Merken

Managing the Cloud with a Few Lines of Python

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
the
soul the 2nd speaker of the session is going to be kind to back up fact that has been developing in existence for for more than 10 years he specializes in the field of of testing and benchmarking the since 2013 he is working for Amazon Web Services and I see a couple of Andersen T-shirts in the room so well flight also has had tutorials and talks at the German pi comes about IPython and John In his spare time which by his own accord is very rare he is producing the German-speaking pop cost in part this to date fuck is going to introduce the lipid portal which makes it very you it's very easy to use 8 W S from Python please welcome fight back up with managing the cloud with a few lines
of Python I want thanks for your interest in in this talk I be already lost 4 minutes so I really have to skip through all this stuff here and as I'm already was kind introduce primary
Darius since 2 thousand 13 so that's fun and his mom and as the operating system group which is based on an interesting kind of kilometers south of where we are now and just the looking for 10 and so on in case you're interested in something that's this talk to us we have used the downstairs and but they also have in all over Europe applications development centers where we have a lot of very interesting work and so yeah if you especially like to work at scale and that's probably a good opportunity and working with Python since about original to do for of yes
I this already mentioned may actually put some effort into this focus here
a friend of mine Marcus let's talk earlier about the Sphinx Organization and he already promised that yeah in the next couple of weeks of a some new episodes from the back to to the to talk the idea of the talk like to get and the local little enough conference was talking to a couple of Stephen developers and they said that we get the credits and actually those guys at least and have any idea what to do with them which is set because you know it up asking about for for them to learn to improve and of I think they can use that so we have checked and much less as many of you probably also they didn't fancy to click through ratification Cervantes's and what would not so they I mean how can we automate that and they also some and background so introduce photos mn but I had the feeling like this is something that's also true for others before we actually
talk about the output of 2 other the things the 1st 1
is of no humans like instructions this the cloud kind of the bottom but let's say it is different people have different opinions about remains I defined it for this talk only you know maybe in half an hour have a different opinion on but I mean by that is that you have a dynamical in AWS speak domestic energy resources that can be some like storage can be something like computes retrieval but that can be networking so all you need to content delivery network it's just they're waiting for you but it also could be some random stuff packed filters what we call security groups you can have these these databases as I will show in a messaging systems and the key here is that you can scale up and all those things and ways to actually have to pay more good on the ideas you don't pay anything for this if you don't use that isn't mentioned previously that has to be scriptable because nobody really wants to do that all the hand and in Python I believe this is the perfect candidate to to do so
and if nothing was eventually right now hundreds of to be uploaded but there is a tool for that already actually eligible to comes with commands into that this is so I would recommend the AWS CLI tool which is also written and Tyson and there is a different time for that so In all of all you need to reduced actually fits in a simplicial script and maybe boulders and isn't that the same and you might be much faster using that of so people to actually
that what was that started with this guy which here and all these 2 criteria so that he left the company and the sum of the project is managed by by which
means we make sure that you know the code is up to date but actually also very happy about contributions so I checked and get up last week so we had nearly 400 contributors to the project we had over 6 thousand commits and that is just the tapestry from some and results intent but the name actually comes from this the bottom here of which brings me to the
1st example and show you so or maybe many of you and the structure services called as 3 some the idea is that you can dump the stuff yeah well would be called buckets he also to think of it like a name space of a directory and there you have to create a key that you can have the objects that attaches key you have an object and this can be a stream of whatever so can be fired but as a set whatever and what make sure is that you actually get data back so the term for that is probabilities n what AWS guarantees you is that you get a 99 . 9 9 9 and that's really 9 times 9 per cent of all of but chance that you that you really get a data vector and if you look up what your heart right gives you an indigenous integration some great arrays you receive that's hard to reach resources number and I
they had to do with life than most but I have a couple of the
books prepared to just to see him because he hadn't 107 at the display here that works so but even if you do it in the 1st phase you the portal and you see the most it so execute that done but I have to file on my my heart it doesn't just like and so what we do here as the 1st right to create the
pocket of the details of the just connected but and then we as I mentioned we have to create a key for of the objective you want to upload it and the sky it was called that interpreted and then we just upload to content so so that this is start tangent to so and know that some a cosmetics so we generate actually what we do here is we get this spike again only goes through the list of peace of but really so much 70 the actually this line here that generates you that is valid only for 20 seconds so what you can do is of course you can generate your as but it forever but sometimes you just want to share files and you do not want you know did at this and I'm to so you just 1 have but it's just want have thing being valid for a certain amount of time and that is what the yes so on and on the right in the dust yeah see signature for others what is attaches you and find begun to the fact that I well you know analysis how to create strong was a double time build to and saying that the music industry get in the wrong but actually very helpful for the college in the me this figure prepared that already out you may not remember that I gave you this thing here
