Âscendant

Production Ready Tomcat

2010-03-25T15:39:00.001-05:00

Production Ready Tomcat - Optimizations, Security aspects

To disable directory listing for the entire tomcat server go to $CATALINA_HOME/conf/web.xml and set the listings parameter to false.

<servlet>
<servlet-name>default</servlet-name>
<servlet-class>org.apache.catalina.servlets.DefaultServlet</servlet-class>
<init-param>
<param-name>debug</param-name>
<param-value>0</param-value>
</init-param>
<init-param>
<param-name>listings</param-name>
<param-value>false</param-value>
</init-param>
<load-on-startup>1</load-on-startup>
</servlet>

Switch off application reloading by setting "reloadable" to "false" gives a small run-time performance increase.

Prevent Locking of Exploded Wars by setting antiResourceLocking="true"

By default tomcat uses the MemoryRealm for authentication. Alter the default username and password in tomcat-users.xml with more appropriate ones.

<?xml version='1.0' encoding='utf-8'?>
<tomcat-users>
  <role rolename="tomcat"/>
  <role rolename="manager"/>
  <role rolename="admin"/>
  <user username="tomcat" password="tomcat" roles="tomcat"/>
  <user username="admin" password="XXXX" roles="admin,manager"/>
</tomcat-users>

Make the tomcat-users.xml as readonly by adding readonly="true" to the <Resource> tag in server.xml

Move to http://svn.apache.org/repos/asf/tomcat/trunk/modules/jdbc-pool/ instead of any other connection pool.

       <dependency>
               <groupId>org.apache.tomcat</groupId>
               <artifactId>juli</artifactId>
               <version>6.0.24</version>
       </dependency>
       <dependency>
               <groupId>org.apache.tomcat</groupId>
               <artifactId>jdbc-pool</artifactId>
               <version>1.0.8.5</version>
       </dependency>

Swallow System.out and System.err and direct them to a log file

Rotation of logs

logging.properties file:

.handlers = 1catalina.org.apache.juli.FileHandler

1catalina.org.apache.juli.FileHandler.level = FINE

1catalina.org.apache.juli.FileHandler.directory = ${catalina.base}/logs

1catalina.org.apache.juli.FileHandler.prefix = catalina.

1catalina.org.apache.juli.FileHandler.pattern = ${catalina.base}/logs/catalina.%g.log

1catalina.org.apache.juli.FileHandler.limit = 20000000 # No more than 5x20mb files

1catalina.org.apache.juli.FileHandler.count = 5

Tuning Sun JVM settings for Tomcat.Set JAVA_OPTS variable to catalina.sh.

Try to adopt these settings :

-server

Better for server side applications

-Djava.awt.headless=true

Headless mode is a system configuration in which the display device, keyboard, or mouse is lacking. Sounds unexpected, but actually you can perform different operations in this mode, even with graphic data.

-XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/tomcat/dumps

Prepare thread dumps when a OutOfMemoryError occours. This way you can analyze the thread dumps to figure out the issue.

Memory related tuning parameters:

-Xms

initial heap size is 1/64th of the machine's physical memory

-Xmx

maximum heap size is 1/4th of the machine's physical memory (up to 1GB max).

-XX:PermSize
-XX:MaxPermSize

Garbage Collection Tuning Parameters:

-XX:+UseParallelGC parallel (throughput) garbage collector, or
-XX:+UseConcMarkSweepGC concurrent (low pause time) garbage collector (known as CMS)
-XX:+UseSerialGC serial garbage collector (for smaller applications and systems)

See http://java.sun.com/performance/reference/whitepapers/tuning.html

Apache Tomcat - Apache Log4j

2010-03-25T12:17:00.001-05:00

Configuring Apache Tomcat Logging using Apache Log4j

I will be covering how to use log4j instead of the default logging provided by tomcat based on the reference http://tomcat.apache.org/tomcat-6.0-doc/logging.html

Objective

Using log4j to customize the logging in apache-tomcat
Separate apache-tomcat logs from the application logs

Using log4j to customize the logging in apache-tomcat

Create a file called log4j.properties with the following content and save it into $CATALINA_HOME/lib.

