Bestand wählen
Merken

How to Drive a Delivery Truck with Insurance in Africa

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
scope often in everybody and the officer's going the is talk around I experience and how we putting shift delivery and some of the fun and games we had and say I talk is how to drug delivery truck in Africa
and this is actually a legitimate picture of a corner was pulled over in Johannesburg also it's a minibus taxi and a little bit about us and
so we're really small team this this 3 of us 1 manager ourself to automation years and then we look after everything we look up the provisioning shaved delivery monitoring in you can think of some somehow our team has responsibility for that uh and most importantly our teams motto is don't get chopped such phenomena as node that means means that it will get a little about
how old so we work standard bank because bank in Africa and we've been around for 150 3 years we've got over 60 thousand employees and and that's that's a really at the big environment that's that's really hard to change things we really siloed changes hard you know when when we think about governance and a
bank and things dioxolan all those wonderful people 500 people have to review change before you can get anything done and we have to we have to stick to PCI and all of that with with core customer data and we got popular over us at the moment or customer then we can and can't show uh you and and lots of just resistance to change and we've got a highly highly regulated and financial environments are reserve banks really strict we can't put any data into the cloud uh we con host anything off sure it's it's really really tricky but but I classic OK
and we'll send this problem so from an IT perspective where but about a child of the nineties for a long time we we never ever tried it internally engineer anything we always would go to 1 of the big players in the like just blank checks so something to solve our problems and generally as you all know that tends to not really go too well um and then you know we generally also bodies these products and from vendors customize them into oblivion and then they can never ever be upgraded again so then I think
there were seen this and we can just silos birds and was pretty much cash occur how things work at centre-back um so the does gave 2 sons to work on stuff Westerners totalled the wall of garlic after granites because what but but actually it's kind of what you wanna get through with your delivery pipeline because you want your data is right occurred and not really worrying too much and making it really easy for them to get the course delivered which I think is wise like the throwing stuff over the wall so if the Government shall pipeline works really well in back than than its ICT get stuff into the into the environment so what is a pipeline
um for us it's obviously rest
since this is a classic deployment pipeline that we will use in the past and so the you brought to have that stuff that a developer would be able to do that premaxilla automated awesome that's red blocks of that's an act shift to give something or ask someone else to do something and then have to wait for them to do before it can continue with whatever introduced so in the sun the developer commits code that work through all then I've Andrew pick it up a lot effects and publishers and then develop reported that their environments make sure it's working really happy he actually test themselves that's not automated because part of our side as 1 of the silences us separates thing division review testing developers not so much so so here because under test usually just some of features written not any kind of regression or anything and then and then the fun against starts the now I want to go into the test environment inactive this Laura change say on a plot discovered is that OK this is what it's going to do this is my change number please approved and then depending on what application that is 1 or 2 or 3 or 1 of many people have to eventually see it's in their workload and I'll prove that change some article do you call it right so then you'll know take that artifacts due to the UPS guy's of Scott's please these releases for me some of cool Ulvskog edits and then was like you and like 0 right this there's something wrong with this change member the schema doesn't match the right number for this environments are you have to you have to get a new change for those who exchange numbers wrong and stuff like that happening yeah but eventually of God was to test then is gonna go back to the test as be like I'm artifacts the blood test the guys please please check if this is working correctly filled the tests same thing and broad log another change if change number word for approval get the artifacts and hopefully at the end of the day unaffected that he builds is the same 1 that gets into environments not always the case that the going for integrated so we decided to sample pipelines let's let's do this so we got a delivery truck growing but we knew that they
