Welcome to is my pleasure and I agree mean it to introduce OPA many of you may know so I'll tell you story when I was at IBM some of you know was the architect of my series which is Message Queuing platform used by most companies to into series and I was most of you that attended election know right about replication integration data and I saw this book and thought this man must be a genius and he must be 120 years old Because this contains all the knowledge I've always wanted to have about message curing and
messaging and well that everybody in the lecture does still this is the best textbook in computer science I'd ever read it's a pleasure to read it it contains absolutely true knowledge absolutely not acknowledged and then by chance I bumped into him absolutely shocked so this was a very very consultancy country and he was doing consultant in this message King arena And the head time as it comes out new travel around and he spent all night in the hotel right into the textbook this is for people and then he changed the sudden and went to Google
now but then that the sudden each hd's but recently I met his former boss the CEO of aliens fantastic and then he went back to Google the game and now he is a freelancer karf it's a couple of days free in the freelancer and knows a lot of stuff of consulting Hi CX type of people right is famous about architectural skills with Google to give a talk on architects elevator most in my advisor oh it's my pleasure thank you very much yesterday three innocent freelancer free to decide so this is finally forecaster the body me
to summer so many many many times I never get any employer to fund any travel to treat is it down the line who wanna take ourselves Problem than happy I probably got me and they're satisfied to be there so couple of things some of the day I'm going to talk mostly love architect so having a very different view of architecture and architects really mean have a tiny little preamble though little history lesson my love of service object important to computing but I don't see at this workshop Asia and here pronto and IOT all this one
kind of stuff right under The essence of it it's still service order computed as Frank so alluded to my history with service from the computing goes back to the book enterprise integration patterns and here some this great website we didn't back then this is I have a 2003 when Frank and I were booted younger and it talks about messaging patters and because he was so great the website and the 2019 looks like this it is basically it's mine capsule we've never changed and it's Basically shows that some things just do master or fee have so
much stuff that's changing a few things on the interstate some fundamental architecture patterns and architectural styles I think a part of that so I have another history in northfield of the folks here I mean the appearance of publishing anything just just hanging out with some pens and he thought that was 2012 years ago long time and I contributed a chapter I think this was the scene here and actually Conversations between loosely coupled systems and the conversation talking is actually from the disturbance something that interested me quite a bit and in 2008 I really started writing
about conversation patterns and the way that is looked at this it's a sort of thermal view to the messaging patterns rather than messaging patters it looks in the same message how to message flows through a system right this is very elegant pipes and filters that are Classic you know architectural style and it's stateless because all the status the messages it scales it is decoupled is impossible in a pond has these really nice properties but when people read our books they realize there's a few more things missing so for example air handler was partly partly live
or so the complex interactions you know back and forth transactional kind of things reservations are sneaking out of style for compensating actions all these kind Of things were missing because basic messaging doesn't have the vocabulary and these messaging is the message was right and that's all it does so I started thinking more about the other half of it rather than following the message to through a set of participants can we look at it extend potential participants and so see how the conversation between those two occurs over time so rather than that looking was so how
things flow it's about how Things go over time between the fixed set of persisting this this whole thing on a different viewpoint and my favorite example is actually different is we talk about publish/subscribe so much right a very basic pepper pops up you saw send one message and every subscriber gets a message while looking at these two different patterns it's typically have a very different view of what is interesting about pops up first Obviously pops up is a misnomer because they usually some pop his first to subscribe you know to get messages and how we
say pops up also the name doesn't match semantically right because subscribing is sort of a meta operation it is subscribe once and then publishing into all the time so pops up man interesting things when we think about it is wrong in at least two ways it's probably ordered and the two words cover two different dimensions so so let's First two puffs up but mostly look at who's subscribing with us the message come I already have a publisher you have a subscriber but when you look at the other side where the conversational camera is there's a
whole different question that you have this life ahaha do these subscriptions actually last rather they get a whole different conversation he also a pun intended about Lisa's and timeline resource allocation right subscription see of Those kind of things so he even actually think about how does a subscriber happen and you don't care so much how many subscribers they are because you don't follow the message as much but you really look at the life cycle of a subscription like do I actively need to renew to a police get reminded to the messages just a whole different
set of questions so having a lot of fun with this unfortunately this was in 2008 and then somehow I spent the next ten years Doing random stuff so the talk today is not a message detective the talk today is about random stuff and my first service is actually my official status my ego status is 900 but because I'm a freelance I'm a bit Singapore where I live my piece I had find a visa so if you wanted to independent things currently I cannot do that in my home country that's quite bouncing around the planet so
technically I'm unemployed As far as legal status in Singapore all right that has some really nice advantages I do what I like right free as a freelance and say freedom is the most important thing I have almost like no but I had all this corporate machine there I need to deal with and finally I have some time today and then front work and I were talking about something else and probably one of the very very few people here this event I never quite made it there I went dissertation right But if you think about it
it also has enormous advantages I can say whatever I choose I don't need to specialize in any feeling all over the map and most importantly I never need any evidence I just make it up as I oh and that's another special form of freedom and I enjoy that freedom very much with Will Turner architects so that's that's what you guys to get to know me a little bit now we're talking about architects not architecture and architects are like Special animals right yeah scream from people but architecture is you probably get four different answers and many
people say the same things about the ID tenants that people that they're all kind of special animals so I talked to the typical celebrity crowd about what it is like to be an architect or chief architect right the first image that comes to their mind is you hit this time right in the Matrix movie and that's the only young Architect right the person who makes all the decisions will knows everything right protects the balance and the university cetera it sounds really good until you realize the movie a little bit that this section of the unity
