Merken

To the Clouds: Why you should deploy to the cloud even if you don't want to

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
that it the could can you hear me 1 but at the cost of so my microphone I will canonical
on vertical due to this reason the government had been applied to develop of about 30 years now and I think or develop about of those years aria released all created unit test 2 and lot of library and which reflects the particular interests my particular passion the test I think protesting practices of the only ways of the world so they are an important part of keeping developed site and this is this is my colleagues and actually my boss to major issue the period of the I and the me the I I work with Michael in canonical form the hopefully next generation of the station framework to system but did do here is that so
it's goes to the class and it's about deploying applications and services to the clouds of white should probably be deployed to the cloud even if you don't want that because we're talking about clouds wouldn't have a lot of buzz words the cloud is there a piece of jargon that have a lot of and over the last decade really I'm hoping that in this talk will reveal the same term a little bit and show how this despite the high is actually useful set of technologies and principles behind so this
was a lot of structure right is actually 1 of the original computers that's Google used for that in the early days of a search engine apology and which I think is actually forced on the mobile out of necessity them directly on entire genius was they the service using cheap commodity hardware rather than the very expensive big on mainframes their competitors we use and a consequence of this was that they were able to scale out very rapidly very easily in very cheaply just by applying new units of inexpensive commodity hardware but because they're so this hardware wasn't as reliable as the the big on mainframes that other people were using that builds on top of it is fault tolerant architecture that serve as well as enabling them to on ads you units of hardware to the system also allows them to take out very hardware they eventually released this Platform as a Service has on the Google App Engine I wasn't undertook this to the next level with their infrastructure that used for running a giant retail websites and they took a slightly different approach providing on a host of our virtual machines as deployment targets for their services and this infrastructure was as a service approach rather than the Platform-as-a-Service as a approach and is really loved by developers because it just gives them a machine and and when they just have a machine they know what to do we don't know how to the point to of this so I when I wasn't made them Cloud public rapidly became the dominant player in the the public cloud markets and all the modern infrastructure as a service public clouds include HP Cloud Microsoft Azure jury and and OpenStack based offers a whole host of the so there
are several problems that are using the cloud sold and these include dependency health resource on the utilization and hardware management on the way the clouds solves these problems is by separating the the deployment targets the you deployment letter from your hardware layer so we deployed to the cloud you deployment of virtual machines without having to care about what physical machines the use of this is actually running on so resource underutilization was also equally was also utilization of the situation you might be familiar with you have to deploy a new on small public service sigh or an e-mail server dev Wikibook track so you need a new machine when you deploy stuff using about 10 % of its past or alternatively you work in the company were getting new hardware is already slow and painful process so what you do is the free service that you already have John everything onto that and everything runs as slow as hell dependency have again situation may be familiar with you have a whole bunch of other applications and services and some of them use the same library that the same dependency for these different versions of the same library so and because you have these applications the can't deployed on the same machine that have to be located on physically different machines or alternatively you do the work you make sure they're all using the same version of the library and then what happens is that you want to use some fancy new feature that comes in the new version of the library so you you deploy that for 1 application forget about the other services and your your deployment breaks on another application not obviously all Python libraries take backwards compatibility very seriously so this never happens in practice but I had to do emergency rollbacks of deployments because we upgraded dependencies we forgot some of the application of the same box using the same dependency and just or alternatively you do that well the new port all your applications at the same time to use the new version of the library and do you do your operation due new releases all in lock step and then there's regression in 1 of your applications and you have to roll back all of them at the same time and there are various ways of solving this you can follow your services to separate machines and then you run into back into resource utilization we use virtual and and to provide an isolated and deployment environments for all of your application again this is something that so that out of the way you then have you shouldn't have been with the same libraries in multiple different