were going to be equal for the delivery at this stage we have a pipeline having some speech at some things that we could not removed from this process um wasn't just gonna be magically perfect somewhere in the 1st so mean the things change management process that was gonna be something that we don't have the ability to change that's not going to change very quickly will have to prove the concept before we can get rid of that and massive governance checklists and also because of the testing insider and where it being things and this no no automated test because that is hinting at because all the testimony so this was different theories stuff that was going slow us down but not sure attractive concept so it the OK so
just a little bit about our journey and we we started trying to do the develops in November 2014 and before that lacks the further the previous process lots of lots of manual changes infrastructure totally different he at custody differently so so we started with chopped up which was kind of like our 1st attempt at building a t where we had some thoughts and their guys and and and I'm goal was to get internet banking into a pilot that could be used by internal staff in the bank by February which using shift using some of the tools of that gas street we managed to do what we did then later managed to get that into project for customers and and you know we then this article this work really well we're gonna do this poll for everything across the back so going back to what is a popular question they we useful kind of figuring out what apartment should look like but you know you got to google was apart plan with what about b and this is the 1st thing that pops up I think most importantly for us is that you you definitely 1 the building something from from version control you you wanna make sure you do some sort tests you 1 ultimately deployed and that the most important thing is you constantly 1 have feedback coming back and saying this works this doesn't this needs to be changed that this can go ahead so this is this is kind of what people told us apartment should be no you don't get burned you shot
slowly about slightly different so we we had 3 main packages a wave front end and unfortunately a Java application back and and we had some sort of configuration data and what we used to do is any time someone made a change to 1 of those we we would take that that code with boulders and with units and integration testing against that and it was all 3 together was successful we would packages now because useful trying to get around the whole governance trick we we had 2 different parts sort of uh a taste pipeline in part to an idea would be if manual testing was happening in in the top part of the which is the green part we wouldn't push changes down to the bottom part and then as as the bad taste is found bugs and stuff what you would then you say well we actually have explicit running in an environment and we would switch switch concepts that environment pretty much going into production you something very similar way someone told us you have to come in at 2 o'clock in the morning to deploy and we like that is the worst idea ever my job description so we just ran 2 copies of prod and idea would be we have 1 active to customers and win when it came time to to deploy new vision readers which it wanted a passive vision and that would not be brought and yes it would greatly had 0 downtime to customers if something did manage to get to production that was that we could quickly roll back and we we started our automated testing Janie and and an apple was fully automated and brought and deployable but things weren't entirely
perfect that the biggest thing not
everything was going through the pipeline so all I infrastructure changes were kind of a little bit ad-hoc and cookbooks people just append random versions or just upload uploaded cookbooks and suddenly you got the latest cookbook that could be anything and the and any other which is in some configuration changes that what happens in environment that that when going through our pipeline the other thing even and it would
probably disagree with us is that we we ran into this big scale problem is that we both as 1 custom pipeline and but there was a manual pipeline we built a fund lots of hacky scripts and you know we did all it could to click drag dragon venue and but the challenge we had was that as soon as we try to add 2 3 4 5 10 pipelines it was there's just no way to scale this up at all of that was really closed impossible and also that the challenge that we had it if you look that up our our our problem that black is you had to grab all 3 or artifact before you connect you run any tests for the 1st time so often useful waiting for a build of some job lot effect when you had a city typo on on your front and as an example also some of the other