<?xml version="1.0" encoding="utf-8"?>

<!DOCTYPE log4j:configuration SYSTEM "log4j.dtd">

<log4j:configuration xmlns:log4j="http://jakarta.apache.org/log4j/" debug="false">

</layout>

</appender>

</layout>

</appender>

</logger>

</logger>

</logger>

<root>

<appender-ref ref="ROLL" />

<appender-ref ref="CONSOLE" />

</root>

</log4j:configuration>

Download Log4J (v1.2 or later) and place the log4j jar in $CATALINA_HOME/lib.
Build or download the additional logging components. See the extras components documentation for details.
Replace $CATALINA_HOME/bin/tomcat-juli.jar with output/extras/tomcat-juli.jar.
Place output/extras/tomcat-juli-adapters.jar in $CATALINA_HOME/lib.
Delete $CATALINA_BASE/conf/logging.properties to prevent java.util.logging generating zero length log files.
Start Tomcat

Separate apache-tomcat logs from the application logs

The start with following log4j.properties can be used in your web application to divert all the logs to application specific file (app_demo.out).

<?xml version="1.0" encoding="utf-8"?>

<!DOCTYPE log4j:configuration SYSTEM "log4j.dtd">

<log4j:configuration xmlns:log4j="http://jakarta.apache.org/log4j/" debug="false">

</layout>

</appender>

</layout>

</appender>

</logger>

<root>

<appender-ref ref="ROLL" />

<appender-ref ref="CONSOLE" />

</root>

</log4j:configuration>

@TODO

Changing the log levels for various loggers when the app is running

Recurrsion using parallel processing

2010-02-25T22:35:00.001-05:00

Recurrsion using parallel processing

As I was google around I found the Concurrencer: a tool for retrofitting parallelism into sequential Java code. One of the cool thing of this tool is, it can convert the recurrsive program into a parallel program. I learned about ForkJoin framework from concurrency-interest mailinglist. So I quickly wrote a program to see how the performance will be boosted.

The ForkJoin framework will be part of JDK7.0. But the following instructions could help to work with ForkJoin framework on J2SE 6.0. First create a maven2 project and under the root create a lib folder. Download the ForkJoin framework jar "jsr166.jar" from Doug Lea website. Install jsr166.jar as a maven dependency using the following command:

mvn install:install-file -DgroupId=jsr166 -DartifactId=jsr166 -Dversion=166 -Dpackaging=jar -Dfile=jsr166.jar

The following program shows a recurrsive program to generate the fibonacci number and the also using parallel processing. ForkJoinPool is a ExecutorPool that orchestrates the threads.

Fibonacci.java

1    package com.blogspot.ascendant.forkjoin; 
2     
3    import java.util.concurrent.ForkJoinPool; 
4    import java.util.concurrent.RecursiveAction; 
5     
6    public class Fibonacci { 
7     
8        public static void main(String... args) { 
9     
10           final int input = 40; 
11           System.out.printf("Input: %d\n\n", input); 
12    
13           long nowSet1 = System.nanoTime(); 
14    
15           System.out.printf("Fibonacci using recursion: %d\n", fibRecurr(input)); 
16    
17           long laterSet1 = System.nanoTime(); 
18           System.out.printf("Computing time for recursion: %d nanoseconds\n\n", (laterSet1 - nowSet1)); 
19    
20           long nowSet2 = System.nanoTime(); 
21    
22           final int numberOfProcessors = Runtime.getRuntime().availableProcessors(); 
23           System.out.printf("Number of processors %d\n", numberOfProcessors); 
24    
25           final FibonacciParallel fp = new FibonacciParallel(input); 
26    
27           final ForkJoinPool es = new ForkJoinPool(numberOfProcessors); 
28           es.invoke(fp); 
29           es.shutdown(); 
30    
31           System.out.printf("Fibonacci using parallel processing: %d\n", fp.result); 
32    
33           long laterSet2 = System.nanoTime(); 
34           System.out.printf("Computing time for parallel processing: %d nanoseconds\n", (laterSet2 - nowSet2)); 
35       } 
36    
37    
38       /** 
39        * Recursive approach 
40        */ 
41       private static long fibRecurr(final int n) { 
42           if (n < 2) { 
43               return n; 
44           } else { 
45               return fibRecurr(n - 1) + fibRecurr(n - 2); 
46           } 
47       } 
48    
49    
50       /** 
51        * Recursive Task 
52        */ 
53       public static class FibonacciParallel extends RecursiveAction { 
54    
55           private static final long serialVersionUID = -6810496708906992778L; 
56           final int numberToCompute; 
57           volatile long result; 
58    
59           public FibonacciParallel(int numberToCompute) { 
60               this.numberToCompute = numberToCompute; 
61           } 
62    
63           @Override 
64           protected void compute() { 
65               if (numberToCompute < 2) { 
66                   result = numberToCompute; 
67               } else { 
68                   FibonacciParallel future1 = new FibonacciParallel(numberToCompute - 1); 
69                   FibonacciParallel future2 = new FibonacciParallel(numberToCompute - 2); 
70    
71                   invokeAll(future1, future2); 
72    
73                   result = future1.result + future2.result; 
74               } 
75           } 
76       } 
77   } 
78    
79

