Merken

Microservices, a bittersweet symphony

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
do in the and the and the observation of
an of my name is
sebastian I even more out as the capital of Columbia South America I you can find me as similar so right in the range I would
1st part of right and we all are reinventing the way people commute to work so you should definitely check a set of 2 they
wouldn't about services so I only give a lecture about like services there's a bunch of them are the really did uh I should check them out you haven't out today want talk about so why why I think sometimes micro-services our by idea and but sometimes like a lot of times so that's our work
by defining what might services or of so much services are like a
particular way of assigning which and minus signs offered the creating and the like a lot of small applications that can be in the plane independently deployed so you will
have has discussed like is much there just service vector and Kennedy's muckers services are just like a subset of all services architecture have is just like with a set of it's just a piece of it with a set of really well defined rules out and you could say that mother services on the service erected architecture of what like its promise to add the only basically and the much resource or not just like the moon and you think that if 2 developers like talk about this far people that have been doing micro-services 14 years is just like of of war I name that was given to that that got subset of rules In the said intrasectoral like no cloud of the definition so there has been a lot of
controversy as to how big a much service should be so the problem with that is that like a services have a failure of a design and the failure its in its name it close the size of the service in the in the name instead confusion because like micro says that they should be really small and this size here is not the most important thing is just part of what 1 might serve as a sort but they could have probably been been given a better name so that people like
to say that I is is as mean that a service can be considered a much service the venting all hallmarks lines of code tests that's the best that can be confusing which is a lot of depending on this of language use and probably some of what some people will say that our universe the servicer might service at based on the size of the gene that's building on my service for example NMS and they have like it to be true and of that makes sense but I I don't really like those approaches this another type of approach of which are like really ugly tries to find rules for example I should follow likes to say that about services have to be this big and this big being like basically like how big your hand is on your screen so if it's bigger elliptical year bigger than that islamic service that's kind of a trading I know I realized that diffusion so 2 per cent of efficient I think that we should all well no like have in mind when talking about is a resist also delivered
about cognitive psychology so that's a concept income except psychology call cognitive load which refers to the amount of information a person is trying to process at any given time and there's another concept allergy introduce that's called kind limit which refer to the maximum number of chunks of information a person can process in working memory at any given time so these are really related so basically that I think they you can say that services my causes at the
cognitive load at service represents is at lower than the kind in the limit so this sucker little bit more about the kind to limit so basically the kinds of limit refers to publish information can hold in years at any given point in time so if you're able to out I completely wrap your mind around the service planets somewhat suing without the need to look into code point mutations you could call that service images out but you have to keep something in mind is that but when talking
about a group of people playing to the Limited Test to decrease as a team grows so basically living in seeing the the lower the coming to limit its this is a totally made our graph so the with 4 numbers or X and Y limits but as physically I think really really explains why I'm trying to say so before we go into more detail some of much services related stuff on still you like a personal story so that personal story
that will give us a little of context so the star is about the 1st l I had and that jobless
and we'll see of my jump insisted on a sunny hot boxing game days Internet I was getting so mad you asked me to still to all still form fingers and so that went OK uh but women you asked me to make some public announcement during and in between innings and some Chris but OK now had to also you that some really enough I was able to do all that and so I you asked me to mold Allowed before its game solid is it doing that also to the 4 each game a brightly manager last employee and so he asked me to help him equity people on the streets after more you log in what we will was like getting into the system for the name
