Anda di halaman 1dari 43

Global Scale ESB with Mule

Andrew D Kennedy
grkvlt@apache.org
March 2011
Who Am I
• Enterprise Java Developer
• Apache Qpid Committer
• Previously Security Engineer
• Over 15 Years Java Experience
• Worked with various Investment Banks
• Interested in Data Mining and
Visualisation
This Presentation
• What is an ESB
• What is Mule
• Mule Applications
• Mule Scalability
What is an ESB
• Enterprise Service Bus
• SOA
– Service Oriented Architecture
• SEDA
– Staged Event-Driven Architecture
• EIP
– Enterprise Integration Patterns
Service Oriented Architecture
• Services
– Shared Business Functions
• Decoupling
– Local and Remote Services
• Distributed Applications
• Directory and Discovery
EIP
• Similar to Design Patterns
– Gang of Four Book
• Shared Terminology
• Names for Common Concepts
• Allows Discussion
• Simplifies Design
(Some) Common Patterns
• Aggregator (268)
• Channel Adapter (127)
• Content Based Router (230)
• Dead Letter Channel (119)
• Selective Consumer (515)
• Message Translator (85)
• Guaranteed Delivery (122)
What is Mule
• Open Source Project
• MuleSoft
– Mule Forge
– Mule Enterprise
– Mule MQ
– Tcat Server
• 2.x - Legacy Applications
• 3.1.1 - Current Community Release
• 3.2.0 - Latest Developer Build
Transports
• Carry Messages between Services
• Connector
– Configuration
– Threading
– Retry
• Endpoint
– Connects Services
– Transform Messages
– Transaction Boundary
Endpoint Examples
<file:inbound-endpoint path="/esb/in"
comparator="org.mule.transport.file.comparator.Older
FirstComparator" reverseOrder="true” />
<jetty:endpoint name="server" host="localhost"
port="60203" path="services/Lookup" />
<imaps:endpoint name=”mail" host="localhost"
password=”hunter2" port="123" user=”adk"/>
<quartz:endpoint name=”batch" repeatCount="10"
repeatInterval=”60000" jobName="job"/>
<jms:inbound-endpoint queue="test.queue"/>
VM
• Communication within VM
• Synchronous or Asynchronous
• Transactional
<vm:connector name="async"
queueEvents="true” />
<vm:inbound-endpoint path="in"
connector-ref="async"/>
JMS
• Queues
• Topics
• Request-Response
<jms:connector name=”jms" specification="1.1"
connectionFactory-ref=”qpid"
username=”adk" password=”hunter2” />
<jms:outbound-endpoint queue=”audit” />
<jms:inbound-endpoint topic=”uk.*.gbp” />
Web Services
• HTTP and HTTPS
– Custom Data Formats
• CXF
– JAX-WS
– SOAP
– WSDL
• REST
CXF JAX-WS
<inbound-endpoint
address="http://localhost:63081/hello" />
<cxf:jaxws-service
serviceClass="org.example.ServiceImpl” />
<cxf:jaxws-client clientClass=”com.example.Client”
wsdlPort="SoapPort”
wsdlLocation="classpath:/save.wsdl”
operation=”output” />
<outbound-endpoint
address="http://www.example.com/services/save” />
File System
• Poll Directory for Input
<file:connector name="input" fileAge="500"
autoDelete="true" pollingFrequency="100"
moveToDirectory="/backup"
moveToPattern="#[header:originalFilename].backup”
/>
• Write Output File
<file:connector name="output" outputAppend="true"
outputPattern="#[function:datestamp]-
#[header:originalFilename]" />
Other Transports
• AJAX • TCP
• jBPM • UDP
• JDBC • FTP
• Quartz • POP3
• STDIO • IMAP
• RMI • SMTP
• Servlet • XMPP
Transformers
• Change message content
• XML to POJO via JAXB
• Applied to endpoints
• Efficient mechanism for processing
• Similar to AOP
Transformer Types
• Encryption and Decryption
• ByteArray to Object
• Expression Evaluator
• XML and JSON
• Compression
• Encoding and Decoding
Filtering
• Apply to endpoints
• Payload Type
• Expression
– XPath, OGNL, JXPath
• RegEx
• Wildcard
• Message Property
• Logical Operators
– And, Or, Not
Filtering Examples
<and-filter>
<message-property-filter
pattern="JMSCorrelationID=1234567890"/>
<message-property-filter pattern="JMSReplyTo=null"/>
</and-filter>
<or-filter>
<payload-type-filter expectedType="java.lang.String"/>
<payload-type-filter expectedType="java.lang.StringBuffer"/>
</or-filter>
<not-filter>
<wildcard-filter pattern=”com.example.*"/>
</not-filter>
Routers
• Control message flow at endpoints
• Used to link services
• Both inbound and outbound
• Specify Filtering and Transformations
Selective Consumer
<inbound>
<selective-consumer-router>
<mulexml:jxpath-filter
expression="msg/header/resultcode =
'success'"/>
</selective-consumer-router>
<forwarding-catch-all-strategy>
<jms:endpoint topic="error.topic"/>
</forwarding-catch-all-strategy>
</inbound>
Idempotent Filter
<inbound>
<idempotent-receiver-router
idExpression="#[message:id]-#[header:foo]">
<simple-text-file-store directory="./idempotent” />
</idempotent-receiver-router>
</inbound>
Filtering Router
<outbound>
<forwarding-catch-all-strategy>
<jms:outbound-endpoint queue="error.queue"/>
</forwarding-catch-all-strategy>
<filtering-router>
<smtp:outbound-endpoint to=”adk@example.com"/>
<payload-type-filter expectedType="java.lang.Exception"/>
</filtering-router>
<filtering-router>
<jms:outbound-endpoint queue="string.queue"/>
<and-filter>
<payload-type-filter expectedType="java.lang.String"/>
<regex-filter pattern="the quick brown (.*)"/>
</and-filter>
</filtering-router>
</outbound>
Services
• Combine Endpoints, Routers and
Components
• Mule 2.0 idiom
• Replaced with Flows and Patterns
– Simpler to use and configure
– Some features still not available
Flows
• Mule 3 idiom
• Inbound endpoint
• Message processors
– Chained execution
– Components
• Outbound endpoint for one-way
• Response for request-response
Components
• Carry out actions
• Business logic in a flow
• Simple components
– Logging
– Passthrough
– Testing
• Spring or POJO
• Web services
Message Processors
• Aggregator and Splitter
• Message Filters
• Recipient List
• Resequencer
• Round Robin
• Wire Tap
Example Flow
<flow name=”OrderFlow">
<file:inbound-endpoint path="/incoming">
<file:filename-filter name=”order-*.xml"/>
</file:inbound-endpoint>
<xml:xslt-transformer xsl-file=”order-transform.xsl"/>
<splitter expression="xpath://order"/>
<!-- The following message processors will be invoked for each order in the xml file -->
<expression-filter expression="xpath://order[@type='book']"/>
<component class=”com.example.BookOrderProcessor"/>
<smtp:outbound-endpoint subject="Order Confirmation" address="#[variable:email]"/>
<jdbc:outbound-endpoint ref=“saveOrder"/ >
<default-exception-strategy>
<jms:outbound-endpoint queue="failedOrders"/>
</default-exception-strategy>
</flow>
JMS Flow
<flow name=”JmsFlow">
<jms:inbound-endpoint queue="in">
<jms:transaction action="ALWAYS_BEGIN" />
</jms:inbound-endpoint>
<component class="com.example.ProcessMessage" />
<jms:outbound-endpoint queue="out">
<jms:transaction action="ALWAYS_JOIN" />
</jms:outbound-endpoint>
<default-exception-strategy>
<commit-transaction exception-pattern="com.example.ExampleException” />
<jms:outbound-endpoint queue="dead.letter">
<jms:transaction action="JOIN_IF_POSSIBLE" />
</jms:outbound-endpoint>
</default-exception-strategy>
</flow>
Complex Flow
<flow name=”TwitterFlow”>
<poll frequency="5000">
<twitter:search query=”example" />
</poll>
<splitter evaluator="json" expression="results" />
<idempotent-message-filter idExpression="#[json:id]" />
<enricher target="#[variable:userDescription]" source="#[json:description]">
<twitter:user userId="#[json:from_user]" />
</enricher>
<salesforce:create type="Opportunity">
<salesforce:sObject>
<salesforce:field key="Name" value="#[json:from_user]" />
<salesforce:field key="Description" value="#[variable:userDescription]" />
</salesforce:sObject>
</salesforce:create>
</flow>
Patterns
• Specific integration Features
• Simple Service
– Exposes Components as Web Services
• Web Service Proxy
• Bridge
– Message Adapter and Transformer
• Validator
– Validates inbound messages synchronously
– Processes asynchronously
Questions?
Mule Application