locations and non-standard places on the file system and that's a security problem and system administrators they tend to have this solution because they want to understand the and preferably be be able to control dependencies so an alternative approach is for every service that run or even every component of every service that you want to have that in his own the virtual machine where you're about to very tightly controlled and specify the dependencies just for this application and then hardware management that's 1 of the most important benefits of the the clouds because we have this separation of about deployment where for my physical and we're able to deploy new services just by acquiring a new virtual machine were able to add new machines the cluster very easily and were able to take out very machines operate machines without shutting down running services can exploit so they may be the case that
your already running a bunch of services you may have hundreds of services you may have just a few hours and you certainly don't want you feel like you don't want the clout and certainly don't want your data or your customers data on someone else's machine maybe you don't want your data located in America for hosted by an American company and and you get some of the benefits that that our of talked about just 1 managing virtual machines on your answer that you get this the isolated environment and will only be nice if there was a framework that provided the dynamics of management aspects of this that is able to provide automatically provide you with new virtual machines automatically provide you with a new deployment targets and make it easier to our machines and taking data and that's what you really want is applied private and if what you want is a private cloud that problem problem and you
want that a PDF experts screwed up all the sort of thing so OpenStack is it's it's basically the private cloud it's it's not entirely the only option but it's it's the giant and 11 is written in Python i is probably 1 of the biggest things going on about the world right now so there are lots of companies by big and small hiring Python developers to work on the clouds either all directly on OpenStack on specific implementations of of the of Clouds for public and private using instead of other stuff it's huge it's huge in terms of the amount of code it's huge in terms of the amount of projects within OpenStack huge in terms of the functionality it provides an huge in terms of the number of people using it and contributing and you can get all the benefits of clouds but without having to use somebody else's implementation there was about to
the loss of that just mentioned that there are other alternatives eucalyptus and use the alternatives and plot implementation they got acquired on the required something by company that has a public cloud offerings using a that so that thing you know for you which is currently exists and there's no alternative data sensor technology that sort of worked with a bit called Marce metal as a service and that's another canonical product and that gives you a lot of the benefits of the plant that the dynamics of a management aspects of it but with physical hardware rather than virtual machines and due to the project the we worked on it be deployed can deploy American step 2 miles or you can deploy deploy directly to Mosul with due to and it's achieves full so density full resource utilization with 3 using let's orcadian consignments to the deployment targets only the physical machines there's an interesting technology sort of a data center level technology sits below the level of stuff so the benefits of the
clouds of solving the problems that kind of dependency hell hardware management resource utilization if you have all along alongside that if you have automated fully automated deployments than the other big benefits you get from the cloud on you get it right rapidly the ability to rapidly scan about and easily deploy new services and you only get those if you have a fully automated deployments and an important principle for fully automated deployment of the which we've just service as livestock rather than pets so if you have a pet your pet is unique repair has a name 2 name servers and that the company you work for so have names and if you are if you practice if you spend a lot of time and money and effort on getting well again where's lifestyle livestock they don't and have names that have numbers and if you're livestock entail well it's a it's a cruel world to ensure that you get another 1 and this is how we should be treating of servers and deploying new services and we ought to be able to turn down the application servers re provision of with a single command without caring about what physical machine that located on without caring anything about machine that located on virtual machine or physical machine without preferably without having to worry about machine configuration now the trouble is that with infrastructure as a service with which is largely beaten what a service in the market by the way except for some specific platforms unlike sales force and infrastructures services large and the app developers like because the paradigm provides is you have a machine to the to what that means is that although we leave some problems behind the problems of what how do we provision how do we can figure out how to manage those machines and we administer them we just take those problems with this 1 goes about now there are lots of tools out there that will help you with machine provision shut up its sold stock and