down here for this presentation and and this of course also comes from street so the only thing do actually is yeah attached to your reading question of different and when you get is the transfer so was the limitations of the violence we have here Cancer can talk to each other but that would be improved downloads like this and and a next example
I talked about this message
queues the services called scarce but it basically dozens and there are many other implementation of that you just down the message in in a Quran and somewhere else here take it out that's the basic concept very useful and and distributed systems and
and but as I mentioned that many open source project that this kind of the same thing but if you want to have that sustainable if you want to have that in high variability and you will find out it's not a simple anymore to set up an and also can be quite costly if you have to distributed services and stuff so this model it's quite easy to to do
that the and they go to the personal and try to make this data so this time use the as guess module at of all to the greater connection to this is always region so we we go to the European region from AWS they use service we create a q which means between label yeah has 14 the said the time out to come to that in a bit and related to the main executed respond the 2nd 1 and In the next block we actually at the message so the input of the message and we instantiate 1 and we set the body imaging and you and you write that to the to the queue being that created for so now let's assume we have some on a on a totally different system we create a remote Q and we get all the messages we present the message body and we present the q counts so I executed again then it has for this way so they're yet of course we get we get the message imaging in humans and we also get a kind of 0 so motivated and actually this time and we see again I'll make use of many messages we haven't Q and big surprise smells 1 so the idea that is that if for whatever reason you're services was actually dealing with the message and receiving had knowledge of Jason and doing something that fateful whatever reason crashed then you do not actually want to message to be done you want actually service than men and there is a message to deleted and that's actually the last block so you just say it again you get the messages iterate over them and you delete them done so you get for 1 of these message truth and next time it's empty and that was extended to this case and now
let's launched and so on that brings me to the next so as not work on again we use the standard and polities two-module of water so Section do that you also see you which kind of helpful to is that you know you do not have to generate all those extreme stuff and you as quasi that the semi lovely so what is it was sampled loving and said that a lot of attention to the back which both pick up again greater connection now I have to bargain and people to work she tells me that found the conflict and didn't actually there are several ways I can put in your of the deadliest he's and in case some of that I have temporary keys for this presentation so but In case you will for the reasons don't right the so and actually on it takes to to actually run this instance now this month don't designed by I have to say parameter 1st parameter is actually image we want lunch the term that's on the Amazon mission images and that what you get you get if you get a linear system but notes systems what's being installed and as I mentioned earlier of forget part about so the to 1 this is sufficient to make sure the transposon nativists columns in the notes but you also can have right Susan Dey beyond what was not into the so the thing had to notice you that's actually all the thing novel to generate sparse to launch so you will have to a country availability sounds and and and the data but they have a common and when security groups and that the architecture wrote devices of we don't really have to have the time to learn to get into that but this thing now is assumed the thing is launched in read a couple of seconds and and then to the system boots and that is kind of the hand of from us to to you as a customer that when we do not intend to anyone and we don't really know if you're instance variables of 1 out so there's a service for that money to that of course but it's not the fault so what you get back in that case is a so-called reservation ID which you can use actually then later on to see if you're instances this turned to the state running and which means that that's every but also instances the same thing but it's an idea of what is the confinement back and that is what will change it's looking can here inspired
him to most on but when the that in that example you that you get like 4 of those instances and yeah and once I have the the object to this instance I have a couple of methods like I want to have the topic in the name of as you see later on our 2nd terminators instances and check them and at this point flight Mexican set
the as the next example have to introduce a few concepts in the so called the so called Virtual Private Cloud which roughly would say it is land but just in the cloud so already talked about regions of region is really you know sentence at the geographical point and that is divided into a so-called availability sounds which means if you want to have available services he was launched in different availability songs and 1 goes down and you the central to the other 1 will still be up and running this idea behind that and and then you you want to have this interview see want have you on natural in that you do not want to see the traffic from all management or from other customers or whatever and therefore you basically knowledge this private IP addresses you have the segments and their assessment of the availability and so on and if you want to have those instances exposed to the internet you can always attach topic of characters and there was a again what you can go through this Internet greatly and wrote in all the different things like source stuff you can use it of it's important when I wanna
do now so this example actually
shows how to just launched 10 of those posts and instances from inside the is and all the stuff you need to to build a new kind and so that this is and then this is seen as a the functionality of a Mexican and find at the nodes and then you can compile and I'm afraid we do not have the time to really show that because and all that takes a little bit on so far images so the rate of oxygen
want when the product but it is
