Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Google App Engine for Java


Published on

Overview of the Google App Engine for Java.

Published in: Technology

Google App Engine for Java

  1. 1. Google App Engine for Java <ul><li>Lars Vogel </li></ul><ul><li>http:/// </li></ul><ul><li>Twitter: </li></ul>
  2. 2. Why do I qualify? <ul><li>Two popular tutorials in the internet about Google App Engine </li></ul><ul><ul><li> </li></ul></ul><ul><ul><li> </li></ul></ul><ul><li>Published an article about robots on the App Engine in the (german) Eclipse Magazin </li></ul><ul><li>Technical editor for a Google App Engine Book written by Google employees </li></ul>
  3. 3. What is cloud computing for me? What is the Google App Engine? Development for GAE with Java Persistence without SQL Tools - A fool with a tool is still a fool App Engine Services Testing on the cloud And of course coding
  4. 4. A typical system setup OS +Application Server + DB
  5. 5. You need scaling... Application Server Database Application Server Application Server
  6. 6. So you are fine, right?...
  7. 7. Designed for peek consumption Time Utilization
  8. 8. Cloud computing tries to solve this
  9. 9. What is a cloud computing not? Putting one or more computer in the internet is not necessary cloud computing That is just a server in the internet
  10. 10. What is a cloud computing? Cloud computing is Web-based processing, whereby shared resources, software, and information are provided to computers and other devices (such as smartphones) on demand over the Internet. Blabla, this means nothing...
  11. 11. Cloud: ....some kind of abstracting from the hardware and providing resources on demand Time
  12. 12. What types of cloud computing do we have? Infrastructure as a Service -> Amazon Platform as a Service -> Google App Engine Software as a service ->, MyERP
  13. 13. Google App Engine Scalable Infrastructure GAE allows you to host webapplications on the Google infrastructure.
  14. 14. Difference to Amazons Scalable Infrastructure Amazon provides virtual servers App Engine provides an interface to program against App Engine give no access to the underlying system / hardware
  15. 15. Run your app on the Google Infrastructure Scalable Infrastructure
  16. 16. Google handles the infrastructure, e.g. hardware failures, security patches, OS upgrades
  17. 17. Sometimes there are issues
  18. 18. and the usage of App Engine is free…. within limits in any case you only pay for what your use
  19. 19. Google App Engine - Limits Scalable Infrastructure 10 Applications per User 5 Million Pageviews are free per month. Approx. 6.5 hours of CPU and 1 Gigabyte of inbound and outbound traffic. 100 hits per secs (non-billing) and 500 for billing enabled applications
  20. 20. Real Costs Example - Game <ul><li>… the site was built on Google App Engine and, according to Google Analytics, it served about 3.4 million pageviews (1 million unique visitors) in the 12 days it was live. It grossed over $1.2 million…. Google sent us a bill for a grand total of $71.56. </li></ul><ul><li> </li></ul>
  21. 21. How does Google run thousands of thousands Java apps at the same time? They don’t… applications which are not actively used will be frozen and saved to big table  Initial startup time
  22. 22. App Engine Architecture App Engine FrontEnd App Engine FrontEnd App Engine FrontEnd AppServer AppServer AppServer Load Balancer
  23. 23. So it scales but what about Platform as a service?
  24. 24. Writing Java Webs application is hard from Guillaume Laforge and Patrick Chanezon
  25. 25. To write a Java Web application <ul><li>Install / prepare the server </li></ul><ul><li>Install / Configure the database </li></ul><ul><li>Setup the webcontainer </li></ul><ul><li>Develop your application </li></ul><ul><li>Package your application into a WAR file </li></ul><ul><li>Deploy it on your server </li></ul>
  26. 26. ...developing a Java Web application from scratch is really, really slow due to the initial setup required
  27. 27. Google App Engine – Programming Languages Python Java-isch Scala Groovy JRuby JPython … Still some issues with Grails....
  28. 28. Servlets and JSPs A servlet is a Java class which answers a HTTP request within a web container. JavaServer Pages (JSP) are files which contains HTML and Java code. The web container compiles the JSP into a servlet at the first time of accessing this JSP
  29. 29. Possible Web Frameworks on GAE Basically all Java Web frameworks, e.g. JSP, Servlets based GWT, JSF, Struts, Wicket