fun and games we ran into is that when we did try and create lots of other pipelines we had different folks updating different topics so for example and where we're big Jabal shop and what would happen is vision 1 of J was great someone would take version 101 and Oracle DB support someone would then take version 1 make a provision wanted to and add support for something else and certainly when we try to go division 1 3 there was just no way to get the 2 to merge back together and we also had a problem that a lot of teams didn't wanna do any environment opening and shifts so they were really just getting cookbook surprises all the time the at end of the the like is it people would fall for cookbooks and I you'd have 3 or 4 or 5 different branches of cookbook and then all wanna try and uploaded to the same shit for also probably out at this the
biggest by being so there were some smart guys it figured out really quickly they needed to print cookbooks but then I can cook books and then ever change them again and suddenly we had a problem with people were 2 major visions behind a cookbook nothing would never like quick can you fix as mediator production of the so often come to ship Conf last year we we were
enlightened we we found only grounding shift delivery all to automate now apparently and I was tossed to to go
do delivery truck mechanic School in Seattle and figure out how we get this working in i environments always always very likely
head silly Mister Fantastic ensure anyone who knows in about deliveries work in the actually in the room at the time he was the came resume and he just magically every now and then we get stuck with something and you hear his voice from the heavens and our problem would magically be fixed we also had some of the celebrities sitting in the front John in the room having to figure out how we deal with this very awesome error message at the bottom of how do you get anything to talk to the internet way kind of delivery when you install it with every cluster at once really James of one's packages at once those of ones that on it would no nitric axis so that was quite challenging and then lost the just you know how you get anything done in office with your own portable be Akaike this and I probably within a month than 100 feet which senseless and we've got some ideas you can contact us after the talk and there are so so can directly but in pipelines together we had some tests on the that and didn't Seattle and that and then the
kind of set like that for long so other has anyone to get on if you haven't you they should beautiful beautiful city 1 of the things we have there is this bridge which is it's right
near the holiday just below the mountain
and was meant to be a however that's connected 1 side of the city to the outside of the city is another portion of it just like that on the other side and just got abandoned and never completed and that was pretty much the state of art delivery and datasets at 1 point spent and we had a bunch of reports going through its we so that there's going Frankfurt the wall there was very little testing and and restore going man of the poison to production so so it kind of and and we started working on other stuff and then after a while we decided to look at it again we started looking around for pilot projects so we went to
summer teens with some applications were running the 1st thing we consider would be a good pilots was considered banking which is the service that we should department for problem was there any had waking pipeline into antiquing change things up and then you get to the rest of Africa guys % but problem that too many stakeholders to many like the from things going on too much regulation and different acquirements all countries so there was wouldn't God mobile banking we had there's a sum the pipelines you had a really belts that winter going production but it was a bunch of smaller services but also that their impending I'm going to prod at that time so eventually came across so crazy guy from insurance have small team and he's is willing to give it a shot so his stack
goes on on running running and hats and Marcus said was in the so and national and Apache of from so the what here was living job that we were using bonded to global artifacts and stuff and there wasn't about support for shift from ships delivery for jobless years and have to do some work here and that's our got office literature
analysts I mean it was an amazing but shows kind of working so we so let us see I happening and then so Dowager concurred but then we would pick up the change from made and stuff but not affected but effective nexus and then once the that it would pump to the version and then the metadata for the corporate that was going to be pushed through delivery and then that would trigger pipeline and then kick from person into the environment which was I mean it was working class but