Mule 2 Application Design


Mule Scalability
• Global Enterprises
• Performance
• Multiple Data Locations, Sources,
Destinations and Owners
• Services
– Software as a Service
– Infrastructure as a Service
– Platform as a Service
Problems
• Performance
– Computation
– Latency
– Throughput
– Infrastructure
• Integration
– Adding Functionality
– Collecting Data
• Security and Identity
• Transactions
Transactions
• ACID
– Atomicity, Consistency, Isolation, Durability
• Easy on one system
• Hard with distributed systems
– Pick your boundaries
• Robust failure handling
Computation
• Parallel Processing
• Asynchronous Tasks
• Orchestration
– Workflow
• Pooling and Threading
• More and faster is better
Bottlenecks
• Bandwidth
– Networking
– Latency and Throughput
• External service providers
• Data sources
– Database
• Messaging
Messaging Backbone
• Sets scalability limits
• Message flow through system
• Infrastructure choices
– JMS
– Web based
– AMQP
– XMPP
– TIBCO
• Hardware solutions available
Clustering
• Various external components
– Database
– Messaging
– Web Services
• Business logic components
– Terracotta
– GigaSpaces XAP
Load Balancing
• Web Services
• Software
• Hardware
– XML Transformation
– SSL
• Round Robin
Questions?

Thanks for Listening


Andrew D Kennedy
grkvlt@apache.org

Anda mungkin juga menyukai