of mn way but I'm sorry couldn't find like a better image of a teaching the Internet and I had to use 1 that's like for 1 direction so up because things were embody enough already at 1 direction showed up for the game had to accommodate them balance that really my left there on the job I had to be
but that's the the and I but Parliament is for you already realized that businesses at resource understand to make a point you out so in that order start I want to ask questions to everyone of this raise your hand if you have a system of floating going over your team cognitive load and you want to break it up breaking down into other systems OK cool thanks so this is not uncommon and this happens a lot when we're
going to the situation where we have a monolithic application so at this stage talked about this and earlier today and I can disagree with his comes about 1 of the competition so let's look at it a little bit let us look at rate so I
think that like a single rules obligation but having a system basically that's a single rose obligation doesn't make that system older than the application of monolithic application at so 1 makes single race obligation and when the link of monolithic application it basically poor object oriented design so let's like support this concept given the fact that the system is only composed but 1 rails add the they real 7 suddenly I see well sad can be like well designed and well built and there's actually like
at the top that was using the atingle review nails but again wants to add work to give a really in an army which was really interesting he talked about how to cook that which I think is a Japanese company have real what's probably there we are the world's largest rails obligation the the gives and on the talk is like really good inside of palliative that and like their mother profit they can handle with 1 rails obligation that's like running on an island and like the line but thing is like 300 servers are not so if you have an beginning it so although like I think they probably like the good to watch to keep the system being like a single rose obligation and I don't know if I go I will go to that extent after you could do it in if that's the way you like to roll up that's fine and that new applications similarly so you can call
the assistance you want that's fine you can give a name you wanted them out but let's be honest I wanna list doesn't
like this I would say I want to really looks
like this that makes more sense that's how well it
loads that's how you feel when you want to when the
monolithic application and the way
that we would throw in a monolithic application
act you probably you you know we
like I can't reuse of part of system
without basically like reusing the whole system
that or you can like easily change the
flow without having to show that
surgery which is really that
so now like yeah now and
now we know like the difference between what's really
what's really a monolithic application and what's
not when should you go for
Michael services our the analysts think a little bit about it it's interesting topic
so but at so should you go out for a monolayer I mean should you go from like services when you don't realize that you have I like a monolith that using the Royal Stat so longer well defined a world assigned ad by it's on Mali
well I would say no this is a
fallacy that a lot of people believing we have think everyone knows that we have a In
use pressured you get a bunch of little but has so so the fact that a lot of publication of the meaning you have to go for like services I found this intruder this is
something I really like and it says that if you can correctly assigned a single obligation you won't be able to like this and like is a is a system that favors based on like a service architecture and I told you it and you read that you 1st need to be able to nail a out a single as obligation before thinking about giving going for my services so when she
really like news like a services I mean is q what when not to but so when she did not so 1 the reason to going
to that direction is we have applied the system that needs to be escalated differently from the rest of the system are so this has happened to us in right out where we add subtle will be with a kind of like a rain application that was going on of things and we'll realize that some parts of the system needs to be up it to have a better performance so we had to do with the decision of the we want to scale the whole like brain obligation which is wonderful like services for the we out but can we let just that extract this and is collated differently because that's specific for this so this is a really good reason to go toward factors on the 1 it's when up apply the resistance
