Ranker jms implementation
- 2. Agenda ● Why Event Processing System For Ranker? ● What are challenges? ● General Overview of JMS ● Walk through ActiveMQ ● Why Messaging? ● Other JMS Providers.
- 3. Why Messaging? ● Asynchronous operation − A client can schedule work to be done and return immediately − A client can be notified in an event-driven fashion when something has happened ● Loose Coupling − Systems connected via messaging need not be aware of each other, or use the same technologies − Key messages can easily be routed to many systems − Message broker is buffer in case of downtime
- 4. Why Messaging? ● Fire and Forget − Message Broker can guarantee messages are recorded and delivered even vs. crashes ● Parallelize Work − Messages in a single stream can be handled in parallel by many client threads or systems ● Throttle Work − A large amount of work can be split into a small number of threads or systems to throttle effort (e.g. due to licensing restrictions)
- 5. Why Event Processing System For Ranker? ● Ranker has a lot of complicated algorithms running behind and user really does not need to wait till the execution of all the algorithms. ● You may wish to implement this because the request take a long time to complete or because several parties may be interested in the actual request ● Eg: Crowd ranked/Vote ranked List aggregation (This algorithm may take 1- 5 mts to complete depends of user contribution.) ● Email sending functionality. ● Moderation of user input (eg. Bad word filtering) ● List Auto tagging.
- 6. What are the challenges? ● Selecting the JMS provider − We are running the app based on Spring 3.0 /Apache tomcat stack, we have to select the appropriate JMS provider. We selected Apache ActiveMQ. ● Reasons for selecting ActiveMQ − Its pretty easy to integrate with Spring Framework. − Open source project. − Pretty good documentation available for ActiveMQ as well as for Integration with spring framework.
- 7. What is JMS? ● A specification that describes a common way for Java programs to create, send, receive and read distributed enterprise messages ● Loosely coupled communication ● Asynchronous messaging ● Reliable delivery - A message is guaranteed to be delivered once and only once. ● Outside the specification - Security services - Management services
- 8. A JMS Application ● JMS Clients - Java programs that send/receive messages ● Messages - Administered Objects preconfigured JMS objects created by an admin for the use of clients - ConnectionFactory, Destination (queue or topic) ● JMS Provider - messaging system that implements JMS and administrative functionality
- 10. JMS Messaging Domains ● Point-to-Point (PTP) ○ Built around the concept of message queues ○ Each message has only one consumer ● Publish-Subscribe systems ○ Uses a “topic” to send and receive messages ○ Each message has multiple consumers
- 13. Message Consumptions ● Synchronously ○ A subscriber or a receiver explicitly fetches the message from the destination by calling the receive method. ○ The receive method can block until a message arrives or can time out if a message does not arrive within a specified time limit. ● Asynchronously ○ A client can register a message listener with a consumer. ○ Whenever a message arrives at the destination, the JMS provider delivers the message by calling the listener's onMessage() method.
- 14. Page 1
- 16. More JMS Features ● Durable subscription ● By default a subscriber gets only messages published on a topic while a subscriber is alive ● Durable subscription retains messages until a they are received by a subscriber or expire
- 17. More JMS Features Difference Between Persistent & Nonpersistent deliveries − Persistent delivery take cares of reprocessing the request if the JMS broker restarted before serving the request, it will use file/db to store the request, this implementation is costly compare to non persistent delivery. − Nonpersistent delivery will just ignore the request if broker restarted while serving the request.
- 18. About ActiveMQ ● An open-source message broker (compare to JBossMQ, or many commercial products) − See http://activemq.apache.org/ ● Generally stable and high-performance ● Can be run standalone, or inside another process, app server, or Java EE application ● Supports everything JMS requires, plus various extensions ● Integrates well into other products
- 19. ActiveMQ Message Extensions ● Virtual Destinations (load-balancing and failover for topics) ● Retroactive Subscriptions (subscriber can receive some number of previous messages on connect) ● Exclusive Consumers & Message Groups (loadbalancing and failover while preserving message ordering) ● Mirrored queues (monitor queue messages)
- 20. ActiveMQ Client Connectivity ● Dictated by the wire protocol a client uses to talk to the broker ● Generally there are two protocol options – OpenWire (binary) and Stomp (text) − OpenWire is the default and has the most history and best support (including SSL) – for Java, .NET, etc. − Stomp is easiest to develop for and therefore has the most cross-language support (Perl, Python, Ruby, ...) ● Also a variety of other special-purpose protocols (Jabber, adapters for REST/AJAX, etc.)
- 21. ActiveMQ Persistence Options ● Different strategies available for storing persistent messages − to local files, database, etc. − or both – stored to local files and then periodically batch undelivered messages to the DB... ● Default implementation changed between ActiveMQ 4.x and 5.x ● May still customize the persistence engine based on specific performance requirements
- 22. ActiveMQ SecurityAnd Management ● OpenWire protocol can use SSL for encryption ● Broker can use authentication (e.g. username/password required to connect) − Uses JAAS to identify the back-end user data store (properties files, DB, LDAP, etc.) ● JMX management enabled by default − Use a tool like JConsole to monitor queues, etc. ● Web Console available as well
- 23. ActiveMQ Testing ● ActiveMQ can easily run in an embedded, nonpersistent, in-VM only mode for unit tests ● Also easily to run ActiveMQ via beans in a Spring context, if you're testing with Spring ● ActiveMQ includes a simple JNDI provider if you want to test Java EE code that relies on JNDI lookups to access JMS resources ● Can use tools like JMeter to load test the broker
- 24. JMS Providers ● SunONE Message Queue (SUN) ○ A JMS provider integrated with the SunONE Application Server ○ http://www.sun.com ● MQ JMS (IBM) ○ MQSeries is another messaging technology ○ Can configure MQ as a JMS provider ○ (http://www7b.software.ibm. com/wsdd/library/techtip/0112_cox.html)
- 25. JMS Providers ● WebLogic JMS (BEA) ○ Enterprise-class messaging system integrated into WebLogic Server ○ http://dev2dev.bea.com/technologies/jms/index.jsp ● JMSCourier (Codemesh) ○ Merging C++ applications into a JMS environment ○ http://www.codemesh. com/en/AlignTechnologyCaseStudy.html
- 26. More JMS Vendors ● Fiorano Software http://www.fiorano.com ● JRUN Server http://www.allaire.com ● GemStone http://www.gemstone.com ● Nirvana http://www.pcbsys.com ● Oracle http://www.oracle.com ● A more exhaustive listing is available at ● http://java.sun.com/products/jms/vendors.html