  30. 30. App Engine can be more then just a web application platform Scalable Infrastructure Doesn‘t have to be a webapplication, can be a backend for Android or iPhones... Can be used to do some serious number crunching
  31. 31. Access <ul><li>Can you from a Google domain or from your own </li></ul><ul><li>You need only a Google User to get started </li></ul>
  32. 32. Develop according to Java standards from Guillaume Laforge and Patrick Chanezon
  33. 33. Session Support Turned off by default appengine-web.xml <sessions-enabled>true</sessions-enabled> Persistent and distributed
  34. 34. Configuration <ul><li>web.xml </li></ul><ul><li>appengine-web.xml </li></ul><ul><li>- allows several versions of your app </li></ul><ul><li>can define static content -> super fast </li></ul><ul><li>can define system variables </li></ul><ul><li>enable ssl and http sessions </li></ul>
  35. 35. Development Tools
  36. 36. DevApp Server Scalable Infrastructure Emulates the Google App Engine, its service and restrictions Based on Jetty
  37. 37. Eclipse Scalable Infrastructure
  38. 38. Demo – A new project
  39. 39. Deployment Scalable Infrastructure Run your application on [email_address] or on your own domain Command line or Eclipse based
  40. 40. Logging <ul><li>java.util.logging.Logger </li></ul><ul><li>System.out and System.err are also logged </li></ul>
  41. 41. Performance
  42. 42. Be fast <ul><li>You have to be fast: </li></ul><ul><li>30 seconds to respond! </li></ul><ul><li>Otherwise </li></ul><ul><li>Google helps you that as its optimize its infrastructure all the time </li></ul>
  43. 43. Example for recent performance work Byte code used to enhanced dynamically, now App Engine does this this statically. JIT and GC improvements Example Work on reflection Data from Google I/O 2010
  44. 44. Limits
  45. 45. Limits <ul><li>No native threads – but Tasks </li></ul><ul><li>No sockets </li></ul><ul><li>No write access to file system </li></ul><ul><li>No Native code </li></ul><ul><li>Not all standard Java classes available (white list) </li></ul><ul><li>Processes must finish without 30 secs (Tasks Chains…) </li></ul>
  46. 46. Compartibility is increasing Whitelist is growing. More and more libraries are supported
  47. 47. Datastore
  48. 48. Storing data <ul><li>Application can store data in memory, memcache, blobstore or the datastore </li></ul>
  49. 49. Datastore <ul><li>The datastore is App Engine's non-relational database </li></ul><ul><li>Uses internally Bigtable which is a key-value store </li></ul><ul><li>Read / Write Consistent and distributed </li></ul>
  50. 50. Datastore <ul><li>Build for fast reading </li></ul><ul><li>Update / Insert not that fast </li></ul><ul><li>Individual datastore query can return a maximum of 1000 results </li></ul><ul><li>Support concurrent queries </li></ul>
  51. 51. Datastore <ul><li>The performance of a query depends only on the size of your result set. It does not depend on the total number of data entries </li></ul>
  52. 52. Data Store <ul><li>Based on BigTable </li></ul><ul><li>Sorted Map, no Joins </li></ul><ul><li>Schemaless </li></ul><ul><li>Transactional </li></ul><ul><li>Low-level APIs </li></ul><ul><li>JDO and JPA </li></ul><ul><li>Blobstore </li></ul>
  53. 53. Datastore via JDO <ul><li>Data Classes </li></ul><ul><li>@PersistenceCapable public class Employee {     @Persistent     private String name } </li></ul><ul><li>PersistenceManager </li></ul><ul><li>PersistenceManager pm = PMF.get().getPersistenceManager(); pm.makePersistent(employee); pm.close; </li></ul><ul><li>Retrieval and Queries </li></ul><ul><li>employee = pm.getObjectById(Employee.class, key); Query query = pm.newQuery(Employee.class, “name = :name”); List<Employee> results = (List<Employee>) query.execute(&quot;Smith&quot;); </li></ul><ul><li>Relations </li></ul><ul><li>@Persistent(mappedBy = &quot;employee“) private List<ContactInfo> contactInfoSets; </li></ul>
  54. 54. Indexes <ul><li>App Engine builds indexes for several simple queries by default. </li></ul><ul><li>If you run a query that filters on multiple entity properties or orders results by multiple properties, you will need an index for that query.  </li></ul><ul><li>datastore-indexes.xml in WEB-INF/  </li></ul>
  55. 55. Indexes <ul><li>Every query you run in the SDK automatically gets indexed. </li></ul><ul><li>Stored in application's datastore-indexes.xml (Java) </li></ul>
  56. 56. Other storage possibilities Blobstore Memcache