civil some people use stuff although that's really about interface workplaces not really but some people using it in this way to help with the problem of machine provisional but these tools will require you to think about machine provisions what's going to live on this machine how do I configured minister this machine and as developers or even DevOps we really want to spend their time worrying about machine configuration administration released that's not how I think about the services that are employed and when I deploying as an application or deploy a service that what I think about it is I think it's about the components I have my application server that I have a message you have load balancers have database and these are the these are the components a role related to each other FIL interrelated they'll communicate this is how I think about my application but the actual deployment we tend to with with forced into thinking about our services in different ways with the forces of thinking in terms of units of machines and juju is is a tool that takes a slightly different approach so this
diagram on shows that this is through the going to give you and you don't have to use the URI of course but it's a great way visualizing services this shows deployed and related services deployed with due to the lines that show the relationships between the components of the services and situated takes a different approach it's it's about service orchestration it's a tool that provides a powerful service modeling language but just happens to use virtual machines as a deployment Todd the the basic unit for deployment for services or service components and you is the charge and the child no codifies and the knowledge required to deployed and configured at application this is a very important principle of double develops principle of automated deployment is that the knowledge about how to deploy and configure your service components rather than living in some system administrators had and then that system administrator gets a better job somewhere else and unlike we and said in his talk this morning they have deployed services using binaries and nobody knows how to get final perhaps if they so there is some disaster because presumably they have lots and lots of fun because that they can't reproduce like that but that they don't have a repeatable full deployment by can't automate and reproduce you need to have your your deployment knowledge codified and chance that due the the codified knowledge about how to not just how to deploy the but how it communicates with other services and and these
chance tend to be written so when you use intuition gets to codify the unemployment knowledge imparted but not just not through the chance all and there's a whole host of our common infrastructure components many of which you your already using unsure of whether existing chance out so sometimes called up get so why did you what
kind of talked about some of these we get to think in terms of and service orchestration think in terms of service components and how they relate to connect to each other which is how we think about our applications that and when we get to and work with Python to do cloud independence is important as well juju will happily deploying using the same exactly the same configuration to use your Durant HP Cloud seeds to Mars and also using lexis containers to your local machine on which is not just you let permission to use the ice so you can take exactly the same production deployment configuration you can deploy that locally for running around new test we'll see ISO can also take the that can pronounced spin up the whole production stuff stock and configure their the communication of the biggest things to to talk to each other and you can run your your acceptance tests on on UCI so using effectively your production configurations and you are running CI tests you are running acceptance Europe regulations and like this that I have a particular interest in testing the good principle we're testing this if it is a test is broken but is tested it might not be brought some the principles of origins of testing doesn't guarantee that things were against it what it does is provide ensures that things might work you need to actually and through the chance to
although there are many other jobs
available things like do set PostgreSQL and gender gender applications my sequel to assess squares of all of the standard sometimes there'll charmed up there and ready to deploy so let's look at an example deployments Django we have 10 minutes left on this slide Demetrius initially and its agenda on the pace using the gender framework charm and this is what we already have to put is not this is a lot of false start on a lecture to me to take over so that is so as you can see here well the
resolution is not very good but you have services represented the boxes and this is all you really care about seniors deployment you don't really care that much about machines you care about the what lots and also how they relate to each other and as you