that shepherd down and of problems we had was that that's that was kind of pointless to MCI portion outside of the literary because you can't actually see with the changes all listeners version but by the same user every time so they knew not too sure what's going on and at the time of the guy store using snapshots so all effective building because of and other from our previous process where there were manually I mean physically give effect like a guy is a platform to deploy this for me and so that means snapshots when issue in that model but now they were becoming an issue because of you the node running union Norris or something and starting from a snapshot somebody change that's not but it's not meant to be deployed and somebody else forces which change to a basic work that node gets converted suddenly from so surprising new environments so and so on and then the other thing will be pushing for purposes of the research could Macy was changing the review portion of the work
so in the figure that we need to learn some stuff so the 1st thing we learned from us version all the
things we needed to envision everything so we needed a proper repository for artifacts we needed to not use snapshots anymore on when it's tag Rosso's source code stuff and we went and all that and then went back and read a little
objects in checked about a bit of a new delivery reach new processing pipeline started working effects I stolen test but it was working pretty wrong are we still have the same thing boundary triggering delivery so artifact bowl was the same projects incurred benefits convolves at and now instead of Persian connects we just pull out effects which have the bill member and was really version from from bamboo it seemed like a pretty good idea at the time we had an acceptance mode and the obvious still and pushing up reports to delivery but it was was working but that the but the trick of
affect us so what happened was banned really not really an artifact repository great for the NCI but those artifacts that persist on the server so if enough builds happened even other teams those artifacts of disappear something 10 3 1 of the basic works and is not affected so that was really bad and the other thing is so figuring things out so there's a lot of experimentation on on the words a lot of hands on their hands on keyboards of figuring stuff are consistent so sometimes you get nodes mood states an arm chair of copper snobbery images so it can really recover that well from some somewhere conflict situations yeah so had some more
learning from the so I went back to the drawing board decided we need automated belts we needed
really good machine so we went and together service where we records kinetochore up machines and sit there and listen environments at that time so that couple configured maybe and part of the pipeline which was really key which most of the things you saw wouldn't have done instead it's kind of added pipeline recent trickery on things I love about the results of this should applied I still another problem
not much testing so went back and focus colonic testing got the guys start writing tests of giving feedback from tests and 1 of things we wanted to do for mitosis fertile as soon as possible so what are the units of sound and stuff happening acceptance
OK so the next version of of of our delivery truck for be about attempts 42 because they were a coupling between that we haven't covered and what we decided he was there a badly adjust that was nothing trying to have a CI system talk to use the system just the work and the fact that delivery had CI and it just kind of made sense to put out and also also the other really important thing here was that we had to make not just books go through them you wheelsets coming through to delivery we had to make the code go through delivery as well so we entirely removed we mean that we and also from a from a testing perspective we implemented some lands and syntax testing which really means on home weights to picking up areas and and getting problems out of the way and and then you know we we sort of added more unit if and and we kept stripe which I integration tests and and this is pretty close to the pipeline we have today and I was pretty good but as always we don't currently a few things from this so what are the
challenges in i environment is we don't have access to AWS and and we don't even just have like a massive internal OpenStack cloud as 1 of the things that that is a really good thing about delivery but for every cookbook or every piece of code pushed through it you need you need an acceptance node or maybe even multiple acceptance nodes and we just couldn't really parent node for every single cookbook or every single piece of code so what we did is we built a service called get off and using some of the automated server both stuff that we had with the previously an idea is now at the re does is a calls can often cities and is there an acceptance node for the spot if the isn't he spends 1 that gives it to the God runs on these tests if that successful that then kills the nodal and that change is ready to be delivered if it's not successful the developers got 24 hours to log onto that box and and use it to troubleshoot the problem the reason being that also are suitable sometimes a little bit different to what you can easily reproducing kitchen so we have a lot of tools that machines have to be registered to but you can't always register kitchen nodes to settle the tricky to sometimes because of our kitchen so we found that worked really well for us to get get off to workload when it comes to