and these are the machines right Matrix movie story back so these are the machines so this is not a human being so trying to be the although an architect is probably a difficult one for France whereas for us too then Another very strange will be more recent movie yes Anthony Royale in highlights and this is the architect to asses grandiose vision of this perfect building the militant absolutely ugly totally completely goes long right I'm sure we seen these kind of architects as well right there this fantastic vision for realities far far away and the two
towns to not meet when people look at me as an architect to get more fear which of this guy Ryan is like In stop ground from Back to the Future I think that he has salute has a little bit to do with it and any son of my way my way of doing sense so it's a different way of looking you know what is an architect in to understand what the role of these people is and why there's so many stories of something personas of different architects the first thing is in order to understand why
we talk so much about the role of an architect Especially about organizations we first need to understand that the role that IT plays has fundamentally changed that's why it's interesting to talk about architects we go from here - couple things you tell my outfits architect elevator since each advertisers at least my domain name by its architect I'll come if you don't remember anything else that one you should remember changing over 30 these things we talk all the time about digital stuff by addition Transformation to generation digital challenges digital giants right like we go to any
large European almost worldwide organization they don't have some digital thingy program of one way or another right can almost be guaranteed they have something for my hands to do that as well making things digital is actually stuff we did my Frank and I were still down a long time ago right which will replace letters with email will be less precise chemical Calculations with spreadsheets right sometimes we do funny things like making matches where you need to push a button just to know the target is right but basically the digital thing we did it's the decades
ago before some of you were even born so this whole digital thing is a little bit of a misnomer right it's not about making things digital it's about using the technologies to fundamentally change the model and courts let me I think Generally some very very clever people have started to sort of draw a line between digitization digitalization my promise I can never remember which one is which socket on use these kind of terms basically say when we talk about digital remember that is the technology which that's cute right but we are interested in changing the
models and a great example about what it means to change the model is looking at an encyclopedia yeah who remembers in Carter I was like ok ok ok an excessive I'm a generous man in capital was a great digital encyclopedia write as much as you can make fun of my encounter right now it's important to remember that in all the mentions it was superior to the printed epithelia right have more volume better not to take multimedia content easy searchability right easy portability like on any dimension you draw it was severely it was a super great
product with printed stuff however Fans who still using it Carter no hands because that's Horrible's but it was a digital replica of the thing that we it was superior in all categories but only in those categories that we have defined it did not question or change the model and of course you know the reason you guys don't use account anymore it's all well known because not use Wikipedia and Wikipedians made example of something that changes it questions some will not Very safe assumption the assumption then knowledge needs to come from experts I guess a small
group of people who just kind of better than anybody else they write it down and share it with everybody and nobody ever thought about this strange idea right people achieve as we were saying I because you don't want some encyclopedia never random stuff right so nobody have a question and Morgan because it served us well until somebody came along and replace That volume is something completely different and then we start to realize how many of these assumptions we actually carry well we don't know this these I can't the believes we have the assumptions we make
we don't really know those until they get challenged the way in this translates back into big business Mike mentioned I worked at analysis a slogan I often told The Onion's leadership I said hey listen people you know the company is a hundred Twenty-five hundred twenty-eight years old pals right so I was sent any business model that doesn't depend on IT you can be pretty sure we found it pretty clever people and we have a hundred and twenty eight years delusional right now if use the same knowledge it is like that's gonna come out it's only
technology that is about changing the model it's all about selling insurance online store that because that's making Things digital right it's rather doing completely different things like preventing things all together through prediction etc right those are things we define what it surance means right those are the things that actually change the model now that was interesting as in how many ways these assumptions are baked into the way organizations and Aunty organizations function that's this interesting thing about these assumptions we don't know we Have them really until they get challenged and B they permeate our thinking
and our doing organizational structures now incentive systems they currently in all these things so if you look at anything in order to get anything done in the traditional IT environment you need one site that's not a project if you don't have a project right nothing will happen you need a project in funding right and that's what this looks like We have analysis and the true P and you make the business case right out of this business case you make a project and you know in time the project is our dear mrs. when the dattari usually
happens the launch party Christine you are happy when the project is finished and we can come back to business as usual and business as usual needs no change peace of mind drive operations run now the change stuff so this a fundamental belief here that change is a temporary Effect that can be bundled into these so-called projects and the result of it can be predicted and ironically the big part you have basically we're the first user ever sees your brother that's when you celebrate before anybody actually has ever really seen it kind of questionable kind of
thing right now how do we break these assumptions or what does the digital the so condition will bring well the digital world as we call it is one of constant Change right it's not new territory but trying things out the experimenting right there easily by our target state so it is fundamentally different right and so projects don't make any sense in this part or country is acceptable you can approximate this but I haven't IT projects right I think it's all about what am i doing today will Springs intubations that's what the ads are people cannot
do right but in the end the picture is you know basically it's Much more moving than resting right and this for these celebrations when wrestling starts I was in the digital world in project ends when it's dead in the market well this a gmail has ended Google Wave has ended as a project right so will these people where the project is they have a happy celebration launch party basically when a project ends here they have the funeral right the project is dead and that shows you that this is not A plain words right but it's