can see the head i and j framework here which is a modified version of the default the channel in the textbook and it's configured run and the based only in the D-tree people of all of them from the that and there are a few other things like that as the proxy and also a couple HA champs which are configured based on application load balancer and cash flow analysis we have from 10 that C and also the backend balls rests on database this is do you who itself which is also a chair that you may or may not be employed in and these here is that the unicorn which is subordinate to our debate what does that mean it means that we each and every unit staff of the base there is unique on deployed alongside so are eat it can be used for example for lobbying for other closely related services which depends on each other and also so this is this is this is the default and here we have the machine you which shows are that we have 8 machines actually lexicon bigger and what services there any longer and if these were physical mission this is all running on a laptop using the local provided that I mentioned and so the lexicon times these services running in which is why I want to tell some of the things we're running on virtual machines that can solve our constraints that we have to change what is I also can connect is that there you can also use the command line to do the same sort of inspection
and management so for example you haven't status months that that's just the gist of what I had in my deployment and if I want to get show a lot more details about it what sort of work all state cities for each channel relations and look at the addresses when it was deployed in so long and there is even more state history that shows for example the right point this charge but it didn't come up or like it eats waiting for some sort of interaction like I need to enter database it that's all I need to do something manually in order to enable it to work also the we can easily go well not quite right now
because the nonexistent but I had shown you will have the same exact employment than in Amazon it's actually
of if you had if you can keep this to say because we but there yeah and I can show you where recent use it to make the uh questions for welcome so what you can do with this you can actually see what are the details of its services and see what's
units surrounding on it's a loss the configuration a lot of things that you can tweet for
each child and this is also practical Chen encapsulates as best practice of 1 of what we can do here is in the soul there was 1 running unit if we needed to scare last or what we would do due to and units think gender whatever whatever this chance go 5 and that would create 5 new new units but because they gender challenge related to the PostgreSQL database server as the as the new units come up there is yeah juju these need to be related to the database and so on chance that comprise the hoax which our code that can run if we're going to get back to the slides will have to fly through the use of just 5 minutes but that there there's a
chance relation joy and so a new within the opening of the new units of the application servers joined the the principles that was so knows the configuration information that they the PostgreSQL of John and what will we don't need to generate a password and user for the full for the act for the database sorry and this is all done as part of the job what went on and so on stuff start up a new new database users created and then as the relationship is joined to the application server that information is passed through it is the configuration is automatically given to the to the application server as we had a new units of the application server that configuration is given information is given to the new units so this the these services are automatically added and configured told through the relationships that we define defined the there automatically configured to talk to the rest of the application so it's not just the the deployment of the information that's codified into the the the the child but the configuration information how to talk to other other units of the service and as we add new units of this reconfiguration is updated configuration happens automatically for this is this is so strong you can see this is in the directory of an of an example John it's just as a part of and we can see here that all of the different under the different hopes that PFA when different things happen that students to a single Python
file and here's installed and that actually act installs a bunch of packages onto the to the next slide so we're actually using
deployable bundle on here which is often a young file you you can keep under version control deployers as a separate so that the functionality is now the world rolled into you call and you can see for the for the gender chart that we're using but there's a whole bunch of configuration information stored in that and I can go
on to the next slide that the relationships between the chance that and the other components of the service that we use and also the defined in 2 . 1 so you can see there's an awful lot to due to personal for what will we talk about the basic principle of codifying I deployment and configuration information in Python and being able to reuse the existing chance using virtual machines deployment targets together so that benefits of the cloud of separating deployment letter from a physical hardware on which you should be aware that just read target that of multiple back and this is a very powerful combination of will
and then this is just an example of how easy it is to do this will will make the slides available as the last if we go to the last slide that hasn't URL out to which individuals we this is
deployed during its stuff 11