be easily replaceable so what happens we have bullets and what has happened like all of really big companies that have been doing that being Naprosyn suffered formative years is that some systems are like used 10 years although it 10 years old or built on languages that probably no 1 in the like the scene at the i given time knows how to right coding and spying server Hadamard dots for a lot that from a fallout of years and no wants to change because they're really scared about breaking stuff so we have stopped when when you have like you've his logic basically up spread about spread in different a small my services this is really it's really difficult for us to have because if you have like a service that it's feeling a some scaling of the ship and has a blog or whatever and itself acknowledges server that know what it seems to me that you understand the can is easily and replace it like this building in whatever language you want you want everything is fast and replace it with a you have liked to replace a runaway you can do it readily in testing and testing and I can ever said that the new service will really be able to replace the old 1 In this recall this is great because you need to do a degree right you have to like troll the 20 you old service the garbage and suffered scratch and probably create all the bugs so where'd effects on a huge application OK so the set about on other recent like you should think about going from my services and this is where you want to be
able to like deployed at some courses system more often than others and is it really common and like when we have a big real sad you notice that there's some tools are actually allow you to check this and the you know that there are some plots that have a greater like sharing that you changing all the time probably because like this is like the Sinulog are probably because it's 40 20 cents a day a lot of margin this specific what's it can be also because laughter the sum of all of reasons but sometimes you he had to change was the system but more often than others in way you have like it a big single at this means that you have to the point where at every time you change small part of a system and this is this costs a lot I also like the fear that the more you're afraid employed to production so you tend to like accumulate more changes the force employed and then when you commit a lot of changes which deployed this all of more possibilities of the planning to go back that way you just like the boy constantly smaller so when this happens to you but all the greatest thing OK so another way of
like the reason why you should look at my services it's when you want to do the same technology sector for every point in the system you have to be aware of that but this can be taken to extremes so there's a blog post out there by some kind of thing when they explain that up when they moved to America service architecture they allow anyone to write a new workers services and then like a service in any language they want to and this and being a huge problem for them because I don't want to services all my basically any like living up language other than a language the sources on the development and everyone to like re language so they basically like creating a new meta-service in that language which to be really cool but can be also really hard to handle so for example a ride and why we do as when we need to create a new services are Weber where was going to work at all who what kind of waves by people where respond to be responsible for it but it creates like a proposal and said OK record pneumothoraces this it's like that's contractors stressful stability and I want to use this text set to build it and then we open it for proposals so sorry for common so we do not see a request for comments and the everyone in the in the in the engineering team has it period of time when they can give feedback proposal and that feedback is some it areas on the sign to it's like it has to do with like the business value that that is regular resist it's like a service at few of like the whole of Part I was holding but adults also around but the millisecond use because the whole thing is good responsible to mean of maintaining that my new much of a system to create so that we're prone wanted to be reading something that makes sense for the problem effects and but also something else that we are mean yes but that we want to maintain something that we that we think makes sense and that at selecting new crystal language something just someone just once a learning that Truman with so that's called of technology at their already in the city I that Edwards confusing to so last but not least
of the most important thing is that the thing that team just working on building a system has to be ready to support this type of architecture and that's the key thing is it's like really important in this and it's was 1 our people moving towards services mess the skill set of the team has to be the ready to support 1 more into micro-services means which will talk about a little bit uh printed in the stock and I also like the team has to be big enough to be able to walking different service atoms rid of Michigan as resistance and some testable over resulting with this type of architecture and that's when they start to see a lot of problems with that so you have to be really aware of the of this so how do you know as it used to be ready to up go toward selection a lecture this visited like this things start key the 1st