exactly how many degrees the opposite of each other when people have had time they are people know the product is no longer successful right and with this kind of course the skill that you need to have this also fundamentally different right this is a scaler there you go good cancer right like in terms of to restate of how they were taken how much money it costs to because to be Honest pay it all right so you'll be a good guesser over my dead body being able to guess it doesn't hurt but mostly you need to
be a fast learner and I call this contrast the economies of scale because he'll building something and you basally can either return on investment but he just being big or heavy slightly bigger you are on the turnover of this project you get so the currency is size of the right-hand side the currency is speed now there's an interesting Adam to This that you can put in this if you only is moving at least from the zoom of leather and if you're always moving absolutes become somewhat meaningless and you will find that these people use a
different language that if vocabulary or red wire green line people here the regular people they talk about budgets they talk about health they talk about how much the server costs 50,000 euros for server they talking absolutes on the right hand side because you're always Moving absolutes become a lot less meaningless minimum less meaningful sorry I was saying one time the German Autobahn the three hundred kilometers an hour trying to look at the map it's exactly where you are isn't that interesting half a second later already somewhere else right so it becomes a matter of you
know predicting how this is going to take how fast are you moving etc so the force on the right hand side talked About bird wings right they don't talk about budgets they take about bird rates also they don't talk about how much the server cost in absolute they talk about how much a server cost per hour because the ants again the rate so then that's one of the reasons to the force on the right hand side cloud computing sees totally obvious because cloud computing speaks their language it fits their mental model I sometimes say these
people on the right Hand side they live in the first derivative because the only constant they have there is the first derivative so they essentially live on the first arisen very different think about it and if you look at computing all the data centers and Amazon and Google and all those folks have Delta all very nice but it'll be the data centers for pretty long time right so the data center cloud is about a different provisioning model and a Different building model and the hastily difference of building model is incremental its elite it's the first
derivative if you wish by D of dollars over time right and that's why it fits so well in the right head boss so I'm sorry money but they're important to understand that this thinking Cavalieri people use is fundamentally different it defines how these organization functions and when you make cloud computing to the lead Line people of what they do is they make a cloud migration project managers that's what they know and they move everything from one side to the other and then they're sharply surprised that the transformation hasn't happened because they sort of just plucked
one infrastructure tracked another one in nothing else has changed it hasn't gotten them into any economies of speed so changing the model is much more important but also much higher this is Another like how different the two things are you can see how he would operate in these respective economies the economy of scale is largely you're automating things that already existed in the insurance business what did you automate risk calculations claims process and new accounts payments right these are all things that insurance companies have done for 300 years so you know what you after and
that's why all the classics are waterfall kind of kind Of things that we have as a business case it's often triggered by cost reduction and to be honest most corporate IT has done fairly well for itself junus mechanics has a tea punch of like 3 billion euro over here that's out month well sorry it wasn't like the market was wrong tested pretty well but in the end is a model that was in a world where you can't see this place the existing processes existing Things with ITV and that includes even rather than going to somebody
in selling insurance you know be an agent if you just do the same thing about with a mobile app and I do change the technology but not the body so it's basically exactly the opposite right you discovering new things different ways you're testing hypotheses right your money experiments by the very companies these experiments haven't really told in terms of how difficult it is to Understand the other world if you live in the tech world and there was a classic that it lobstermen financial services organization let's say somebody came up and said we need to fake
it a touch I because they hear about these digital kind of companies the experiment that they fail and said well we already atmosphere the culture look at all these projects you know that tank like 10 billion euros right we don't need a failure culture we come because of the Right hand side if an experiment fails base if it's a hypothesis is falsified and a custom relatively little money later so it's not about failing right this shows you how easy it is to misunderstand how these different things work it is a memory and yes most organizations
have a birthday the culture is the bernam of money projects that are hard to believe too big to fail have you seen all that kind of stuff right so that's the IT now in that world Is a very nice model about what more fundamentally articulate and it's interesting because most people would say 90 is important but as so often we see important isn't such a simple word as it sounds because important can mean two different things it won't tell me in Stata that's certainly true for anybody fatigue disappeared instantly an average insurance company or bank
would not have a lot of lifespan right couple Days maybe right it's not going to happen but there's another axis of important that is is it a differentiator it's also important the power is on if the power goes out is attentative convenient but it's not a differentiator is just like a commodity and this is what we call in this model for the shift from economies of scale economies of speed the activity moves out of the parity everybody has some IT but it's not particularly Interesting to it is actually the most interesting thing because all future
business partners come from Deonte all right again it's a very very different world Brodie and I can choose of why being an architect in that environment is actually very interesting so it's about not about making things digital rights about changing the model changing the model means also we go from these economy scales and economies of speed Rather basic lid for the first derivative and we have a very different sort of mode in which we operate and I think that also very much influences the robe of Architects because I checked off some of the hooking point
the pivot point of the RT and a lot of decisions right they talk to a lot of people they make a lot of plans so I believe the architects are smack dab in the middle of this kind of thing and what I believe them the most Important skill and architect needs to have these days is why I call riding the elevator and it relates exactly double with what we've just seen most organizations have a lot of leaders right and that has served when to the dislike there's a penthouse somewhere here right but the decisions get
made right and then you some of the other additional capital punishment the middle middle mention and the lower management of the alcohol industry trend the middle Mention of the local management program manager project manager Lee acquired Corner South engineer and you know here is the 40 student who does actually decoding casino a thousand or twenty polyols right so this is what your classic IT organization looks like or maybe the external consultant and it's only the only one who can do is qualify because the other side for some dumb models again not so fast and we