As we are alternat version of java.util.concurrent package we need to alter the bootclasspath using -Xbootclasspath/a:/home/ramesh/Workspace/ascendant/forkjoin/lib/jsr166.jar option.

java -Xbootclasspath/a:/home/ramesh/Workspace/ascendant/forkjoin/lib/jsr166.jar com.blogspot.ascendant.forkjoin.Fibonacci

ramesh@ramesh-laptop:~$ grep "model name" /proc/cpuinfo
model name : Genuine Intel(R) CPU T2500 @ 2.00GHz
model name : Genuine Intel(R) CPU T2500 @ 2.00GHz

Here is the output on my laptop:
Input: 40

Fibonacci using recursion: 102334155
Computing time for recursion: 1919286038 nanoseconds

Number of processors 2
Fibonacci using parallel processing: 102334155
Computing time for parallel processing: 33836924737 nanoseconds

Process finished with exit code 0

As you see the results using parallel computing is not attractive. I guess this is due to the overload of the thread infrastructure. As the computation becomes more complex, this parallel processing should result better results.

Here is the completed project:
http://www.box.net/shared/egrj42i8xf

Update:

Other people also facing the same performance issue. See http://cs.oswego.edu/pipermail/concurrency-interest/2010-March/006913.html

Upcoming JDBC-Pool -

2010-02-12T18:30:00.001-05:00

Upcoming Tomcat JDBC-Pool an alternate to Jakarta DBCP Component

The Jakarta Commons DBCP internally uses Commons Pool along with other dependencies. It worked me pretty well all the time. As I recently encountered Tomcat JDBC-Pool at Grails.org. I thought of exploring it to see, if it works as a standalone projects. And certainly works with out tomcat and usage is very straight forward and also an inplace replacement to Commons DBCP with very minimum change i.e., use org.apache.tomcat.jdbc.pool.DataSource instead of org.apache.commons.dbcp.BasicDataSource.

You can download the jars from http://people.apache.org/~fhanik/jdbc-pool/v1.0.8.5/ or build
it from the svn repository located at http://svn.apache.org/repos/asf/tomcat/trunk/modules/jdbc-pool. As it is not available in any public standard maven2 repository I manually updated local repository using:

mvn install:install-file -DgroupId=org.apache.tomcat -DartifactId=jdbc-pool -Dversion=1.0.8.5 -Dpackaging=jar -Dfile=tomcat-jdbc.jar

For apache maven lovers, below are the dependenies required:

        <dependency>
            <groupId>org.apache.tomcat</groupId>
            <artifactId>jdbc-pool</artifactId>
            <version>1.0.8.5</version>
        </dependency>

        <dependency>
            <groupId>org.apache.tomcat</groupId>
            <artifactId>juli</artifactId>
            <version>6.0.20</version>
        </dependency>

JDBC properties in app.properties file looks a below:

# Oracle configuration
jdbc.driverClassName=oracle.jdbc.driver.OracleDriver
jdbc.url=jdbc:oracle:thin:@//<<Server>>:<<PORT>/<<Service Name>>
jdbc.username=<<Database User>>
jdbc.password=<<Database Password>>

Spring applicationContext.xml

<context:property-placeholder location="classpath:app.properties" />

<bean id="datasource" class="org.apache.tomcat.jdbc.pool.DataSource"
          destroy-method="close"
          p:driverClassName="${jdbc.driverClassName}"
          p:url="${jdbc.url}"
          p:username="${jdbc.username}"
          p:password="${jdbc.password}"
          p:initialSize="5"
          p:minIdle="5"
          p:maxIdle="10"
          p:maxActive="20"
          p:maxWait="6000"
          p:jmxEnabled="true"
          p:jdbcInterceptors="org.apache.tomcat.jdbc.pool.interceptor.ConnectionState;org.apache.tomcat.jdbc.pool.interceptor.StatementFinalizer;org.apache.tomcat.jdbc.pool.interceptor.SlowQueryReportJmx(threshold=10000)"
          p:removeAbandoned="true"
          p:removeAbandonedTimeout="60"
          p:logAbandoned="true"
          p:testOnBorrow="true"
          p:testOnReturn="false"
          p:testWhileIdle="false"
          p:useEquals="true"
          p:fairQueue="true"
          p:timeBetweenEvictionRunsMillis="30000"
          p:minEvictableIdleTimeMillis="30000"
          p:validationInterval="1800000"
          p:validationQuery="select 1 from dual"
          p:accessToUnderlyingConnectionAllowed="true"
          />

I still need to see how it works when dealing with vendor specific JDBC features.

Concurrent URL checker using Completi...

2010-02-12T16:57:00.001-05:00

Concurrent URL checker using Completion Service

From javadocs:

"A service that decouples the production of new asynchronous tasks from the consumption of the results of completed tasks. Producers submit tasks for execution. Consumers take completed tasks and process their results in the order they complete. A CompletionService can for example be used to manage asynchronous IO, in which tasks that perform reads are submitted in one part of a program or system, and then acted upon in a different part of the program when the reads complete, possibly in a different order than they were requested.

Typically, a CompletionService relies on a separate Executor to actually execute the tasks, in which case the CompletionService only manages an internal completion queue. The ExecutorCompletionService class provides an implementation of this approach."

The program:

Takes set of urls as input. This is typically done by a producer.
The CallableUrlChecker class checks one url and produces the Result.