that's about all we have time to the end of the study available and there is also as documentation and all the different the concepts to explain to you the of course there's also like them that can be used as a sandbox to do like experimental deployments you can then export that as the yellow file and then deployed as we shall we quickly did you deploy to your now for your local machine and Vermont few minutes symposium the a question it when I was in I see you like is something like a moral thought that a bad salt for any of these things that I don't really see a how it complicates the B because there is something that when you have some that simple applications usually comes from the word them if we 3 parameters so that it's only the applications that it used in my combining we don't a really really companies where we need to have a like for those in beyond all 3 exams it's as it is to that it's in the city and they are the completely differently so at the end of the it's cases you have are really really conflicts in how it would you still with the complexity of this so on there are quite a few very complex challenges that have certainly in the order of tens of configuration options which are things that you cannot in the the the the deployed gamble but obviously there will have same defaults but there are many tens of configuration options that you can tweak and the following on what we tend to see people doing it is for their own applications and they will write their own nature the proposal which I can let the Django framework traumatic that a basic you deploy agenda well and you can have thoughts challenged the constraint of the the application bundle itself and extra configuration information to you merely specify the path to the configuration falls within the charm bundle of and so you have your own that is specific to your application and then he would reuse existing jobs for the other component pieces that are for the 1st of point 1 side and the good thing about terrorists is that we should admit an example and like but they're actually language-agnostic so you can for example if using shared up at the and so you could use the same language you can use whatever you like but do we have a very good support for like like
like schools libraries unit
best support cover region and match so I mean 1 of the things to note is that
under the hood of course preferred individual service component is doing machine provisions and the difference is that the model unless you're not thinking in terms of machine provisional so due to really works alongside pixels and you can see that this particular job has a flat file as uncivil drop on this particular gender framework job identified it's still at the point that we were created with the system was originally created without using uncivil and fabric for to doing in the service installation machine configuration I can see that but you can be useful for provisioning services where the basic building block is like a container or something or like anything that runs on top of the M machine that's already running at stake but is due to the right tool or is it intended to be tool that you can use for actually speeding up all of the ends in the cloud and talked to like Google Cloud API or 8 of Mississippi or any the this this task is that step so well so if you just want something to manage your your your clouds but yeah I mean if you compute engine AWS all of this and you do choose you other machine it will create a new virtual machine for you you can then use that you can specify the machine you can create contains on those virtual machines and deployed components to those change due due SSH machine number gives you a necessary connection to that specific virtual due to the just want something to spin up virtual machines to users alternative what target their then juju actions the lights are on specific demands on those machines so you kind of hands it not like it so highly intelligent so it has the integration with to model API and has with all the cloud providers that supported them Google Compute Engine a diverse student in which love this your job Durant yes it housed within its knowledge of how to talk to present and got out to create virtual machines to destroy the virtual machines which connects some our aspects about the network this kind of stuff here has to be that has to do that because it's during the machine provision for and the question on you mentioned that your intuition is treated like like stock there are have you deploy configuration changes you take on machine in most of you would not want to reapply right so the idea is that you Our goal away from the actual most local figuration instead called what configuration your service needs within the charm and then if you need to change that integration you specify the tweaked charm corporate settings or you related to something else which provides functionality needs so the the actual configuration happens and automatically this depending on how the terms so if for example your time is because Volpe as a stand-alone application and as a clustering solution and if you just have more units the venture will automatically discovered all southern units and set clustering in the same way but there is a concept of charge on moderate where there's a new version of a charter right to the point yet individual units can be upgraded to the next version of the yeah the question yeah where do you follow is getting no nature and that belongs to the ongoing deploying new settles on the demand so this is Canada as a feature propofol due to