come to this in more detail the Model has a lot of advantages right separation of concerns for example right the people in the penthouse don't need to know how many talker containers exactly running in what version of kubernetes you have deployed right so it has some convenient properties but of course as inconvenient properties and that is a communication takes time ahead because you go through all these levels and like I said the people in between above Qualified so you have the telephone game the still cost kind of problem right what comes down in the engine
room is generally much different than was started in the tenth house and vice versa and as you listen these economies so speed as you mentioned things get faster you start to realize the disadvantages of this layer architects regulation we start out way key advantages now what do you do the one thing you can do is try To collapse this thing right because the digital companies the so-called digital right the Google's Facebook's Amazon's they don't have that many players deliver a bundle of I always say highly technical management right short decision paths they don't have this
now the problem is if you blow up the skyscraper you don't get a bungalow you can imagine trouble right if you say I pull out of the middle of a divorce right that is not a new building that's A mess so you cannot do that well it's not not easily right I always say if you're trying to reform middle management you know there's a whole village of vacation homes and at least a few hundred children's education at play right so that is not kind of eliminating that you not get eliminated very quickly so perhaps it
doesn't work so what is the next destination to tax the rich [Music] connection between the levels that's Exactly what I think the architects are best equipped to do because the architects are the ones who understand the technology but I also understand the context and can communicate out and these things I just the value of an architect by how many levels they can spare I don't care about where they start probably always have this the people in the 10th house always think they're more important rather than the others for me That doesn't matter where purely Magnus's
county force then if you go to the top out and you have a great buddy who goes to the second half perfect as long as you talk to each other that works now as always there are a few high standards and you've seen this those who people who might be haunted elevator but only wants to the top and they stay there and they easily identified by their favorite slogan which is or used to be technically is always like the Technicals great but even you would really like what's wrong and I often say if so many
companies it is an answer to say part of the technique of it have you ever seen a chief financial officer who said what a not so good at math you know like what is numbers of you know division stuff maybe Exim is Excel but if they said hey you know those math stuff you know and I mean I used to be good at math and now the chief financial officer a lot of people But somehow imitated seems to be liable to like a fee but the technical right so it was the books are only by
the elevator what's up a few conditions as I said looked in order to make this work you've got to look at both the organization as well as the technology all right I said the time technology gives of the new tools but the organization holds these models these believes this project thinking all this stuff is baked into The organization which is whose people and processes at the contrary to it so you've gotta do both and the other thing is riding the elevator and telling each floor a different story doesn't count either right you're supposed to connect
it so again you've gone and have the same story but told in a different way rather that's what makes that interesting and challenging now once a architect needs to do is they need to be able to reason about complex systems Every new process aligns right we deal with systems and we reason about what these systems do the same is the case here and you will find some very interesting very interesting certain events and what you may find is that even though the overall state is inefficient right all these levels of this publication of this middle
management you will find that most of these parties are actually happy with the current state so the system has Settled itself into a local optimal that you know not to be a global of as it generally is the sum of mobile-optimized our global optimum but each party in its own way is happy and it looks like this management lives a happy illusion that there are tears actually are fairly capable of weather you know the using computer bananas so they're like cloud enabled family Google watch out here we come right on it marriage ventures believes in
everything Is going family well likewise these people each other different because management absolutely no clue what they really do right you just tell normally building a cloud infrastructure big data of machine learning analytics kind of thing that sounds really cool event to whatever you want and the base of that Boris is because they clearly separated that 24 that was electorates we call that decoupling we call it loosening up with systems right it's one of one of Those architecture principles realize now respectively that's all happy with the current situation now the problem is when you come
in as an architect right you burst all their bubbles because you didn't say you know what yeah once you stop running you replace the illusionist transparency but it's like yo says I turning the lights on in the basement you see all the dirt and a first reaction has started and it's full of dirt so that's what you're in there you Bring some governance to the edge you look like a few checks and balances tonight gonna happen and unfortunately what are you bringing to the middle level we are the most somebody sneaks out yet so this
is one of the things that need change so interesting and what makes it so important so I have an architectural mindset to see how these things function all right so we have this local Optima you go in any completely bust the model With a time okay so I ordered the smart guys to deal with it those are the things that you know you need to be able to do with the different levels all right you got your business IT strategy is correct that's the fundamental premise of the elevator if those two things don't match it
doesn't matter how much automation how much efficiency you put in right it's like digging the two tunnels faster but they don't need and if they're not connected nothing else Can help you right and then the order to make this work right generally you need to drive change because organizations aren't built this way they're built with hybrid towers of all these kinda levels and one of the biggest inhibitors is the complexity we have our stuff just is difficult and our software can do a lot of things that it wasn't able to do a while ago and
we think these days is infinitely scalable certainly resilient we can do updates without downtime I can Raise I mean you name what we can all do but the one thing we did do is make it simple and even admit that in the topic uber name is the other fancy stuff simple it is love so you need to combat this complexity is not going to disappear and one of the best ways I know to view the complexity is explain things better that are all those demystify and I think through all the little bit guilty of this
because you know life in jammu can be tough as Our other stuff we need to deal with and we always love for management to understand all of the complexities we have to deal with right but what this of course does is carrying all this complexity to the levels of people just say I understand these artists are Sunni to have strike them all away and that's the best currency you have to some I get things going Ichigo miss a little bit easier right we know how Star Wars get fetched out of the system you know make
