory)Power (physics)Single-precision floating-point formatPhysical systemGame controllerDifferent (Kate Ryan album)Perspective (visual)MereologyGame theoryCloud computingState of matterArithmetic meanComputer animation
31:29
Cloud computingPresentation of a groupComputer clusterBlock (periodic table)Point (geometry)
32:22
Point (geometry)Cloud computingMathematicsGame controllerVirtual machineWave packetComputer architectureEnterprise architectureCartesian coordinate systemPhysical systemTwitterFigurate numberArithmetic meanComputer animationMeeting/Interview
33:41
Control flowGame controllerCloud computingHuman migrationInformation technology consultingEnterprise architecture
34:33
Physical systemVirtual machineMixed realitySoftwareService (economics)Internet service providerProgrammable read-only memoryData centerPerspective (visual)Prisoner's dilemmaBit rateIntegrated development environment
35:47
MultiplicationNumberService (economics)Internet service providerDatabaseEnterprise architectureInheritance (object-oriented programming)Combinational logic
36:33
Parameter (computer programming)Execution unitBitPlastikkarteSoftware developerInternet service providerState of matterPay televisionInternetworkingOpen sourceExistenceMereologyGame theoryStandard deviationVariety (linguistics)
37:47
Open sourceCloud computingEnterprise architectureInstance (computer science)Pay television
38:23
Universe (mathematics)InternetworkingResultantField (computer science)Diagram
40:00
Process (computing)Online chatData managementBuildingInternet service providerMereologyExtension (kinesiology)Goodness of fitProduct (business)Differential (mechanical device)Expected valueCloud computingDifferent (Kate Ryan album)Multiplication signStrategy game2 (number)Software developerDigitizingSymbol tablePlotterNetwork topologyCovering spaceCausalityNP-hardData structureWater vapor
43:55
Cloud computingJSONXMLUML
Transcript: English(auto-generated)
00:07
Welcome, Bernd, he's going to tell us a little bit how convenience is killing open standards. Yeah, good evening and thanks for the introduction. So yeah, my name is Bernd.
00:23
I give a little bit of introduction of who I am and what I do. So I'm working as CEO for a company named NetWays in Nuremberg and also for the open source monitoring product named Isinga. I'm heavily involved in the DevOps days movement.
00:41
Obviously, because of COVID-19, there are pretty much no events going on, actually, but usually I'm really tried to visit a lot of places and try on DevOps days there. Yeah, I do this talk in English. Obviously, I have a little bit of a Franconian accent, which is related to Nuremberg, of course.
01:02
And if you would like to get in touch with me, Twitter definitely would be the best one. So yeah, just to give you a little bit of my background where I'm coming from, which perhaps makes hopefully sense in some way for you, why the topic is important to me. So I started my IT career at the end of the 90s
01:23
as a Solaris admin at a company named Kvelle. The older ones of you know that still was a big retailer in Germany and Europe as well. At some point I came into the Oracle business, which primarily the reason was, I think I was the slowest person to hide when they asked for somebody
01:41
who's able to do the Oracle training. So I was into that and started doing Oracle. And I did that for over seven years, did Oracle consulting and traveled pretty much, yeah, Europe and Germany mainly. And then I ended up in an open source company, which is Netways,
02:01
which I'm still there for over 12, 13 years now. And so the different perspectives I had and the different industries, I think are one of the reasons why I'm in some way worried about where we're heading as an industry and pretty much also how my topic comes together.
02:20
So yeah, there's a pretty high chance I will retire there because I really like it there. It's still an overseeable team. And I really liked the job I'm doing and it's one of the biggest privileges you can have, I think. Yeah, so the motivation for the talk. So when you think about your own and think,
02:41
okay, what catches me when I think about this topic? I think one thing is that I'm old, which means I'm old, is 43 years old, which is perhaps not super old, but it's old enough that you see that our industry comes back to every wave and every hive every now and then. So we have a very short term memory in our industry
03:03
that we usually when a new hype curve and a new thing comes up, we tend to forget everything we learned before. So that's one thing when I'm looking back to, let's say now 25 years in that industry, I've seen a lot of things. I would not say that I understand everything,
03:21
but you get kind of an idea how the industry in some way works. Of course, I'm still curious. So I'm still, even nowadays, I'm not so much involved in the technical stuff from the project doing aspect, but I'm still very interested in technology, but also more on the level where the industry is going
03:42
and how different business models come together into that way. So I'm still very interested in trying to understand where we're heading. So of course, we would like to do something great and that is something never changed. So when I started my career in IT,
04:00
there was pretty much two vendors we were working with still in my shop at this time, there was also a big IBM shop for the whole host world. But we, at the time, we were the total newbies coming with Unix systems like Sun and HPX, we were for the host people, we were the very unserious play dudes
04:23
and they really didn't take us serious. And at the time we had pretty much Sun Microsystems and HPX, so hard end software. So I was mainly in the Sun team and we did big warehousing with E10 case and all that stuff.