currently provides their plans to do things like it was scaling while balancing as part of the the set of features that would provide that's not a what I'm not getting there but it's being worked on it's also as we mentioned that the source work 2nd contribution sits on the fact that we have so but I think the standard answer is landscape who's going on top of the right at the look how is it different then OpenStack's heat run if FIL of each of these different that it's not very being and not quite familiar with both of those in the that our but I think that being this have to do is not a prescribing as a suitable service called that cannot restrict from for example I know he'd like industry told that whereas if you do you can orchestrate all back and other notes just as well with the same set of features and instead published that and these just a quick question how popular is due due respect to you know other kind of similar to see Section 4 possible that situation where it's fully open-source developed by canonical and waging it's a greater without cost this is how and why you exists we have planned customers with large deployments and often but not only large deployments of OpenStack where we set up for the more manageable manage the deployment for them and so this is how we use in our Institute USA score battle-tested getting refined based on those experiences and there is community uses particularly community around the generation don't think in terms of the wider community it's as widely used as just above I think the death what kind of my that is still used to thinking in terms of machine provisions of that look for tools to do machine provision and the idea of social illustration what you do you around for what service orchestration and so as a concept and way of thinking I think it matches more and more directly the way developers think about their application but it's something that something so gaining traction but yeah because we say the community of charm and pressures there's there's lots of people using to and and things recently like last year we had a lot of these contributions like we will be working with our partners called which actually all that called need for Windows and the same cost as well as I think there is a change in their associated with this this is main company who were using due to their their customer needed Windows workflows so they
do the work you do now points Windows workloads and usually once falling on sentences was going to be for Windows you need and want to jujitsu on state so that's manage the applications we deployed to Windows versions of a sentence you can have from samples Jugendstil him on and that's on the server-side and client-side acquired the set available for Windows Linux and Mac and and pretty much any platform that go around so it's it's it's go the worst been here in this and
Softwaretest
Web Site
Komponententest
Jordan-Normalform
Fächer <Mathematik>
Physikalisches System
Frequenz
Framework <Informatik>
Computeranimation
Vorhersagbarkeit
Generator <Informatik>
Benutzerschnittstellenverwaltungssystem
Arbeitsplatzcomputer
Mereologie
Programmbibliothek
Vorlesung/Konferenz
Web Site
Punkt
Klasse <Mathematik>
Green-Funktion
Kartesische Koordinaten
Computerunterstütztes Verfahren
Term
Systemplattform
Computeranimation
Übergang
Fehlertoleranz
Virtuelle Maschine
Web Services
Einheit <Mathematik>
Datenverarbeitungssystem
Suchmaschine
Google App Engine
Datenstruktur
Softwareentwickler
Hardware
Automatische Differentiation
Physikalisches System
Ausgleichsrechnung
Großrechner
Modallogik
Menge
Rechter Winkel
Hypercube
Server
Wort <Informatik>
Computerarchitektur
Personal Area Network
Streuungsdiagramm
Subtraktion
Prozess <Physik>
Quader
Zurücksetzung <Transaktion>
Versionsverwaltung
Kartesische Koordinaten
Framework <Informatik>
Computeranimation
Virtuelle Maschine
Weg <Topologie>
Multiplikation
Datenmanagement
Web Services
Datenverarbeitungssystem
Lineare Regression
Programmbibliothek
Dateiverwaltung
Kontrollstruktur
Zusammenhängender Graph
E-Mail
Trennungsaxiom
Nichtlinearer Operator
Hardware
Diskretes System
Computersicherheit
Softwarewerkzeug
Systemverwaltung
Server
URL
Programmierumgebung
Streuungsdiagramm
Umwandlungsenthalpie
Expertensystem
Lineares Funktional
Softwareentwickler
Besprechung/Interview
Red Hat
Zahlenbereich
Implementierung
Dichte <Stochastik>
Term
Code
Quick-Sort
Computeranimation
Konfiguration <Informatik>
Projektive Ebene
Softwareentwickler
Streuungsdiagramm
Subtraktion
Bit
Einfügungsdämpfung
Physikalismus
Besprechung/Interview
Implementierung
Zahlenbereich
Kartesische Koordinaten
Term
Systemplattform
Computeranimation
Lastteilung
Übergang
Rechenzentrum
Virtuelle Maschine
Einheit <Mathematik>
Datenmanagement
Web Services
Programmierparadigma
Äußere Algebra eines Moduls
Zusammenhängender Graph
Softwareentwickler
Konfigurationsraum
Schnittstelle
Umwandlungsenthalpie
App <Programm>
Softwareentwickler
Hardware
Diskretes System
Datenhaltung
Softwarewerkzeug
Systemverwaltung
Red Hat
Ausnahmebehandlung
Plot <Graphische Darstellung>
Quick-Sort
Dichte <Physik>
Forcing
Server
Projektive Ebene
Streuungsdiagramm
Message-Passing
Systemverwaltung
Prozessautomation
Kartesische Koordinaten
Binärcode
Entwurfssprache
Computeranimation
Metropolitan area network
Virtuelle Maschine
Diagramm
Web Services
Einheit <Mathematik>
Prozess <Informatik>
Server
Zusammenhängender Graph
Gerade
Leistung <Physik>
Softwaretest
Telekommunikation
Stochastische Abhängigkeit
Güte der Anpassung
Kartesische Koordinaten
Konfigurator <Softwaresystem>
Biprodukt
Term
Computeranimation
Teilmenge
Virtuelle Maschine
Web Services
Prozess <Informatik>
Zusammenhängender Graph
Konfigurationsraum
Stochastische Abhängigkeit
Regulator <Mathematik>
Web Services
Quader
Fortsetzung <Mathematik>
Kartesische Koordinaten
Framework <Informatik>
Computeranimation
Rechenschieber
Metropolitan area network
Virtuelle Maschine
Task
Quadratzahl
Menge
Geschlecht <Mathematik>