Systems kimball etc but let's actually sort of have the elevator we looks like and what we do is take a quick thank you but but see what an architect things like and the different leaders of this penthouse part of this disaster for every story in the penthouse in the house you TV with organizational structure it's important for you to understand how reality functions and what is expected from unity and generally corporate Pieces are really cio and achievement formation officer sometimes people also joke it's called career is over because it's a very difficult job and we
go see how you're supposed to do everything right this is supposed to save money transform the company be innovative right now the use of horses know the latest technology right see how there was a tough now a does this or into if you are in the power how this organization is wired right we talked About some of our modern what believe these organizations have up and the org chart gives you an enormous here and obviously this is architecture and is finalists ready looking at the system except this time it's not servers you know and that
where it distances organizations you looking at the into tendencies and what do you find is that we face cio reports to a chief financial officer which is the classics of a model their main KPI is gonna be cars right Because the CFO was the chief financial officer is the person who writes the checks for Martha much right they trying to prevent the family from writing too many chunks so they look at Cost Reduction now whether you like it or not the logical right thing to do if that is the model is outsourced if I tears
if your cost if you put it somewhere where it's cheaper you performed in line with you Canadian you did exactly the right thing You reduce the cost now we know many reasons why in a model a verified but if an organization has that reporting line you run a suit against they they care about cost and let's say you come with all the fancy doctor cumin a misty or whatever a lot kind of stuff and you sell a jolly continues to the redeployment spinnaker they are all those kind of things you basically sell it speed to
somebody who only cares about cost and the effect Of that is architects and management is stupid and understand that I'm gonna listen they don't get it because you're selling the wrong thing to the wrong person so it's very important to understand what's in the heads and you can read this other they are China and that I saw always see all my ancestors for example the CIA all reports of the co-ceo is the chief operations officer and sometimes say is the chief optimization officer and then make sure Everything runs efficiently right and this takes a big
step it's still not where we want to be with our traditional style right but they found by specialized most important thing is they managed by return on investment the difference between managing my cause and managing my return on investment is a simple car analogy if you manage for a cost you don't get an oil change because it was money fifty pounds if you manage to return invested you very much again No change because of each other investments is extremely good so it's a big step ahead so this is what people talk about refactoring and clean
coal many of the right but you see it cost some money but it has a high return this organization would be much more likely to engage in that conversation than my first point however since we won a transform right wickets interesting for us is on the right hand side the CEO is a special Character Beauty the CIO would not report to the CEO CEOs chief digital officer with all the caveats we have a digital media where the punishment we saw so the chief digital officer is some of the other key of the CIO this is
often an attempt where the company says it's not reason for us to do all these things at once so I'm going to take the economies of scale and leaves and with the classic CIO probably recording it to the CFO I'm going to take the economies Of speed stars are moving under the seal is one of those classic things that sounds good on paper but isn't such a great idea unreality because now we have a new fancy fast idea old crappy Aranea come straight light and now the question is we're really good people want to work
and the opportunity is the one who keeps the lights on it's a bad recipe that sounds good so I think a lot of companies the CEO static a man with people experimenting with it of course The last bubble is IT is just like anything else reported by the a to the CEO why should the CIO be some of a secondary product second Debra TS else important is anything and what you would find in you know things like Google's a classic example in eight years at Google nobody ever talked about nobody ever said we need to
talk to the business or we need to talk to our team like that distinction just did not exist right and sometimes people say always Because it's a technology company [Applause] does the money come from they sell advertising they actually wish they saw what technology this Google Cloud right we're trying to catch up with Amazon right so Google is not selling another technology selling advertising it just realized their voice for it for it to be successful IT is not some of a commodity thing that does the bookkeeping in the payment I teach the Core differentiy have
less than change tomorrow fundamentally right and that makes them have such technology company as a beelzebub as much an advertiser insurance company yes they have different way of thinking about it so has a right understanding organizations you will see many see Eric was moving to the right amount of tons that we did that you can find many others are big companies making these Shifts and it's always a sign that somebody at work talk has understood that this movement and while it takes time to trickle down I will give this companies a lot of credit for
actually making big organizational changes based on that it's a very good sir and you know things are moving that suffix illegally the nice little messages I have 18 years is the census architect as a modern architect you need to deal with technology systems and organizational Systems I only mention them if you want to be successful you can't separate the two at the end of the day the technical ones are the easier ones but they won't really take effect if you don't fix the organizational systems right you're all the C ICD but you have a six
week approval process to approve a server process for new Socrates you've got nothing now most technical architects don't like to deal with organizations all of that much because you have to Deal with people and politics and all those other tedious stuff like technical systems are just lot easier to deal with so now we have the challenging to do with all this organizational stuff now the silver lining is that organizational systems are technical systems actually very much follow similar rules unless that's your saving grace as an architect you already have the right way of thinking to
influence organizations because the dynamics a little to us just Different now isn't it let's take this weather this one whatever planet what architectural panel has always helped us is lately right we know like Kabir science is the people in every problem can be solved by one more way of interaction and so lately always force and why is the later stuff well right this is probably much much longer and we have 20 very smart audience so I take that as a minus 100 examples right a separation of concerns clean interfaces Replace utility right clear dependencies overall