contract 1 more time so we we we we react on delivery track and so that you know any new awesome updates from shift that were part of the community cookbook which will be only get those but that we could kind of make global updates across all our pipelines without having to try and teach its cookbook an intron retrofit their changes every time they came back also once
we made our application code through the pipeline we realized tracks great morally people to application code so we bow copy track because well copy trapped lives around and some cool stuff it does is that it's got separate maven profiles so that you can run because of how horrible maven is the way it does once through its testing at once now we can break up the tasting into all the different phases which is great and we also use deliveries version management system for managing and version numbers in artifacts makes life a lot easier now that those opened into the environment and also we push all our results to so under which we'll talk about it in
and we did school can run into us they are same favorite problem the shift Gospel Henry shame when they talk about this but they are we there was this the usual change management theater kind of boys won the day and it is no way that the guys mannerliness go straight into u production at the 2nd challenge we also had is that for us since we run J. Boston job and you know we deploy summing this down time you have wait for restarts it's a bit of a mess so for anyone has looked at
this what I would definitely recommend is don't waste your time with all those wonderful products from IBM and look at something simple like cheap proxy and we we did this body or on a much higher in the banking platform on a much simpler
scale way we we just bought something simple where you had sort of statically active environment in a passive environment and we would just kind of spatial games with them and switch between the 2 but for this scenario we needed something a little bit more complex so this is where our current that looks like and we treat delivered is like pre-project sigh In fact changes that she or going straight to production and they can be looked at and and we can make sure that the 100 % we we decide that change is good and we run a switch way those we idea proxy nodes now start sending traffic to or those nodes basically new up environments to i active environment and H. approach start sending in life customer traffic and what was in active use into into our passive environment in bed so that you know you know in case you've accidentally delivered something and you need to go back you can and and then what happens is our nodes that word airbag keep the Commission we both and then they get put into the delivery past environments not not exactly pure pure continuous delivery but we find it works pretty well for us yeah the answer of soulless changes on the delivery process from the fall that because the more like us so here we got liver I mean developer checks and God islands deliberatively and then there was land syntaxin Access and Delivery preserve the code review some yes come 1 of the other guys in the team check the button check button but then we have a good future that both artifacts and publishes the nexus and also then will open their vision in the environment using will handle should guys rights but then Elliot shift applying that until then environments run our integration tests have been acceptance that node gets deleted by and off and that change weights for someone to click ladder upon which goes than your normal delivery awesomeness shuffled applied at test run all the tests and stuff deployed abroad passive and then we have our notes until somebody actually it's ready to be released I want big things here that in my 1st pipeline and the last thing the last red states that what happens after prod release than the guys to do their business verification and see this is what we release the correct thing and doesn't where correctly now it's really cool with edge approxiamtely knowledge that releases actually we can do business verification of before because the customers describing your 1st a mom analysis gone way down and wearing vandals but there's nothing really a sentiment Monitor release at the press the button so since you're really awesome and delivery a saved us a lot of a lot of headaches that vision management staff at on con states how much that helps us managing aka protrusions and especially managing appeared possible books and applications and stuff so I mean so that the quality
products we have no and relics some machines there's no passing of flash drives and big knobs goes to deploy artifacts and then that the pretty awesome so now that the working really well and if you remember back but we had that production feedback stuff so we started looking now we're delivering what's actually happening how good are these pipelines we want to know when things are happening in delivery on some things we don't of plot we have we can't use images WebSockets handshake it's broken so we can't actually at those delivery notifications listed e-mail but the terrible to configure each each pipeline individually so right there so