thing I you have to be able to like is the provision new servers how to rot there really raised the shipper provision servers whenever you need to get out the 2nd 1 is the data to be able to set a basic withering for every new much of you create an existing ones LC in the unit to be able to respond to failures that that monitoring the use of our backs and last there is a need to be able to like rapidly deploying applications so that if you have a numerical service that tend to reduction issue legatine should be able to get it into production in a few minutes at a few hours few days few weeks so that's really important if you do not there yet you're probably better off not going towards much of architecture so the now that we have in let's let's say that team is ready to handle all of this over the summer which is great of you could say OK I'm ready let's go over market services I really love the idea of it I think we give it a really long taught and discuss about it so let's do it but you have to consider the downsides of it and
they're not trivial so Our let's get to let's see would like the most common stones size of what services or so 1st of it is that that also
intensive so measure before you seems to to be good loves and to be ready queue up support all our services flying all the time and this can kind of be mitigated by using a a plethora observers which will probably do a lot of the work for you and and as good as that's we use it to go the is fine of another downside of having
more than the downside that's like something you need to be aware but let's talk about this so you have a log files for every services trying and this can make really there in this can make dividing really complicated hot so you to have a central like surprise loss for the whole system so that you can like watch what's going on there and also something as really important is that you probably want to have a unique and fires for every request as being that that's going to this to other systems so that if something goes wrong can issue traces and figure out when were thing so belonging how the fix this is really important so I onset of 1 services is that you have to be
prepared for failure so I know I mean this is not this is something you should only that you should only worry about where you have my services like you should also be prepared for the year we have I'm like a single rail a single at if it's real sort of this month of but the thing is that we have like a services this is hard to help you manage you need to have a lot more stuff into consideration to be prepare for failure so let's look at what are the most common failures when you're dealing with AltLex's objection so the 1st one is never of
partitions so that is something that will events individually have you know I mean sorry this is something that you will inevitably have to deal with when working with this real systems event if they're not like services out so this is very important because there's no such thing as and they were undevoted doesn't fail on services that never goes down will not sometimes it's like sometimes these things felt failed so we need to reverse for so for example
let's say we have a system that looks like this out essay that communication with between out node or service in the those that and then we get a request and no aid of also mold a has up to the decide between basically 2 options no they can respond to the clients telling them like Hey this like something wrong going on I can't process request right now out all know they can also I said OK I'm going to do with this no respond to something and then I will internally deal with the consequences of that having no communication to node C will bring to the system so this is something a century or something and you need to fight had but it so that the thing is that this another problem
where you can only you can also have services that are of some dozen available so even if you have the perfect network that never goes down services will eventually be unavailable so this is formula for at
says that basically like the probability of failure requiring the existence of this equal to 1 minus the probability of a node of service not failing the poleward to how many of those or this is like well I discovered it when I
1st saw this it was really like hard to get my in my head around so let's just see example let's
say you have so which much services all of them have enough time of 99 . 9 % which is the act but you have for like services 30 40 my customers this means that the system applies formula was a saw of
out there is 1 the system would