  57. 57. Memcache <ul><li>Cache data to avoid re-doing expensive operations </li></ul><ul><li>Fast </li></ul><ul><li>Developer can define a time-limit but App Engine may remove the data earlier -> do not rely on the availability </li></ul><ul><li>Cache Values can be up to 1 MB </li></ul>
  58. 58. Blogstore <ul><li>Store, Upload and serve large files (<=2GB per blob) </li></ul><ul><li>Create by upload a file via HTTP POST </li></ul><ul><li>Read only </li></ul><ul><li>Application can read blobs as if there are local files </li></ul><ul><li>Requires billing enabled </li></ul>
  59. 59. App Engine Services
  60. 60. <ul><li>Caching </li></ul><ul><li>Channel API (push to client, comet) </li></ul><ul><li>URL Fetching </li></ul><ul><li>Mail </li></ul><ul><li>Instant Messaging (XMPP) </li></ul><ul><li>Image Manipulation </li></ul><ul><li>User Management </li></ul><ul><li>Background Tasks </li></ul><ul><li>Map (no Reduce) </li></ul><ul><li>XMPP (Chat) </li></ul>API’s
  61. 61. <ul><li>Lets apps issue HTTP and HTTPS requests and receive responses </li></ul><ul><li>Really fast (uses Google Infrastructure) </li></ul><ul><li> HTTP API </li></ul><ul><li>Port 80 and 443 supported </li></ul><ul><li>1 MB limited for request and 32 for response </li></ul><ul><li>Supports both synchronous requests and asynchronous requests </li></ul>URL Fetch
  62. 62. <ul><li>Resize, rotate, flip…. </li></ul><ul><li>Compose image from several images </li></ul><ul><li>Image data from the app or from blobstore </li></ul>Image Service
  63. 63. <ul><li>Background processing by putting tasks into a queue </li></ul><ul><li>GAE dispatches the tasks automatically </li></ul><ul><li>Gets data and a URL as parameter, the URL represents the code which is executed </li></ul>Task Queue
  64. 64. Demos Lets write a chat bot for the App Engine using XMPP - Extensible Messaging and Presence Protocol
  65. 65. Demos Lets send out emails...
  66. 66. Demos Receive email
  67. 67. Testing...
  68. 68. Testing Local Testing - equal functionality as the cloud - not equal in scale - LocalServiceTestHelper(….services to be tested) Cloud Testing - serving HTTP requests - scalable and fast - all limits apply which apply for standard application -
  69. 69. Local Testing Similar to standard Java testing - JUnit, TestNG, Selenium Difference: Some classes require a certain App Engine setup, e.g. the datastore AppEngine Testing API
  70. 70. API.Proxy Register your own class as proxy for the GAE API calls and interfere them.
  71. 71. GAE 1.4
  72. 72. App Engine Recent Updates Running JVM (3 instances) WarmupRequest via appengine-web.xml The Channel API is now available for all users. Task Queue released, Deadline for Task Queue and Cron requests has been raised to 10 minutes. Added a low-level AsyncDatastoreService for making calls to the datastore   asynchronously. The whitelist has been updated to include all classes from javax.xml.soap. Source:
  73. 73. Google Cloud for businesses 99.9 % SLA pay 8 Euros per user up to a maximum of 1000 Euro (==125 Users) Hosted SQL SSL (Secure Sockets Layer) for the domain Multi-tendency
  74. 74. Leverage the greatest IT infrastructure of the world Fast development of new functions Just to great to ignore Google App Engine
  75. 75. Photo credits Please add to the number if not specified <ul><li>Cloud 1309726 </li></ul><ul><li>Agenda 187747 </li></ul><ul><li>Guy pulling the white wall </li></ul><ul><li>Balloon </li></ul><ul><li>Critical guy </li></ul><ul><li>Question mark in the box 1084632 and 1084633 </li></ul><ul><li>Hammer 604247 </li></ul><ul><li>Server 175983 </li></ul><ul><li>Turtle 1198861 </li></ul><ul><li>Thinking Guy 130484 </li></ul><ul><li>Picked Fence 695054 </li></ul><ul><li>Performance 765733 </li></ul><ul><li>Tools 1197009 </li></ul><ul><li>Open Door </li></ul><ul><li>Paper Chart </li></ul><ul><li>Binary </li></ul><ul><li>Footprint </li></ul><ul><li>Old Computer </li></ul><ul><li>Carton </li></ul><ul><li>Eye </li></ul><ul><li>Guitar playing man 894247 </li></ul><ul><li>Brown bag 250762 </li></ul><ul><li>Future 1139530 </li></ul><ul><li>Guy reading book 406547 </li></ul><ul><li>Money House 1097251 </li></ul>
  76. 76. Thank you <ul><li>For further questions: </li></ul><ul><li>[email_address] </li></ul><ul><li> </li></ul><ul><li> </li></ul>
  77. 77. Hands-on Tutorial The following tutorials give an intro to GAE/J using servlets and JSP‘s
  78. 78. License & Acknowledgements <ul><li>This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Germany License </li></ul><ul><ul><li>See </li></ul></ul>