you have strike the details away with us but yeah and the texture would be so interesting if they were that size right everything has ups and downs so it also brings some overhead right so for example yeah I've seen lately diamonds so each of these lines has like so XML written on it from I think Frank is partly responsible for that in certain so do you carry a lot a translation Overhead right also this thing can get its own complexity and all of the changes propagate right this is classic thing all I want to do
is add a small field to my user interface but I've got to change the URL of the application via the API there the persistence of the air the techniques later the change end up changing five layers so most of the downsides now these two things that are very interesting here nothing of these things say whether this is a technical Or organizational system but they apply equally well Otacon responsibility outsourcing right an organization being able to lay around put something else in having clear and he can do that because he dependency that's the premise of outsourcing
and actually if you don't have a clean later if the last thing you probably want to do is outsource right because the dependencies will kill you okay so this applies and we've seen all this Time why for education local optimization right people who optimize locally and as we said you know some divine is not a global optimum right and that is true for technical systems as well as organizational systems why City office or something fill out this 24th what because they optimized their own internal processes once you have the form filled out they're very efficient
right and their balance optimization end right there and the cost if you have Filling on the starting out doesn't show up at the chemistry that's local optimization have we seen the same thing people need a fantastic API fantastically easily from them to implement but neon fantastically difficult for you to use local optimization right so that's message hello already this applies to organizations and technical systems equally much mobile message number two if we use how architecture grades and so Soon out a little bit let's imagine with one of those IQ tests like which Eldon doesn't fit
the series kind of thing right what kind of nature are these sets of properties we find that the left hand side is much more structural kind of things rise of dependencies details interfaces in mostly talks about structure the right hand side is different the right side is more about latency runtime propagation rate the right society blue fuzzy or fuzzy on the Right hand side is much more about dynamic properties of the system there are the different things and that gives us a hint as to why organizations became this way because there is slow moving we
wrote the structural properties are very handy and this Terminator and just don't hit you as much this your world is very slow-moving so you know propagation and latency and this kind of stuff doesn't really hit too much and that's why organizations dr. have that Many layers it's rarely the case that a model that is used is that it's usually that it was appropriate for the way things were and the world has changed in its moment of appropriate and that is the case yet you can see some of once you turn the speed dial up the
things on the right-hand side become of a problem if you build the layer architecture you need to add new UI element three times a day you're gonna be cursive a lot write about all these layers so you know Someone is in this right technical organization architecture sensing right I can changed the heavy does needs a friend you can apply this equally right now thinking of I suppose right in this case of memory we see this difference between structural damages and dynamic disadvantages and though most of the things that we use maybe when we are architects
in the penthouse and those are the things we understand the way we think about very different kind of Architecture but remember me at the penthouse are we dealing organizational structures senior executives etcetera nothing that we realize something up in the penthouse is that we need to look at the individual parts but we much money to look at the system overall and this is I have a famous saying and in architecture diagram that nine years to me is questionable right because the line is defining how the System functions okay I saw a few earlier so you
have another name any names well if you didn't have other lines in the architecture diagram think again miscellaneous defining the structure of the system and here's a classic system many organizations this will remind you of many organizations right how fast can these cars go unlike about 300 kilometers an hour well how fast do they go on a 25 and we've seen Organizations like this again a few very funny picture but a few very important serious thoughts behind it the first thought is that it's not about the individual pieces is about how the systems put together
that defines the performance so the good news is if you fix the system Cassation you can often speed things up without replacing the people of the cops right people can learn how the system works the stock in the service system so you can fix the System you can even move the externality the other thing you learn from this is how much does it have to put more horsepower in the car what nothing okay now what do all organizations would want to become additional what do they all say oh we need better people all we need
to hire somebody from Google right we need people that Google right they want to put more horsepower in their engines and it's not a time to have the systemic or administer is completely Pointless right and you know what you know is it driving around the people who Ferrari are gonna leave very quickly because they don't mind driving 25 km/h so you gotta fix the system before you can fix the ingredients again one of those architectural points of view that we know them but we merely implied an organization up in the 10th person so time to
time to go to the middle so now let us see how positions were this Layering assistance right all this organizational stuff we have a pretty good idea but now we need to deal with some real architectural stuff right we need to deal with our complexities and all these kind of things so it's a good time to ask what actually defines architecture when we all know some engineering institute has all been signed all the definitions of software architecture are listed right yes I usually its components relationships Constraints PHP I could say whatever they write is like
something like them write some rough definition so if we use doubles this was 25 all the components the constraints of the relationships are there I mean I put a lot of thought and its people I made sure the windows are in the wall we move this on the top right so this was snowing or raining right the door goes all the way to the ground so it's easy to go in enough so I did a great job I Talked about the components the relationships and constraints right so I should be architecture certified on the spot
however we talk about architecture in the business and the business has a different cross care so much about what the definition of architecture is they have a very different question is why would I have paid an architect for this and the answer is probably not so much and the Answer is it doesn't have any significant decisions it's like some plain vanilla and it's like this is what you do however is the architectural pictures like this girl these are the nature pictures are the equivalent of I have a front-end application takes all the evidence that most
reigns the relationships is all they write they only think that's massively missing is any meaningful decision and meaningful Decisions also have data sites right the decision that has no downside is probably not grading for CSS for house equally to a drawing you know credits no journey this house has max walls and because generally houses there are B number two halves are called dream analysis analysis we live in this house has a low overhanging groove to keep the Sun out however in the winter when it's actually not that hard it doesn't mean we should summarize that