have 0 under validity of 19 6 . 1 % which I would mean saturday you will have around 4 per cent unlike most of the time you have a probability of of of office and of some of things going wrong and this is only considering like that if a node fails or service fails only that's service will fails but the depending on how you sign your system if it's if a node fails on a lot of other related fields can fail to so this again really that you need to be prepared to handle this set of failures so let's look at another 1 bad
inconsistencies so when then when network it's petitioned was services are available there will most probably a little more forestry cost they had to be inconsistent and this requires planning you need to know know how you deal with this so that they face eventually consistent or if you don't care about consistency in like maybe event like how to deal with this help you would you in of inconsistent use system and this is not trivial because
it can be it can really make a system like not work
as it should at OK and I like this 2 problems with this talk about energy the point of
something it's called the she keep hearing more capturing at which is which talks about like dealing with this type of failures in this systems and really explains warlike not only have how they can affect the system but also Prodi how you can do with them depending on the distance requirements that's the most important thing so basically what CAP sense for consistency of Bolivian partition boards and we have this real systems such as a maximal expected you have to deal with the petition tolerance because the system is partitioned it's not something you can choose but you have to balance how to how to about be available and also consist like have consistency they are at same time so but this is really important to get all services if you haven't heard about the capturing before you should really read about I recommend of work we recommend to you stranger saying but this is off topic big enough to do and I'll talk about it so this basically about the
like what to do like when you want to build my services let's say up you tears ready you can deal with all the bells itself that a you're ready for it so how'd so so 1st of
all of when you're trying to be to break out of the obligation to a small with small there is this use the barley kind expired which basically deals with large models by giving them the ability to have specific behaviors based on the context or the combustor on so like in this
image you see that I we have the customer model and a product model in 2 different contexts 1 of them is like the sales can't they can't texts and it's word context depending on the context it's of this model as Bill Prodi have very labor and you probably once you have access to different data that's related to figure out string of out so also something you really need to do that I
found really useful it's having specification 48 micro service and being able to run acidification as executable tests so 1 of the like it can be a pain in the average of the 2 the testing with of so so 1 way you can handle this and make it easier to work with s by having specs that you can run as executable conference so that when you test it's what customers you can test again that execute will come contract and assured that when everything's trying together but everything's deployed it will do what find together and you want to be like only realize you have box and you you go to collection which is really that and it doesn't mind if you're using an like a to the purest way PI or you're using like communicate for communicating using them as a broker or or perceived this is something you really need to go through the 3rd well save you a lot of headaches so you should also be aware that she should go
for synchronous communication or synchronous communication whatever you whenever you can this will make you make it easier to deal with any type of failure some of it is we just saw the capturing stuck tech about and the other thing that I found really useful in and that we've learned in right that you should based formats over
stimulus ones so there's so like to give you a little of convex for example like a stimulus of formant as Jason and the this is the most basic of the formant can be up prickle buffers are kept in over or under the surely a bunch of them on of the so the good thing about this is
that of steamer based formats allows you to extend the you pass between systems as new cuirasses Rosen changes and it'll it's also I mean it also have some ages who and will also all of you to be backward-compatible and grow and this is something that's really important and will talk about something called the golden rule fighters services in a few minutes so
what you need to start but creating new creating because was when you need not be aware that the perfect as any of the get so you need to have the perfect design before going to much of this is the start of the afraid to it's a it's a solution
so what should you do now that we know what to so 3 boys