easy to is that is important at this time you what to have just 1 instance I wanted 10 that's this thing same exclusively which in society we want to have so this history such as a little bit of compute power but you have to give himself never talk about that area and we want to have monitoring this time and he was standing at a couple of instances is used to separate 2 you know as agent to them and so this step by step let's again Dennis's fabric I actually kick off to the compiler and after after in less than 2 minutes postings things over and Mexican should on everything and that I and very quickly
this through the last example but I could say well maybe 10 instances of the too match so I want to have more flexible and let's say you have composed services something the key there is another tool called up the scale of what it
basically is that you have a so-called ons configuration where you define you know against the size of your budget holds which underwent use and all that you need the so called all skating group which defines the availability zones for instance the minimum size of your of your class the maximum size of and yet this knowledge configuration that is all started indicators of you could imagine that knowledge of what instances been launched and that is exactly what you see on the bottom of the slide here was that gets activities yeah yeah that then you have
to have a standing policy of for scaling up and down and you have to kind of triggers off and on for that and you will and that's here and those alarms begin at the threshold for CPU utilization which have to be has figure for a certain amount of time so that's why 60 seconds and this triggers then it actually scales by this formula here but 1 instance in this we go on and on until you reach the data instances
alright and but such down again 3 lines of Python and
through my slides and the for services to see here both of I think he was assistant would use thank you be we don't
think you think you know we get for 1 very weak questions 1 question please and or a view of a given that I may I ask the question again how long would it take you need to set the service running on little from scratch my experience what you have to keep in Italy this account which means some verification that you you can you be so what and from there you get and a key so you have to configure portal that's with just 2 strings into pieces and you have to give it is shown by whatever and then you use 1 of those lines here in the running thank you of that thing is over there
Softwaretest
Datenfeld
Mereologie
Gerade
Gerade
Streuungsdiagramm
Computeranimation
Web Services
Sinusfunktion
Metropolitan area network
Physikalisches System
Zentrische Streckung
Netzbetriebssystem
Gruppenkeim
Vorlesung/Konferenz
Kartesische Koordinaten
Softwareentwickler
Computeranimation
Web Services
Sinusfunktion
Physikalisches System
Vorlesung/Konferenz
Softwareentwickler
Fokalpunkt
Computeranimation
Data Mining
CDN-Netzwerk
Information Retrieval
Filter <Stochastik>
Schlüsselverwaltung
Datenhaltung
Computersicherheit
Datenmodell
Gruppenkeim
Computer
Digitalfilter
Physikalisches System
Computeranimation
Energiedichte
Mooresches Gesetz
Minimum
Datenspeicherung
Speicher <Informatik>
Schlüsselverwaltung
Streuungsdiagramm
Funktion <Mathematik>
Web Services
Gewichtete Summe
Euler-Winkel
Skript <Programm>
Projektive Ebene
Ausgleichsrechnung
Gerade
Computeranimation
Expertensystem
Resultante
Arithmetisches Mittel
Metropolitan area network
Minimum
Projektive Ebene
Reelle Zahl
Code
Computeranimation
Inklusion <Mathematik>
Internetworking
Videospiel
Namensraum
Euler-Winkel
Logarithmus
Zahlenbereich
Vektorraum
Term
Computeranimation
Integral
Objekt <Kategorie>
Metropolitan area network
Streaming <Kommunikationstechnik>
Dienst <Informatik>
Rechter Winkel
Datenstruktur
Verzeichnisdienst
Schlüsselverwaltung
URL
Array <Informatik>
Offene Menge
Elektronische Publikation
Datensichtgerät
Zwei
Einfach zusammenhängender Raum
Mailing-Liste
Elektronische Publikation
Arbeitsplatzcomputer
Elektronische Unterschrift
Speicherbereichsnetzwerk
Computeranimation
Objekt <Kategorie>
Metropolitan area network
Rechter Winkel
Gruppe <Mathematik>
Notebook-Computer
Figurierte Zahl
Große Vereinheitlichung
Phasenumwandlung
Gerade
Gammafunktion
URL
Analysis
Subtraktion
Euler-Winkel
Inverser Limes
Wärmeübergang
Computeranimation
Lesen <Datenverarbeitung>
Warteschlange
Metropolitan area network
Message-Passing
Dienst <Informatik>
Warteschlange
Message-Passing
Computeranimation
Gammafunktion
URL
Einfach zusammenhängender Raum
Bit
Open Source
Physikalisches System
p-Block
Ein-Ausgabe
Modul
Computeranimation
Warteschlange
Metropolitan area network
Trigonometrische Funktion
Message-Passing
Dienst <Informatik>
Informationsmodellierung
Variable
Web Services
Arbeit <Physik>
Betafunktion
Endogene Variable
Client
Projektive Ebene
Warteschlange
Message-Passing
Bildgebendes Verfahren
Punkt
Euler-Winkel
Wasserdampftafel
Gruppenkeim
Oval
Kombinatorische Gruppentheorie
Term
Computeranimation
Metropolitan area network
Variable
Web Services
Radikal <Mathematik>
MIDI <Musikelektronik>
Bildgebendes Verfahren
Gammafunktion
Einfach zusammenhängender Raum
Parametersystem
Booten
Logarithmus
Computersicherheit
Machsches Prinzip
Zwei
Quarkconfinement
Schwach besetzte Matrix
Physikalisches System
Gruppenoperation
Digitale Photographie
Linearisierung
Warteschlange
Objekt <Kategorie>
Menge
Rechter Winkel
Mereologie
Datenerfassung
Garbentheorie
Direkte numerische Simulation
Computerarchitektur
Schlüsselverwaltung
Standardabweichung
Lesen <Datenverarbeitung>
Instantiierung
Aggregatzustand
Kernel <Informatik>
Subtraktion
Punkt
Natürliche Zahl
Quellcode
Netzadresse
Dialekt
Menge
Computeranimation
Internetworking
Metropolitan area network
Dienst <Informatik>
Datenverwaltung
Compiler
Streuungsdiagramm
Instantiierung
Kernel <Informatik>
Metropolitan area network
Lineares Funktional
Knotenmenge
Bit
Compiler
Biprodukt
Bitrate
Menge
Bildgebendes Verfahren
Computeranimation
Instantiierung
Gewöhnliche Differentialgleichung
Kernel <Informatik>
Metropolitan area network
Zentrische Streckung
Mooresches Gesetz
Dienst <Informatik>
Bit
Matching <Graphentheorie>
Flächeninhalt
Schlüsselverwaltung
Computeranimation
Instantiierung
Inklusion <Mathematik>
Zentrische Streckung
Schwellwertverfahren
Ontologie <Wissensverarbeitung>
Extrempunkt
Zwei
Klasse <Mathematik>
Softwarewerkzeug
Gruppenkeim
Zentraleinheit
Zeitzone
Menge
Computeranimation
Ausdruck <Logik>
Rechenschieber
Metropolitan area network
Minimum
Zentrische Streckung
Indexberechnung
Figurierte Zahl
Konfigurationsraum
Instantiierung
Rechenschieber
Metropolitan area network
Dienst <Informatik>
Uniforme Struktur
Zoom
Gateway
Notepad-Computer
Bildschirmsymbol
Benutzerführung
Gerade
Computeranimation
Sichtenkonzept
Web Services
Dämpfung
Code
Programmverifikation
Vorlesung/Konferenz
Gerade
Zeichenkette