04:41
It was at the first time when I saw that machine, I was very disappointed because it was just a terminal and coming from Windows because that is what I used for gaming and something when I was young. I was very disappointed that there are machines for millions of dollars and you pretty much see just a couple of text terminals. But anyway, it didn't take long
05:00
that it really catches me, I like the system. But we were very limited to the vendors we can choose from. So at the time the department I was working in, we had big goals to achieve. So we would like to start with a web shop at the time that was a new thing, a new project they would start on that systems
05:21
with Solaris and all that stuff. So we had a lot of stuff to do. And what we did is that we made a lot of architecture stuff, how we can build that system. So at the time there was no identity system, there was no central file sharing. So we came up with something like NIST Network Information Service at the time,
05:40
which was at the time was very cool. I don't know if somebody is still using this. So we make a lot of architecture designs, how we can grow our Solaris and also HPX landscape. But the promises which were made by the vendors we had, and that was one of the problems
06:01
that we had a very limited software pool which we can choose from. So every time we said we need a web server, or we need a database server, or we need some, I don't know proxy or whatever architecture we tried to play out. Every time I remember we were in a meeting with usually somebody from Oracle, or sometimes from some micro systems or any like software vendor at the time.
06:23
And that it was always the same thing. Like, yeah, of course we have some software. So we have some whatever suit which you can use. And at the beginning we were very happy with it. That's fantastic. We have something to achieve and there's a software. But even when we hoped so, this is where we ended up.
06:44
Going back to slides, so we plan it this way. It takes a little bit into the stream until the picture comes up. So we want to do it this way, but with the tools and the solutions they promised us, we usually ended up like in a total nightmare that the software didn't work
07:00
and we had to install some patches and worked around that stuff a lot to get it working. So it was really frustrating in some way because every time the supposed improvements made it worse. So every time when we get a new version of something, because of course we had to wait for it because it was not open source in some way.
07:21
So the companies promised something every time it was really worse. And I don't know if you are speaking German, you know that of course, for Schlimb-Besserung, that was the typical thing where we ended up. So every time we tried to improve it, it was even more horrible afterwards.
07:40
So it was super frustrating at some point because we were very limited. We had a very controlled hardware stack, but I have to say the hardware itself didn't make any problems. And also the operating system was not a big deal because it was still the POSIX standard and there was some kind of freedom, but then there was pretty much no software we could use.
08:00
And this was so closed and we had no impact and we couldn't see it and was in transparent, was very frustrating. So one of the things at the time came up and was almost there was that in somewhere there were standards made it possible to achieve more. And one thing is the POSIX standard, which was at the time and all the Unix systems over there in POSIX is defined,
08:24
for example, that you have an NBI in place, that you have SED or all these basic tools and fundamentally APIs of the operating system were defined in the POSIX standards. And it made it possible in some way to interchange and open something which was new to us
08:42
because it brought a change. Because what for us POSIX open up that the new tools come in place. And the new tools at some point where for me at the time and also for my colleagues, the first thing where we can use tools which were open source, or we can have a look on the source code,
09:00
it was done by somebody else. And even the feedback loop was way more faster in that community than we had with our million dollar software vendors. So that was super exciting that we get way more options to choose to build our platform. Also Linux started at the time.
09:22
So Linux started pretty close to where the company was in Ferdinand Bavaria. And the older ones, if you remember these big boxes with a lot of CDs. And I remember when our colleague at the time started with Linux and he pretty much spent weeks compiling the kernel
09:41
and including the right modules. And we pretty much treat him the same way like the host dudes treated us. And we're saying, this is really not a serious system. You will never make it to production. We ended up at the time that the Prokord Solaris system we bought to run the web shop.