you don't shared database written letters of assist sharing databases is like calling on another object problem and that's that's really that something you want to sort of rank encapsulation so don't do this there's a lot of ways around this and if you're doing it is folly because you want hard enough of how to fix the problem if you if you and if you use 7 situation where you'd like temporarily hopefully have to deal with this just make 1 because it was in charge of writing to the database and the rest of them just ring from it but still you'll have a lot of coupling because in that schema changes all of this like is a resistor reading from it will have to change too we like stayed
on you share like a database it will you'll feel like that's if the letter moving on
menus rounds so what something we're
solver for today as to up to avoid being biased by the like always log n this is basically it up this like law was coined by some something that barring Cohen CAN and 1968 he said that
organizations for the systems with the sign is a copy of the structure of the of the organization so this should be really aware of this and not try to just make replicated their way organization words into how the system works because that's following that's only gonna have you know creating what is a position or the same the wages and most importantly we don't
break I mentioned before don't break the golden rule services and the visit assess that need to be able to employ any
service any time without changing any other service and is that through literature because if you failed to to achieve this you're losing most of micro-services benefits if you are you find yourself in a position where you are doing like lockstep deployments we have to employ more than 1 service at a time so the system doesn't break down it doesn't really make sense to have services that they really have to make sense to have things that you should be able to the point that only change as much as you want well the SOC snowing so we are aware of this if you find yourself in the situation where you have to fix a real crack or maybe think about going back into like a single thing about it so true like corrupt everything and been saying up but I will like to bring 2 quotes might Michael Feathers out and the 1st 1 says that
he strongly like she's frontally believes that there's a lot of complexity conservation suffer which means that for example monkeys services
out we tube aware in to be aware that when we go toward Federation out we're pushing complexity into the interaction Over services and we need to be prepared to the with that so my
recommendation s goal for like services if they allow you to better manage the complexity of the system that you are building
that's all got for today's they things on their
well known number of some of the things in and out of the more
Spannweite <Stochastik>
Menge
Rechter Winkel
Mereologie
Besprechung/Interview
Kommutator <Quantentheorie>
Computeranimation
Dienst <Informatik>
Computeranimation
Ebene
Kartesische Koordinaten
Schlussregel
Dienst <Informatik>
Vektorraum
Menge
Computeranimation
Teilmenge
Serviceorientierte Architektur
Software
Dienst <Informatik>
Web Services
Menge
Vorzeichen <Mathematik>
Softwareentwickler
Unternehmensarchitektur
Streuungsdiagramm
Web Services
Softwaretest
Formale Sprache
Gebäude <Mathematik>
Schlussregel
Quick-Sort
Code
Computeranimation
Arithmetisches Mittel
Dienst <Informatik>
Web Services
Mereologie
Datentyp
Grundraum
Gerade
Touchscreen
Bit
Punkt
Prozess <Physik>
Extrempunkt
Physikalischer Effekt
Informationsverarbeitung
Zahlenbereich
Systemaufruf
Ausnahmebehandlung
Code
Computeranimation
Inverser Limes
Dienst <Informatik>
Web Services
Informationsverarbeitung
Last
Festspeicher
Inverser Limes
Information
Bildgebendes Verfahren
Softwaretest
Dienst <Informatik>
Graph
Gruppenkeim
Zahlenbereich
Inverser Limes
Kontextbezogenes System
Computeranimation
Wechselsprung
Bildschirmmaske
Quader
Prozess <Informatik>
Spieltheorie
ABEL <Programmiersprache>
Identitätsverwaltung
Physikalisches System
Störungstheorie
Bildgebendes Verfahren
Richtung
Internetworking
Bit
Punkt
Informationsverarbeitung
Last
Kartesische Koordinaten
Physikalisches System
Bitrate
Ordnung <Mathematik>
Computeranimation
Singularität <Mathematik>
Server
Einfache Genauigkeit
Schlussregel
Kartesische Koordinaten
Physikalisches System
Maßerweiterung
Binder <Informatik>
Gerade
Computeranimation
Systemprogrammierung
Rundung
Last
Desintegration <Mathematik>
Besprechung/Interview
Notepad-Computer
Kartesische Koordinaten
Mailing-Liste
Computeranimation
Subtraktion
Chirurgie <Mathematik>
Rundung
Mereologie
Besprechung/Interview
Vorlesung/Konferenz
Kartesische Koordinaten
Physikalisches System
Datenfluss
Dienst <Informatik>
Bit
Vorlesung/Konferenz
Statistische Analyse
Computeranimation
Arithmetisches Mittel
Dienst <Informatik>
Druckverlauf
Web Services
Dienst <Informatik>
Physikalisches System
Unternehmensarchitektur
Computeranimation
Umwandlungsenthalpie
SCI <Informatik>
Kartesische Koordinaten
Physikalisches System
Teilbarkeit
Computeranimation
Entscheidungstheorie