the Sun Actually shines so what you become here from an architecture or regarding architecture one thing we learn is that good architectures are complicated we could architectures have like a thousand things relations handful of major decisions we make the walls of the class we could have all been a big roof so the Sun doesn't last night of the summer but in the winter we're having the Sun comes in it helps us warm a little bit second Thing we learned from this is is that context right this house may sense in certain area it doesn't make
sense in other areas Singapore for example we don't have summer winter right we on the equator this whole thing doesn't work right so context is a very important equation and if you think this is a little bit simple or simplistic this is for class architecture this these houses of them bone this is one of the case study homes in Los Angeles polarized I was called Jewish northern famous photographer you see in the picture every architect student will know this thing Hollywood Hills I call it most iconic building at Southern California right famous famous more things
right so Google architecture is simple to the point define something right and that's what basic things and great architectural time little things very very important right so it's about making decisions right in making simple Clear decisions understand the context and that's what makes great architecture right right now that's the tough message for architects people actually don't care about your architecture so much sorry guys right especially me I'm an architect right because people actually don't see you architecture the end users see the properties that your architecture gives the system does it scale is it easily modifiable
right all these Kind of things about the actual architecture and nobody sees so instead of the architecture keep up here about the properties your architecture gives the system now that's where the boxes and the lines come in we see two diagrams here example this they have the ingredients ABCD but they put together differently now do you think these two systems have different properties of course Tony it's almost the opposite of each other on the Left-hand side easy responsibility but maybe single point of failure long path right hand side is exactly the opposite short path has
a remote resilience but difficult to replace something you can make a list like this it's always exactly the opposite now this means people when you ask people what is the architecture and they see or equipment project or whatever it is right or via TV and something rather coach is right you realize how meaningless that is People list the ingredients but it says nothing about the architecture is I go to the restaurant and it says dream come look and tomatoes and you don't nobody can look at a pizza what are you gonna get a customer to
make the sauce or some sandwich or whatever mother right people need the medium of the ingredients right and likewise good ingredients I have four but the chef makes the restaurant right and the architect is like the chef are You going to cook you're going to put it together right five ingredients good idea but putting it together right that's what makes the world class chef and the world class architect misconceptions if you always think about that the developers Ryan is one of people who say of the functional staff and the architect with a non-functional idealities like
scalability maintainability all these kind of things And I'd say my new coil that model is a little bit too simplistic right the wood the job of the archetype is not that easy and I hinted at that with a house with the content there's this whole space surrounded of non requirements and the requirements are not things that are not required these not requirements of things that are never been to domineer assumptions tacit knowledge right contacts like all these things right that we're in somebody's Hit somewhere but we've never written down and that's where my view the
architects engage they see things the other people didn't see and they don't deal just with commitment on requirements they allow us to deal with the things that would never go down and that again makes the job of an architect so interesting but also so challenging it's a very different view especially he's changed transformation scenarios that we're talking about now Yeah I have one formula in my talk so when I tried to describe what really the role of Architects in a large organization is I sell options but this is my show us really drives nail products
right for pretty clever folks options work options financial options are and we to defer a decision [Music] by the financial option to buy our stock in one year for 200 euros today except maybe 180 190 so in one year Exactly I can find a career winery a couple of things to consider right a what one here comes the decision whether to buy it up is truly obvious because when the market is more than 200 of course I take the option to buy it for 200 at money the bank if it's less than 200 I won't
take the option so that teaches us that these decisions to the future is patentable because I have fully furnished a different decision as well let me fix the parameters is almost Always that people because I don't lose any parameters but I know so much more and this formula of course vanishes and the options have a price because they have a value now compare this to an architecture option but what's a great example of an architect option horizontal scaling right in the past people used to do services and the problem is surface sizing is that L
- I - possible comes right to make what you small and to make this money wasted CMOS Unhappy too small this is a guessing is a guessing right kind of game well in the deal world what we do is we scale horizontally as architects we know how to make this option very easily stateless elastic infrastructure cloud you know automatic deployment we all have this so we can sell this option now the option isn't free right there's some architectural complexity some work to be done and so the question now is should be behind This option as
I attack I can send this off supporters of the scale now the question is should the business behind it are free and not just a money containment money you take complexity also now the answer is relatively straightforward and the answers also the former he is the little Sigma square then remember this is like Sigma square obviously have a significant and that is the volatility right and you can easily see this their different decision into The future is more valuable the one circles you have if nothing ever changes make a decision today or tomorrow makes no
difference right but if you have no idea what comes tomorrow you'll be a lot smarter than today so volatility drives up the family of this object and is intuitively Mia clear for this horizontal scaling right if you look at the general application for three users or even these horizontal skinny if you've been in an online app That has spikes and Peaks and campaigns and all the kind of stuff right you have ten years of store today ten million users tomorrow while this option is usually damnable okay now the interesting thing is we live in a
world of high uncertainty and just having gone through this kind of thought in the world of science certainty architectures work for seven and sometimes people believe that an age of agility not certainty you cannot pack anything time People feel like everything I just saw go along rather just kind of loosey-goosey just kinda be see how it flows rate is exactly the opposite in the world but the high level of uncertainty architecture is even more important because architecture can give you those options that are more valuable and that takes us to the basement right to the