Server
Benutzerführung
Stochastische Abhängigkeit
Standardabweichung
Bildauflösung
Chipkarte
Nebenbedingung
Proxy Server
Punkt
Stab
Adressraum
Versionsverwaltung
Interaktives Fernsehen
Kartesische Koordinaten
Extrempunkt
Framework <Informatik>
Computeranimation
Lastteilung
Metropolitan area network
Virtuelle Maschine
Task
Einheit <Mathematik>
Datenmanagement
Notebook-Computer
Front-End <Software>
Zoom
Default
Große Vereinheitlichung
Hardware
Schreib-Lese-Kopf
Analysis
Caching
Inklusion <Mathematik>
Web Services
Datenhaltung
Relativitätstheorie
Datenfluss
Quick-Sort
Server
Bildschirmsymbol
Ordnung <Mathematik>
Aggregatzustand
Web Services
Data Encryption Standard
Metropolitan area network
Task
Server
Extrempunkt
Web Services
Schnittstelle
Einfügungsdämpfung
Elektronische Publikation
Datenhaltung
Menge
Code
Hoax
Rechenschieber
Metropolitan area network
Einheit <Mathematik>
Task
Geschlecht <Mathematik>
Server
Polstelle
Konfigurationsraum
Innerer Punkt
Gammafunktion
Caching
Subtraktion
Datenhaltung
Relativitätstheorie
t-Test
Kartesische Koordinaten
Elektronische Publikation
Dateiformat
Computeranimation
Rechenschieber
Einheit <Mathematik>
Funktion <Mathematik>
Prozess <Informatik>
Offene Menge
Mereologie
Server
Installation <Informatik>
Passwort
Binäre Relation
Information
Verzeichnisdienst
Konfigurationsraum
Caching
Web Services
Lineares Funktional
Elektronische Publikation
Hardware
Versionsverwaltung
Elektronische Publikation
Computeranimation
Rechenschieber
Virtuelle Maschine
Multiplikation
Web Services
Geschlecht <Mathematik>
Proxy Server
Zusammenhängender Graph
Information
Konfigurationsraum
Faserbündel
Streuungsdiagramm
Nebenbedingung
Server
Punkt
Jensen-Maß
Wort <Informatik>
Natürliche Zahl
Formale Sprache
Versionsverwaltung
Kartesische Koordinaten
Aggregatzustand
Komplex <Algebra>
Framework <Informatik>
Computeranimation
Metropolitan area network
Virtuelle Maschine
Interaktives Fernsehen
Prozess <Informatik>
DoS-Attacke
Statistische Analyse
Zusammenhängender Graph
Addition
Konfigurationsraum
Default
Gammafunktion
Caching
Web Services
Parametersystem
Elektronische Publikation
Zehn
Singularität <Mathematik>
Güte der Anpassung
Stellenring
Web Site
Programmierumgebung
Elektronische Publikation
Dateiformat
Konfiguration <Informatik>
Rechenschieber
Funktion <Mathematik>
Bootstrap-Aggregation
Wort <Informatik>
Binäre Relation
Information
URL
Bridge <Kommunikationstechnik>
Ordnung <Mathematik>
Faserbündel
Subtraktion
Punkt
Euler-Winkel
Gruppenoperation
Mathematisierung
t-Test
Versionsverwaltung
Zahlenbereich
Term
Framework <Informatik>
Computeranimation
Überlagerung <Mathematik>
Task
Metropolitan area network
Virtuelle Maschine
Web Services
Einheit <Mathematik>
Prozess <Informatik>
Programmbibliothek
Zusammenhängender Graph
Cluster <Rechnernetz>
Tropfen
Figurierte Zahl
Konfigurationsraum
Caching
Web Services
Einfach zusammenhängender Raum
Lineares Funktional
Pixel
Datennetz
Matching <Graphentheorie>
Krümmung
Gebäude <Mathematik>
p-Block
Physikalisches System
Elektronische Publikation
Dateiformat
Cloud Computing
Integral
Diskrete-Elemente-Methode
Funktion <Mathematik>
Menge
Rechter Winkel
Geschlecht <Mathematik>
Server
Reelle Zahl
Binäre Relation
Streuungsdiagramm
Caching
Mathematisierung
Versionsverwaltung
Automatische Handlungsplanung
Kartesische Koordinaten
Quellcode
Systemplattform
Term
Computeranimation
Beanspruchung
Virtuelle Maschine
Generator <Informatik>
Druckverlauf
Web Services
Menge
Rechter Winkel
Stichprobenumfang
Mereologie
Bildschirmfenster
Vorlesung/Konferenz
Garbentheorie
Binäre Relation
Softwareentwickler
Normalspannung
Aggregatzustand
Vorlesung/Konferenz

Metadaten

Formale Metadaten

Titel To the Clouds: Why you should deploy to the cloud even if you don't want to
Serientitel EuroPython 2015
Teil 33
Anzahl der Teile 173
Autor Foord, Michael
Lizenz CC-Namensnennung - keine kommerzielle Nutzung - 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 10.5446/20165
Herausgeber EuroPython
Erscheinungsjahr 2015
Sprache Englisch
Produktionsort Bilbao, Euskadi, Spain

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Michael Foord - To the Clouds: Why you should deploy to the cloud even if you don't want to Do you deploy your Python services to Amazon EC2, or to Openstack, or even to HP cloud, joyent or Azure? Do you want to - without being tied into any one of them? What about local full stack deployments with lxc or kvm containers? Even if you're convinced you don't need "the cloud" because you manage your own servers, amazing technologies like Private clouds and MaaS, for dynamic server management on bare metal, may change your mind. Fed up with the cloud hype? Let us rehabilitate the buzzword! (A bit anyway.) A fully automated cloud deployment system is essential for rapid scaling, but it's also invaluable for full stack testing on continuous integration systems. Even better, your service deployment and infrastructure can be managed with Python code? (Devops distilled) Treat your servers as cattle not as pets, for service oriented repeatable deployments on your choice of back-end. Learn how service orchestration is a powerful new approach to deployment management, and do it with Python! If any of this sounds interesting then Juju maybe for you! In this talk we'll see a demo deployment for a Django application and related infrastructure. We'll be looking at the key benefits of cloud deployments and how service orchestration is different from the "machine provisioning" approach of most existing cloud deployment solutions.
Schlagwörter EuroPython Conference
EP 2015
EuroPython 2015

Ähnliche Filme

Loading...