Richtung
Dienst <Informatik>
Umwandlungsenthalpie
Maßstab
Rechter Winkel
Mereologie
Randverteilung
Mereologie
Punkt
Gewichtete Summe
Web log
Formale Sprache
Mathematisierung
Automatische Handlungsplanung
Kartesische Koordinaten
Mathematische Logik
Code
Computeranimation
Demoszene <Programmierung>
Web Services
Dämpfung
Softwaretest
Soundverarbeitung
Zentrische Streckung
Gebäude <Mathematik>
Plot <Graphische Darstellung>
Physikalisches System
Biprodukt
Programmfehler
Skalarprodukt
Dienst <Informatik>
Minimalgrad
Forcing
Rechter Winkel
Mereologie
Server
Hadamard-Matrix
Zeitzone
Rückkopplung
Subtraktion
Stabilitätstheorie <Logik>
Bit
Punkt
Web log
Wellenlehre
Formale Sprache
Ordinalzahl
Computeranimation
Datensatz
Web Services
Vorzeichen <Mathematik>
Trennschärfe <Statistik>
Endogene Variable
Datentyp
Softwareentwickler
Soundverarbeitung
Quellcode
Physikalisches System
Frequenz
Dienst <Informatik>
Flächeninhalt
Menge
Mereologie
Extreme programming
Unternehmensarchitektur
Normalspannung
Dienst <Informatik>
Einheit <Mathematik>
Web Services
Server
Biprodukt
Unternehmensarchitektur
Ordnungsreduktion
Computeranimation
Eins
Zentralisator
Einfügungsdämpfung
Dienst <Informatik>
Elektronische Publikation
Datenlogger
Luenberger-Beobachter
Warteschlange
Oval
Physikalisches System
Ablaufverfolgung
Figurierte Zahl
Einflussgröße
Computeranimation
Objekt <Kategorie>
Partitionsfunktion
Dienst <Informatik>
Reelle Zahl
Physikalisches System
Partitionsfunktion
Ereignishorizont
Quick-Sort
Computeranimation
Web Services
Telekommunikation
Dienst <Informatik>
Knotenmenge
Client
Web Services
Prozess <Physik>
Datennetz
Endogene Variable
Physikalisches System
Computeranimation
Konfiguration <Informatik>
Knotenmenge
Web Services
Existenzsatz
Knotenmenge
Computeranimation
Schreib-Lese-Kopf
Relativitätstheorie
Validität
Physikalisches System
Dienst <Informatik>
Frequenz
Computeranimation
Ausdruck <Logik>
Office-Paket
Dienst <Informatik>
Knotenmenge
Web Services
Datenfeld
Menge
Dienst <Informatik>
Datennetz
Automatische Handlungsplanung
Physikalisches System
Widerspruchsfreiheit
Hilfesystem
Widerspruchsfreiheit
Ereignishorizont
Computeranimation
Energiedichte
Dienst <Informatik>
Extrempunkt
Datentyp
Abstand
Physikalisches System
Partitionsfunktion
Whiteboard
Widerspruchsfreiheit
Computeranimation
Kugelkappe
Gebundener Zustand
Mustersprache
Dienst <Informatik>
Modelltheorie
Kontextbezogenes System
Kontextbezogenes System
Computeranimation
Softwaretest
Umwandlungsenthalpie
Schreiben <Datenverarbeitung>
Subtraktion
Quader
Dienst <Informatik>
Kontextbezogenes System
Computeranimation
Design by Contract
Arbeit <Physik>
Softwaretest
Web Services
Mittelwert
Produkt <Mathematik>
Wort <Informatik>
Modelltheorie
Bildgebendes Verfahren
Zeichenkette
Puffer <Netzplantechnik>
Rechter Winkel
Konvexer Körper
Datentyp
Bildschirmsymbol
Dateiformat
Systemaufruf
Synchronisierung
Computeranimation
Eins
Dienst <Informatik>
Code
Mathematisierung
Dateiformat
Physikalisches System
Computeranimation
Objekt <Kategorie>
Rangstatistik
Datenhaltung
Geheimnisprinzip
Quick-Sort
Computeranimation
Datenhaltung
Lesen <Datenverarbeitung>
Datenhaltung
Unrundheit
Oval
Gesetz <Physik>
Computeranimation
Dienst <Informatik>
Ortsoperator
Vorzeichen <Mathematik>
Selbst organisierendes System
Kontrollstruktur
Wort <Informatik>
Physikalisches System
Datenstruktur
Computeranimation
Punkt
Ortsoperator
Mathematisierung
Kolmogorov-Komplexität
Physikalisches System
Komplex <Algebra>
Computeranimation
Software
Dienst <Informatik>
Web Services
System-on-Chip
Erhaltungssatz
Kontrollstruktur
Energieerhaltung
Physikalisches System
Dienst <Informatik>
Interaktives Fernsehen
Gebäude <Mathematik>
Kolmogorov-Komplexität
Physikalisches System
Kontextbezogenes System
Komplex <Algebra>
Computeranimation
Videokonferenz
Zahlenbereich
Vorlesung/Konferenz

Metadaten

Formale Metadaten

Titel Microservices, a bittersweet symphony
Serientitel RailsConf 2015
Teil 31
Anzahl der Teile 94
Autor Sogamoso, Sebastian
Lizenz CC-Namensnennung - Weitergabe unter gleichen Bedingungen 3.0 Unported:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen und nicht-kommerziellen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen und das Werk bzw. diesen Inhalt auch in veränderter Form nur unter den Bedingungen dieser Lizenz weitergeben.
DOI 10.5446/30683
Herausgeber Confreaks, LLC
Erscheinungsjahr 2015
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Like an espresso, code is better served in small portions. Unfortunately most of the time we build systems consisting of a monolithic application that gets bigger and scarier by the day. Fortunately there are a few ways to solve this problem. Everyone talks about how good microservices are. At a first glance an architecture of small independently deployable services seems ideal, but it's no free lunch, it comes with some drawbacks. In this talk we'll see how microservices help us think differently about writing code and solving problems, and why they are not always the right answer.

Ähnliche Filme

Loading...