Metadaten

Formale Metadaten

Titel Managing the Cloud with a Few Lines of Python
Serientitel EuroPython 2014
Teil 20
Anzahl der Teile 120
Autor Becker, Frank
Lizenz CC-Namensnennung 3.0 Unported:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen 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.
DOI 10.5446/19977
Herausgeber EuroPython
Erscheinungsjahr 2014
Sprache Englisch
Produktionsort Berlin

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Frank - Managing the Cloud with a Few Lines of Python One of the advantages of cloud computing is that resources can be enabled or disabled dynamically. E. g. is an distributed application short on compute power one can easily add more. But who wants to do that by hand? Python is a perfect fit to control the cloud. The talk introduces the package Boto which offers an easy API to manage most of the Amazon Web Services (AWS) as well as a number of command line tools. First some usage examples are shown to introduce the concepts behind Boto. For that a few virtual hosts with different configurations are launched, and the use of the storage service S3 is briefly introduced. Based on that a scalable continuous integration system controlled by Boto is developed to show how easy all the required services can be used from Python. Most of the examples will be demonstrated during the talk. They should be easily adoptable for similar use cases or serve as an starting point for more different ones. ----- One of the advantages of cloud computing is that resources can be enabled or disabled dynamically. E. g. is an distributed application short on compute power one can easily add more. But who wants to do that by hand? Python is a perfect fit to control the cloud. The talk introduces the package Boto which offers an easy API to manage most of the Amazon Web Services (AWS) as well as a number of command line tools. First some usage examples are shown to introduce the concepts behind Boto. For that a few virtual instances with different configurations are launched, and the use of the storage service S3 is briefly introduced. Based on that a scalable continuous integration system controlled by Boto is developed to show how easy all the required services can be used from Python. Most of the examples will be demonstrated during the talk. They should be easily adoptable for similar use cases or serve as an starting point for more different ones.
Schlagwörter EuroPython Conference
EP 2014
EuroPython 2014

Ähnliche Filme

Loading...