10:01
It was pretty much running CT at home and listening to space for any other thing. And the production web server was running on a Linux box under his desk. So this is where we ended up. And this was also the way when I came into the Linux business in some way. So it was fantastic because if you remember that the old Linux CDs
10:23
you had so many tools, it was so option. So it was really mind blowing how much stuff you can do. And you had communities. So the first time for me at the time, you had the feeling that there's more than just a limited vendor offering. So there are other people fighting the same problems
10:41
and try to find the best solution for them. And it was really, it was also big diversity. It was just fantastic. So we were really motivated. I can remember that even in our team at the time and more and more Linux come into it and we were also allowed to use more open source tools
11:01
in some way was really fantastic. So time flies, a lot happened in the time. So open source is of course, more than ever. So where we are, I think in my opinion, and that, I don't know, short or 16 minute, whatever the history would like to set the foundation
11:22
for where I see where we are as an IT landscape currently. Perhaps it's not so worse than the pictures showing and we are not running completely. But means what happened is in the last decade, I would say that really changed the whole open source landscape.
11:43
Things like Microsoft loves Linux, which is, it's fine. It's great, but still going back 10 to 15 years, it's really hard to imagine that where Microsoft is now and they doing, I think in most of the business cases doing really fantastic, it was really hard to imagine.
12:01
So the landscape and the vendors and the big players in that game where we are now changed dramatically of course, because IT landscape changed dramatically. And I would never expect that if you ask me that 10, 15 years ago. So we have more open source than ever. If it is on GitHub or GitLab or wherever
12:22
your favorite system, we have a lot of open source. So open source is produced on a day-to-day basis. You can pretty much get information about anything, a lot of communities. We have a changing market, which is the cloud report is a little bit outdated, let's say six to seven months,
12:42
but what we can see the numbers in kind of so true. So what we see is that the on-premise business is gonna be reduced in a way that more and more companies and people go to a public, in some way perhaps hybrid cloud module, but of course it's not overseeable
13:02
that many, many companies move their business into an AWS and Azure, Google Cloud or other vendors. So that's the trend. And I think it's pretty much no surprise that this will not change in the next year. Next years, it would definitely surprise me, but I think the trend is clear that people more and more
13:23
give up on their on-premise IT infrastructure and move their services into any kind of cloud system. So where's the problem? So where's the problem in the thing that people move their business around and going to the public cloud providers?
13:42
I think one thing is that we often have a mismatch between open source and open standards. So the problem is that we have, like I said before, we have more open source than ever, but open standards really is a thing. And the problem is that they're often the mismatches
14:02
that an API is not an open standard. So most of the providers out there, they have the API. Sometimes their API became kind of a standard. Think about the S3, for example, which AWS in some way brought up. And then there are other things like you can have like a router's gateway for your set cluster and pretty much speak the same,
14:23
use it like in the same interface. So they are APIs became kind of standards, but most of the things we see out there are not standardized in any way that they have some. We have also in the cloud environment, you have standards like the OCI,
14:40
open cloud interface and all that stuff, but pretty much nobody really knows them or deal with them. So that's something which is really different. What we see now that companies promote, we have an API, but more and more everything they do behind is of course hidden behind that API. And of course the API is usually not an open standard.
15:01
So we have a very unbalanced market. And if you remember the first picture, which I had with Solaris and HPX on top, which is pretty much not a real picture. The picture you're seeing now is a real picture. So this is like Seattle Tacoma Airport
15:20
where employees of the name companies have primary access to the gateways. And this is the market where we're seeing right now that pretty much the big players in that business more and more put all the others on the side. But with that, that they control the majority of services and see it as like storage or compute
15:41
pretty much more and more companies and enterprises go into their systems. They also control more and more of the internet because they own the highway, right? If you drive on their streets, they control who's first and who's last. And that worries me a little bit that we have a very unbalanced market.