that spam bots service that's as so does very well in the sense that is pretty amazing out it's basically all set up Jason filters basically Jason schema and also when an incoming are those data comes and if it matches the schema do is these actions so it and then we have notifies it also Jason schema-based purple so you construct tokens of the incoming the incoming message formulating adjacent objects and send out on through a bunch of various hand with you know all of such as Archer all I have is a bunch of other things you have to do with them and so we actually also configure nexus security scans on the boxes in sperm of systems all yeah um and then the really cool
thing is and if you haven't seen hide yeah but it's so awesome open-source products by because capital 1 if you haven't seen you should really is really great inspired this tool for us and just because having had here was about my my levels no that so here we can see it now we getting some information what pipelines how healthy other a doing hot frequently honored leaving stuff off the quality of the code we can see when deliveries actually doing stuff in the pipeline so no big box in the top left you can see that connects and the convergence and stuff was the old and then testing the conceptus coverage you can see on a test there are small integration close to another they're actually several went to this screenshot and quality of something that really keen on so the speaking to you to Geoff about artifacts and we have little talk about measuring code quality and how the development of there's a lot of great stuff and so on but 1 of things he was saying is that we need a follow-up to rescue looking for stuff choose high complexity and high touch so it's bit small but that's all follows digital idea API service or something uh we can see that very complex file with conservatives also touched a developers very frequently the problem here is that if you got some complicated control logic in new methods and then you got people having little extra things or extra possible time if it's a very complicated it's can be way easier for them to just find a way to assess the point where my gonna need to differentiate what happens for me than what the Normans so this is a novel if statements or control control for a statement that so and that leads to really and maintainable code and and poor quality so that's something that we can highlight sport and try get us to work on at the bottom we can see all the modifications from the literary and whatever other services are sending notifications fast and but so those actually slack notifications that delivery would be sending to slack just reformatted stripped off and posted here and then I professional really keen on is the thing on the right of maturity unfortunately it's not that clear that's basically the idea is because we have such a governments and and care processes we try really get as much as possible but we need some kind of tool some kind of evidence to give to the change control guys so that we doing all this stuff why do we need all this change control so will it's apart pipelines and then we cannot with a bunch of metrics that we can measure pipeline based on so it's got so continuous integration as a category infrastructure automation environments that testing code quality and resilience and then we've got the scores for each of those things which I go that that automatically from the environment all running a lot running through spam what's on but just reading data from the reprocessed pointed out or whatever and that we call that give the team a school and then overall score would be all good we think apartments of frequency with God produce how much change than and then depending on what the overall scores we can we can remove those hurdles and I only need 1 Dutch and real hopefully if get someone who's for resilience to citizens got proper automotive and testing in data generation and tests and you know all this stuff down through automation those hopefully not can it be any need for change but we'll see we haven't quite got anyone that following the Jr and there a friend
of the that's encirclement FF and any questions
Bit
Spieltheorie
Office-Paket
Verschiebungsoperator
Knotenmenge
Datenmanagement
Mathematisierung
Endogene Variable
Programmierumgebung
Standardabweichung
Prozess <Informatik>
Momentenproblem
Perspektive
Mathematisierung
Prozessfähigkeit <Qualitätsmanagement>
Speicherabzug
Biprodukt
Programmierumgebung
Streuungsdiagramm
Programmierumgebung
Subtraktion
Prozess <Physik>
Kontrollstruktur
Mathematisierung
Programmverifikation
Zahlenbereich
Sprachsynthese
Kartesische Koordinaten
Code
Division
Physikalische Theorie
Softwaretest
Datenmanagement
Code
Lineare Regression
Stichprobenumfang
Softwareentwickler
Verschiebungsoperator
Softwaretest
Soundverarbeitung
Softwareentwickler
Prozess <Informatik>
Matching <Graphentheorie>
Klassische Physik
Ruhmasse