engine room let's say and the sub layer Sinatra stuff so I just have a few examples right all of the things that I Would say has fundamentally changed in the basement is how do we look at systems we always like to make robust systems right system stepped on brake systems organ break our drink until they break right if they break we are the state in Germany very good at this any call customer service they say ah I think ok what did they say this is impossible say what there you go man so robust this is
ready until something breaks some water systems are Resilient driving or a common feeling of a now what's important is that you build these systems very differently you build these on the infrastructure and this is mostly application and middle now we can see the more interesting if you have systems that we come up very easily what you can do is you inject failure to make these systems more easily because the more failure you have then the more failure scenarios you play through and Then you system actually becomes more easily now this is not just application base
this is also such the base which includes roots the people in the organization and the medicine what are they by all these is so vital energies for architecture realize one of the most famous articles is Starbucks does not use to face commitment right because they have a high-throughput messaging system and so Singh is here you know when you talk about the bus listen it's Always have the analogy of the firefighters you wonder the firefighters to prove when there's no fire they play cards and they polish the truck right but the other thing they do is
they set fires right they inject disturbance into the system now luckily they don't do this on the production system here this is my pre-production like staging right but the ante on the inject disturbance to make the system more resilient so when IT architectures this is the chaos Engineering the chaos monkey right we always feel it we invented all this stuff I feel just just get out more people just look at the real world right the real world is asynchronous inconsistent right it is resilient I just or forgot about IT architecture right another great example human
body right what will be two against diseases to begin injections immunization shots what does that do it ejects the disturbance To make the system more ism exactly the same thing and human part is probably one of the most successful systems we ever created so when I come to my impression of business Italy are in terms of changing both of our teeth you know the architect elevated what does it mean to think like an architect at a different levels of the penthouse now this swath of the dimension behind it as I said these modern architecture for
the first derivative right just all about Bringing change in wave of change one of the most difficult things is being that this earlier is that it's easy to go to an organization where things are not going very well and to just believe the people must be doing the wrong thing somebody must just like they don't work very hard at the dominant don't understand this stuff or you know some religious feels you must find the problem with something if you just poke around a lot you must Find the problem the catch is that most of the
time that's not how it works and here's a classic example hope everybody does exactly the right thing but the system is broken so let's say you have fan fiction in a system right just like a typical to get approval for arms of and proposals to be written when love kind of things right so what that means is doing things cost money and when doing things cost money there's more at stake Now what is the exactly right thing to do Morris at stake while you should have a few more checkpoints right if you were still too
many meetings a few more projects down something it's right is exactly the right thing to do now what does that lead to yeah more friction so just assume the friction is high enough that running a project here is $100,000 away and she's not such an unrealistic Puncher I need the project manager approvals but units like all that kind of stuff cost on the brain will people make $100,000 projects no right because he could pay a hundred percent tax on top so the right logical thing to do is make bigger projects one of the Nutella projects
now you can see how the story continues readiness though everybody has a billion dollar project well the logical thing to do is you have no controls right so that message I Really want to give to you is when you look at this I promise stuff right it's not the pieces is the sister that's on working everybody's absolute doing the right logical thing so you need to break the system and once you break the system you can get this running in the right direction so the question is why do we even deal with all this stuff
right a change is such a pain in the rare it's so difficult All these complexities people don't wanna change right and here's the toughest the toughest message is you see a mountain of gold right you know a better world when everything is fully automated DevOps agile in your cloud we are secure you know a better world and you have an organization not there the fact is you can make their life force in the short term you're gonna lead them into the swamp and Before they get to the bottom of them so therefore the question is
why did it do it and I have two concluding messages for any other site the first one is in obvious worse before it gets better but the second most important one is the words only becomes worse in my real example is why did he need people to leave that amendment global warming right the ocean levels are rising so it's much better to go through the smog today than to swim tomorrow because the Worst only becomes worse so the only way to deal with changes started today and tomorrow [Applause] okay in other words that where instead
of having the full stack of the operating system having a unit which is basically multiples all the time so there's a couple great to stick models that have they break the spirit right I'll start with an easy one easy Ones platform I like to make a bigger platform that's how all the common pieces are and on top you build the different chambers right we actually call that trap these days and if you think about it cloud is an organizational system and a technical system rain cloud is the other commercial outsourcing right that's one example so
in the unity kernels I would say it's a little bit more differentiated the key thing is all These transformations rely on other properties right and having these smaller loosely coupled things gives you a lot of desirable properties in this environment very much so and you see this people become free agents right by me right so so that model is becoming more popular what is important as an architect understand is what's the precondition and the precondition for them is actually lower friction right if you have low so the reason large Organizations work well and you know
this interesting post about life in the half organizations right we do in efforts why we just might not up and on our own the key argument is merging because if I'm gonna do business with shower then we have nothing to do together we need to negotiate make the contract money we have a lot of stuff to talk about versus forged buddies behind his foster just whatever right I just have to do or Vice versa so the friction is the key element and I think if you can get the organizational friction down then these microkernel kind
of organizations become possible and they deal with the department well last all is though if you probably that more if you say oh we should be like organization but you don't take the friction out of the organization you were so right that's why this is so dangerous people copy some other Structures they see without understanding the enabling factors and actually totally fine respect words and that's why this is an interesting so thank you for the question is there one final urgent question that could not be answered and close off the next two days thank you
[Music]