16:01
And this is just to explain it in a more detail. We have at the end, we have three big players in that market. Azure was doing way better in the last years from the growth rate like AWS was, which I don't know the reasons. I assume that Microsoft still has the way more better connections into enterprises
16:22
from the sales perspectives, but they were able to grow massively. Then we have Google. Then of course we have some others, IBM and we have Alibaba, which is perhaps regional limited of course. But we see that pretty much we end up like 70, 80% of the public cloud
16:41
will be in the hand of three companies, which is in some way honestly worries me because yeah, again, we have no standards. They have it, they have the API, they own it, they control it. And I think that's not a good thing. One thing is, and this is something we can see
17:01
because there is no real open standard tools like Terraform for instance, that's one of the reasons why they are so successful. Because Terraform, if you don't know, a Terraform is kind of a declarative way to describe your IT infrastructure and then you can run it against multiple cloud providers
17:22
or even your local OpenStack or VMware. So Terraform is the most popular config management tool out there because it gives people a way to abstract their non-standardized cloud API in a way that they can describe it with the Ashikov language.
17:41
And I think that's one of the reasons because we don't have real open standard and we have to decide on one thing and there's a lot of articles and blog posts and books about it. What is the right decision? If you go all in and say we go with AWS or we go with Google Cloud and we try to keep out that multi-cloud layer, which is of course,
18:02
it's perhaps more work because also like a Terraform environment cannot go from AWS to Google Cloud without modifying it. So we have to take care of multiple vendors. So it's a tough question if your strategy is cloud in. If you do it, if you really have a multi-vendor strategy
18:21
and a lot of people just skip it and say we go all in with one vendor because it's simpler, we don't have to take care about the other specific requirements and limitations and it's just easier for us. But the impact on open source, in my opinion, is really big.
18:41
Because one thing is that what happens is that the cloud providers we see take more and more open source out of that ecosystem. Some of them give money back like Google does. For instance, if you buy an Elastic Stack at Google,
19:01
they have like a back contract with Elastic to give them money and you can open up a ticket in a single interface, whatever it is, but they try to work together with the producing industry and there are other companies who don't. So they take out the open source software, sometimes modify it, sometimes not, make their business on top of it,
19:20
but not contributing back, which is a big, big problem because it ends up in fear and protection because what other open source companies do, and this is something we see right now, like Redis did it, they try to protect their property because for a lot of open source companies, I would say the 90% plus,
19:42
this typical business model is usually having an open source like OpenCorp, whatever you name it, but it's pretty much the same that it's some things you can buy when you give them money. We have some enterprise feature or some clustering and you give them extra money, then you get enterprise, you get support. And this is how an open source company can make money
20:02
because they pay their people. But for an open source company, that changes dramatically. If like an Amazon take, and we keep that example with the Redis, for instance, take that open source software and sell it, there's no revenue stream back
20:20
to the producers of that software, and that's a problem. So Redis, for example, came up with new licenses and said, okay, we have a license that you use in a SaaS environment, you cannot use specific modules. So they did it modifying the licenses which I also think is not a good idea that we come up with more and more licenses because if you ever have to deal
20:40
with open source licensing, it's complicated enough so we don't need more going to the business. So I think that's a bad situation and that's some market situation also we see currently that more and more companies don't know how to move forward with this. The new licenses, yeah, I said it already.
21:01
So in some way, one concern I have is that open source and open standards in some way need support. If it's like a one person project, if it's like a side project which a company has, in some way everybody needs a funding for his resources, it's some way to pay his own bills.
21:23
And the change of the market that pretty much most of the people who produce software, especially in open source, they don't have access to their customer anymore. And since a lot of the big players in the game, like Amazon are so big, they have no requirement that they need
21:42
the engineering capability of the original creators of the software. So they really cut the whole revenue stream down to that people, which is something I think it bites us because we will end up in having
22:00
a more unflexible industry in this way. And for the other ones, it's really hard to also survive then with an open source business model in some way, because in every stream is caught up when in let's say three to four years with the 80% of the people are running on cloud services. So where's the future in that or where's the leading us?