Plot <Graphische Darstellung>
p-Block
Checkliste
Beanspruchung
Grundsätze ordnungsmäßiger Datenverarbeitung
Mereologie
Wort <Informatik>
Programmierumgebung
Rückkopplung
Subtraktion
Bit
Prozess <Physik>
Stab
Mathematisierung
Applet
Versionsverwaltung
Automatische Handlungsplanung
Wellenfront
Kartesische Koordinaten
Smith-Diagramm
Code
Deskriptive Statistik
Einheit <Mathematik>
Prozess <Informatik>
Telebanking
Konfigurationsraum
Maschinelles Sehen
Verschiebungsoperator
Softwaretest
Internetworking
Gebäude <Mathematik>
Ähnlichkeitsgeometrie
Biprodukt
Quick-Sort
Programmfehler
Integral
Mereologie
Projektive Ebene
Programmierumgebung
Bit
Mathematisierung
Versionsverwaltung
Konfigurationsraum
Softwaretest
Soundverarbeitung
Zentrische Streckung
Subtraktion
Verzweigendes Programm
Versionsverwaltung
Division
Maßstab
Spieltheorie
Offene Menge
Prozess <Informatik>
Skript <Programm>
Luftreibung
Programmierumgebung
Maschinelles Sehen
Verschiebungsoperator
Biprodukt
Maschinelles Sehen
Verschiebungsoperator
Softwaretest
Kraftfahrzeugmechatroniker
Selbst organisierendes System
Minimum
Abelsche Kategorie
Content <Internet>
Kartesische Koordinaten
Figurierte Zahl
Programmierumgebung
URL
Schreib-Lese-Kopf
Office-Paket
Fehlermeldung
Internetworking
Eins
Punkt
Rechter Winkel
Bridge <Kommunikationstechnik>
Biprodukt
Verkehrsinformation
Aggregatzustand
Metropolitan area network
Subtraktion
Dienst <Informatik>
Gewichtete Summe
Prozess <Informatik>
Mobiles Internet
Mathematisierung
Grundsätze ordnungsmäßiger Datenverarbeitung
Kartesische Koordinaten
Biprodukt
Regulator <Mathematik>
Verschiebungsoperator
Office-Paket
Soundverarbeitung
Metadaten
Knotenmenge
Informationsmodellierung
Prozess <Physik>
Forcing
Mathematisierung
Versionsverwaltung
Speicher <Informatik>
Systemplattform
Programmierumgebung
Softwaretest
Soundverarbeitung
ATM
Bit
Prozess <Physik>
Dokumentenserver
Versionsverwaltung
Quellcode
Objekt <Kategorie>
Teilmenge
Randwert
Projektive Ebene
Eigentliche Abbildung
Figurierte Zahl
Verkehrsinformation
Schnelltaste
Knotenmenge
Dokumentenserver
Server
Versionsverwaltung
Wort <Informatik>
Whiteboard
Bildgebendes Verfahren
Teilmenge
Softwaretest
Resultante
Rückkopplung
Virtuelle Maschine
Datensatz
Dienst <Informatik>
Einheit <Mathematik>
Bit
Mereologie
Automatische Differentiation
Programmierumgebung
Bit
Gewicht <Mathematik>
Mathematisierung
PASS <Programm>
Code
Virtuelle Maschine
Knotenmenge
Multiplikation
Perspektive
Softwareentwickler
Softwaretest
Güte der Anpassung
Magnetkarte
Einfache Genauigkeit
Systemaufruf
Ruhmasse
Physikalisches System
Quick-Sort
Integral
Teilmenge
Beanspruchung
Dienst <Informatik>
Grundsätze ordnungsmäßiger Datenverarbeitung
Server
Programmierumgebung
Streuungsdiagramm
Resultante
Trennungsaxiom
Softwaretest
Videospiel
Mathematisierung
Versionsverwaltung
Profil <Aerodynamik>
Zahlenbereich
Kartesische Koordinaten
Physikalisches System
Code
Design by Contract
Weg <Topologie>
Datenmanagement
Mereologie
Programmierumgebung
Phasenumwandlung
Verschiebungsoperator
Magnetbandlaufwerk
Proxy Server
Bit
Datenmanagement
Prozess <Informatik>
Mathematisierung
Biprodukt
Systemplattform
Verschiebungsoperator
Rückkopplung
Proxy Server
Bit
Prozess <Physik>
Gewicht <Mathematik>
Desintegration <Mathematik>
Stab
Mathematisierung
Green-Funktion
Kartesische Koordinaten
Code
Virtuelle Maschine
Flash-Speicher
Knotenmenge
Softwaretest
Datenmanagement
Spieltheorie
Code
Proxy Server
Softwareentwickler
Maschinelles Sehen
Bildgebendes Verfahren
Verschiebungsoperator
Analysis
Softwaretest
Zentrische Streckung
Videospiel
Softwareentwickler
Prozess <Informatik>
Programmverifikation
Plot <Graphische Darstellung>
Biprodukt
Quick-Sort
Integral
Teilmenge
Rechter Winkel
Grundsätze ordnungsmäßiger Datenverarbeitung
Wort <Informatik>
Normalvektor
Programmierumgebung
Message-Passing
Aggregatzustand
Maschinenschreiben
Punkt
Prozess <Physik>
Quader
Gruppenoperation
Mathematisierung
Impuls
Abgeschlossene Menge
Sprachsynthese
Mathematische Logik
Komplex <Algebra>
Code
Übergang
Minimum
CMM <Software Engineering>
Softwareentwickler
Konfigurationsraum
Softwaretest
Filter <Stochastik>
Befehl <Informatik>
Linienelement
Kategorie <Mathematik>
Open Source
Computersicherheit
Kontinuierliche Integration
Systemplattform
Token-Ring
Physikalisches System
Elektronische Publikation
Biprodukt
Frequenz
Integral
Softwarewartung
Dienst <Informatik>
Generator <Informatik>
Einheit <Mathematik>
Komplex <Algebra>
COM
Rechter Winkel
Digitalisierer
Grundsätze ordnungsmäßiger Datenverarbeitung
Gamecontroller
Energieerhaltung
Information
Programmierumgebung
Portscanner
Message-Passing

Metadaten

Formale Metadaten

Titel How to Drive a Delivery Truck with Insurance in Africa
Serientitel ChefConf 2016
Autor Breyer-Menke, Stephen
Talken, Marcus
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 10.5446/34634
Herausgeber Confreaks, LLC
Erscheinungsjahr 2016
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract The journey of how Standard Bank implemented continuous delivery in a highly regulated industry using Chef, Chef Delivery and a host of other tools.

Ähnliche Filme

Loading...
Feedback