22:23
Honestly, I don't know, because I don't have this crystal ball, but I think we are in some way still just as the beginning. So this all came into place with infrastructure, the service and platform and services. And now we see function as a service more and more. So the big providers have their function infrastructure
22:43
where you don't have to deal with your virtual machines or your storage, you just call functions to, I don't know, store or analyze pictures or just write something into the queue, which are very, very hard to be abstract. So if you build your application on a function as a service platform of Amazon,
23:02
it's really, really hard to transform it and put it somewhere else. So you go into a full vendor lock-in, which is a risk you can take, but it's anyway, it's a risk. There are some open source alternatives away, like open fast. So there are some open source function as a service,
23:21
but what I see a lot of big industries and enterprises go more and more in a way that they use public available function as a service at one of these providers and also committing to a full log into their platform, which is I'm always wondering because I was one of the big things when it was about
23:43
like procurement, what platform are we choosing? How we move forward with IT architecture? One of the big things was like, how can we prevent vendor log-in? That was always a big thing. So how can we stay flexible as a company? How if we don't like to work with them anymore, or they don't like to work with us anymore.
24:00
So what are our changes to do something else? That was always, we were always worried about the thing and I'm still am. So every day, if we are in a position where we have to decide for something and how we move forward, I always try to think about if that doesn't work, what can I do?
24:21
Where are my alternatives? And I'm honestly really wondering why this, in some way, in my opinion, priority is not a big thing anymore. Or for a lot of people, it's not a thing anymore. So the question is, is it all worse? Perhaps is it because that's an argument if you talk to people in real life
24:43
or you have a conversation about it, that people say, perhaps this is how it is. This is how the industry moves forward and we have to live with it. And perhaps that's also true. So I think that's not all bad. In some way, of course it makes no sense for everyone to build a data center.
25:02
So I think in that world we currently having that we have in some way access to a lot of resources also in a flexible way is of course an achievement. So it's not all bad. But it's a risk. It's a risk for the price of convenience
25:22
to go into a full vendor login as a company. Even usually it's super hard to identify what you have to pay for it because that usually can nobody tell you except you get the first bill to pay. It's a risk. Login is a risk. So even if you see it from the IT perspective or from renting out your office
25:41
and you cannot go out if you have to, if you log in for a decision-making process and especially if you're an IT-driven company, of course it's a risk if you have to stay with someone and build your platform based on function as a service which are provided by the vendor and it's super hard for you to go somewhere else. But of course, loneliness is a risk too.
26:01
I get that argument totally when people say, I have a great startup idea. I would like to be on the market very quick and the need for speed in our industry changed dramatically in the last years. I totally get it that people cannot wait like for a year and build their engineering team and buy their own, rent their own Colu space
26:21
and build the IT stuff. They will use like whatever favorite open source, not open source, public cloud system they can use and get and start with the application. So of course it's like a thing you have to think about what is important for your business. What we see also is that the cost thing is really
26:44
something where people take care of most. So that's the number one priority and this is crazy how many companies built their whole business model around how to save cloud costs. That's really crazy because people figure out that of course resources you use on an everyday basis,
27:03
especially if you're not able to scale them up or down, gonna be crazy expensive no wonder because these other enterprises need to run machines as well. And another thing which comes up and this is something we see in some areas but not so much still that I'm a little bit wondered
27:20
that the whole market is controlled by three to four big vendors. We have an oligopoly which happens there. So take another example when we talk about an insulin which is perhaps it's a little bit far away as an example to IT resources. But the thing is in 2006, the dose of insulin cost about $67 roundabout.
27:44
There are different vendors and different companies. In 2019 it was about 285. Even the production of insulin itself gotta be cheaper and cheaper over the years but the problem is that there are only three to four vendors who are selling insulin and they totally control the market.
28:02
And even perhaps you cannot compare the dose of insulin price to cloud resource offering right now. But imagine the world will end up in having three to four windows where you are able to buy IT resources
28:22
and also get in some way and I think this is something we will see in the next years they will control the access to that highway. The access to the highway, to the internet which was still right now controlled decentralized will be more and more limited with their own protocols and own stuff.
28:42
And I guess we will see that in pricing as well because if just a couple of vendors control it it's not a law of IT, it's a law of economy that pricing with a little bit windows they have no benefit in like gonna be cheaper. I think this is interesting. The future will show if I'm right, I'm wrong
29:02
with that example, but anyway, I can do it. And there's another example like Nestle which is very popular. I don't know if popular is the right word for it. Like buying water resources all over the world and pumping out waters like crazy and then selling it to a crazy amount of price
29:21
selling people back. So because they control pretty much of the market it's totally up to them to create pricing structure. And that's just like two examples. Water is a good example. And if you see now, water is a fundamentally problem we see now around the world that people don't have access to water.
29:40
And there are companies like Nestle sitting on water in South Africa and pumping whole villages empty and then selling the bottle for 50 cent which is a lot of money for them. So this is the same in some way, the same way market movement I think we see here perhaps it's not comparable 100%
30:00
but I think the economic story is the same. So the abuse of power comes as no surprise. So if just the limited pool of companies controls that IT market, IT resources they in some way will lead to abuse
30:21
which sometimes it already does. But that's something which is no surprise to me. Government impact of course is something means that's a picture of the German Bundestag but means right now the situation what we have is that 70, 80% of that public cloud resources
30:43
is under control by an American government which of course you can doubt. You can have different perspectives on politics which is not the topic of my talk but in some way it's crazy because we as Europe missed it to be part of that game.
31:05
But the whole industry pretty much ends up to be controlled by a single government. And I think that could not be in any other's interest. And I don't get it why German companies or like state companies go into public cloud
31:24
which is totally under control by other system and they use the control. So this is an example. You see that the Chivo which is in Venezuela and when it's just an example last year there was an act against Venezuela
31:42
that from one day to another companies and agencies in Venezuela were not able to use any public cloud offering which the United States companies offered. They were not able to use it from one day to another. So there was an act came in place. And for example, Adobe, they have their accounts in there
32:00
and there's an article it's also linked in the presentation. From one day to another they were not able to access the documents because by that act they were not allowed to grant any company in Venezuela access to the data. And in my opinion, that's totally nuts
32:20
but this is a situation where we are right now. So what can you do or what is perhaps my point in this? I think perhaps it makes sense to not follow every hype. And I'm not saying that public cloud is a hype because it's definitely more than this because it's a change of our industry. But it perhaps always makes sense to think about
32:42
is it really something what we need? Is it really something we should go all in with? There was a tweet by Kelsey Hightower a couple of months ago already when he said monoliths are the future. Means Kelsey, for example, is one of them who promoted a lot of that microservice architecture.
33:02
And there are tons of articles out there like Uber or big enterprises. They pretty much cut their whole applications into pieces and have like hundreds or thousands of microservices that totally went pretty much out of control. And now they now try to gain control back
33:20
in building more monolithic systems again. So, and that's also, that's something of the IT industry we have that every now and then that we go everything to big machines then smaller machines and the landscape will always change. And sometimes if you have the feeling that you miss a hype train, you'll figure out that the next train is still coming and is on the way.
33:42
So cost control is the thing, especially because in the cloud it's so important that people worry about their costs. I don't get it why so many people don't really have a good control about their costs when they're on on-premise, which is interesting because I know a couple of enterprises who are doing public cloud migration
34:03
and for some wonder they figure out that it's crazy expensive and then they put other consultancies in place or any cloud cost control company to help them to reuse cloud costs. And I'm wondering why they weren't able to do that before and perhaps they have an idea about their own costs.
34:23
Why these interest and sense for cost of resources is something which is not coming into place if you run on-premise or your own infrastructure. So be reasonable. I think absolutely in a lot of perspective,
34:42
the solutions available on the market make sense for people, especially when I'm saying you're a startup company, you have an idea would like to start, of course you go with a software as a service offering, you go with any public cloud and try to be on the market as soon as possible because the market will not wait for you.
35:00
That makes total sense. But does it really make sense to put your whole IT you have, which is wherever if it's like a provider or somewhere or your own data center, especially if you have a base performance you have to deliver. So you don't have like a deployment system
35:21
where you have to spin up 300 machines and shut them down five minutes later. So just the basic thing you have to do, which is usually surprisingly cheaper. So what makes sense? So why don't people look for a mix and say, where does it make sense to be flexible? For example, in a build environment,
35:41
because also then a vendor lock-in is not so problematically, so you can go somewhere else. I think another thing we should take care of that interoperability is really an important thing. That we demand from the vendors, we have the companies, your enterprises, buy services from that we demand,
36:03
a good way to interact and combine multiple services and solutions together. Because then you as a customer have a chance to go from one provider to another, from one solution to another, from one database to another, if you have the chance to work with open data.
36:20
And surprisingly, most of them, it's super easy to bring data in. At some ways, super hard to bring data out. And that's something I think you should demand, especially if you're a big customer of them, we should demand more of that. Prevent vendor lock-in. Again, like I said before, I know a lot of people say this is not an argument anymore,
36:41
because preventing this costs you speed on the market and you cannot afford it. I think it's still an argument, perhaps not for everyone, but definitely I would like to bring it up because for me it is. The variety is the thing, which I'm really worried about a little bit,
37:01
which makes the internet and as you see the IT landscape when we work together with data, which is like transferred over the world, the variety that big vendors, smaller shops, everybody in some way has access. And the same thing is true for software developers
37:21
is that variety is really one of the superpowers. We won with open standards, we won with a lot of open source and with everyone having a fair access to the data highway and having a fair access to the resources. And I think this is something we have to take care of,
37:43
that these parts of the industry have a chance to survive. So demand contribution. So if you buy on your cloud provider, so try to figure out if they give back, when they use open source tools, which you previously would be perhaps forced
38:00
to buy an enterprise subscription. So perhaps you can convince them and that they do that as well, that they buy back support. And I know if you have three instances in AWS, they don't give whatever. But if you have a big company, you can of course try to bring this up and make this as a topic. And perhaps it will change in some way.
38:23
So doubt always, if it sounds too good, it's always a reason to doubt, especially if somebody is telling you that this is the only solution if you don't do it, there's no chance for you in the future. Sometimes it's definitely a reason to doubt
38:40
that's really true. So the diversity is up to you. And I think the diversity is up to all of us because we talk about the diversity in our industry a lot, which I totally know that there's other importance for other diversity, but also if we don't wanna end up
39:02
with three companies controlling IT resources and IT traffic and the internet pretty much in some way, I think we have to change our behaviors in some way, whoever we is, but otherwise we don't have that diversity and will be super hard for a small shop
39:21
or for anyone else in that industry to enter it. So I think if we move down the path, it will be super hard for a company to go into that internet business in three to four years from now, because everything is sold and all the fields are given away and it pretty much will be impossible
39:41
to have a fair chance to participate in that industry. And I think that I'm worried about that situation and I'm honestly also worried that in my opinion, so many people don't care about it or don't think about it, but this is what I'm seeing coming next to us.
40:00
Thank you. I hope it had some value for you and perhaps we have a chance to ask questions and I'm happy to answer it. Again, thanks for listening.
40:21
All right, thank you very much. Are there any questions? There's one question in the chat. Which percentage of the costs of a typical company
40:42
you know is server costs? In other words, to what extent is it costs or getting products to market quickly that drives the move to third-party clouds? Oh, that's a good question. So one thing is, in my opinion, one big differentiation is of course, if you make your money with IT only
41:04
or if IT is your business, let's say if you are a bank or let's say if you're building cars that are companies I know or you're building parts for cars, then IT is required in your business because without IT you cannot do the whole process stuff,
41:24
but you're not selling IT. Another thing is if you're an e-commerce shop and you pretty much don't have your own warehouse, you pretty much just build the platform, let's say Shopify, that take that as example, for them, the percentage of the server or infrastructure IT costs,
41:41
let's take like development out of it, has a way other impact. So I think it's a big difference between being a producing company of whatever technology or goods or pretty much selling a digital product
42:00
like a streaming provider. I don't know what percentage the IT costs in a traditional producing company to say, but I think that's the big difference. What I know is I know three to five companies where I really know that they, for various management reason,
42:22
for let's say forever reason, they are forced to go into an all cloud strategy and even if it in some way, and there were a couple of examples where it works technically, but cost wise, none of them hit the expectation they had before.
42:41
Perhaps they were not skilled enough, perhaps they didn't read enough and they didn't doubt enough, I don't know, but any of them were totally surprised how crazy expensive that is. And they learned in a hard way that perhaps it would be way smarter to keep some infrastructure on their own
43:00
and go with a high volatile infrastructure to a cloud provider. This is something I've seen a lot. I don't know if that's true for anyone. So of course you ask others, they have a different story. I know that's not a real 100% answer, but I think that's a big difference between how the market works, I guess.
43:24
All right, thank you very much. Any more questions? I'll give people a few seconds to type. No problem.
43:44
But it doesn't look like there are any more questions left. No problem. Thanks for the time. Have a great evening. Thank you for your talk.