UrlChecker.java

1    package com.blogspot.ascendant.urlchecker; 
2     
3    import org.apache.http.HttpHost; 
4    import org.apache.http.HttpResponse; 
5    import org.apache.http.StatusLine; 
6    import org.apache.http.client.HttpClient; 
7    import org.apache.http.client.methods.HttpGet; 
8    import org.apache.http.impl.client.DefaultHttpClient; 
9    import org.apache.http.protocol.BasicHttpContext; 
10   import org.apache.http.protocol.ExecutionContext; 
11   import org.apache.http.protocol.HttpContext; 
12    
13   import java.util.Set; 
14   import java.util.TreeSet; 
15   import java.util.concurrent.*; 
16    
17   public class UrlChecker { 
18    
19       public static void main(String... args) { 
20    
21           final Set<String> urls = getUrlsToCheck(); 
22    
23           // The executor that acutually executes the Url verification.  
24           final ExecutorService pool = Executors.newFixedThreadPool(10); 
25    
26           // The completion service that manages completed task queue 
27           final CompletionService<Result> completionService = new ExecutorCompletionService<Result>(
28                   pool 
29           ); 
30    
31           // Submit the url check tasks to the completion service 
32           for (final String url : urls) { 
33               completionService.submit(new CallableUrlChecker(url)); 
34           } 
35    
36           Future<Result> futureResult; 
37           Result result; 
38    
39           // Total urls submitted 
40           final int totalUrlsToCheck = urls.size(); 
41    
42           int urlCounter = 0; 
43           try { 
44               while (urlCounter != totalUrlsToCheck) { 
45                   try { 
46                       // Retrieves and removes the Future representing the next completed task, 
47                       // waiting if none are yet present. 
48                       futureResult = completionService.take(); 
49    
50                       // May result a null... 
51                       if (futureResult != null) { 
52                           // get the result with a time wait... 
53                           result = futureResult.get(3, TimeUnit.SECONDS); 
54    
55                           // further process the result 
56                           System.out.println(urlCounter + " : " + result); 
57                       } 
58                   } catch (ExecutionException e) { 
59                       System.err.println("Failed to check the url: " + e.getMessage()); 
60                   } catch (InterruptedException e) { 
61                       Thread.currentThread().interrupt(); 
62                   } catch (TimeoutException e) { 
63                       System.err.println("Failed to check the url: " + e.getMessage()); 
64                   } 
65    
66                   urlCounter++; 
67               } 
68    
69           } finally { 
70               // Shutdown the executor pool if you are done 
71               pool.shutdown(); 
72           } 
73       } 
74    
75       /** 
76        * Set of urls to check 
77        * 
78        * @return non-null set 
79        */ 
80       private static Set<String> getUrlsToCheck() { 
81           final Set<String> urls = new TreeSet<String>(); 
82           for (int i = 0; i <= 5; i++) { 
83               urls.add(String.format("http://test%d.com", i)); 
84           } 
85           return urls; 
86       } 
87    
88       /** 
89        * The callable task that actually verifies the URL and prepares the result... 
90        */ 
91       public static class CallableUrlChecker implements Callable<Result> { 
92           private final String urlToCheck; 
93    
94           public CallableUrlChecker(String urlToCheck) { 
95               this.urlToCheck = urlToCheck; 
96           } 
97    
98           public Result call() throws Exception { 
99               HttpClient httpclient = new DefaultHttpClient(); 
100              HttpGet httpget = new HttpGet(urlToCheck);
101              HttpContext hc = new BasicHttpContext(); 
102   
103              Result result = new Result(); 
104              result.setUrl(urlToCheck); 
105              try { 
106                  HttpResponse response = httpclient.execute(httpget, hc); 
107                  HttpHost target = (HttpHost) hc.getAttribute( 
108                          ExecutionContext.HTTP_TARGET_HOST 
109                  ); 
110   
111                  StatusLine status = response.getStatusLine(); 
112   
113                  result.setStatusCode(status.getStatusCode()); 
114                  result.setStatusPhrase(status.getReasonPhrase()); 
115                  result.setRedirectUrl(target.toURI()); 
116   
117              } catch (Exception e) { 
118                  result.setStatusPhrase(e.getMessage()); 
119              } finally { 
120                  httpclient.getConnectionManager().shutdown(); 
121              } 
122   
123              return result; 
124          } 
125      } 
126   
127      /** 
128       * Result encapsulates the data that is returned after completion of the task 
129       */ 
130      public static class Result { 
131          private String url; 
132          private String redirectUrl; 
133          private int statusCode; 
134          private String statusPhrase; 
135   
136          public String getUrl() { 
137              return url; 
138          } 
139   
140          public void setUrl(String url) { 
141              this.url = url; 
142          } 
143   
144          public String getRedirectUrl() { 
145              return redirectUrl; 
146          } 
147   
148          public void setRedirectUrl(String redirectUrl) { 
149              this.redirectUrl = redirectUrl; 
150          } 
151   
152          public int getStatusCode() { 
153              return statusCode; 
154          } 
155   
156          public void setStatusCode(int statusCode) { 
157              this.statusCode = statusCode; 
158          } 
159   
160          public String getStatusPhrase() { 
161              return statusPhrase; 
162          } 
163   
164          public void setStatusPhrase(String statusPhrase) { 
165              this.statusPhrase = statusPhrase; 
166          } 
167   
168          @Override 
169          public String toString() { 
170              return "Result{" + 
171                      "url='" + url + '\'' + 
172                      ", redirectUrl='" + redirectUrl + '\'' + 
173                      ", statusCode=" + statusCode + 
174                      ", statusPhrase='" + statusPhrase + '\'' + 
175                      '}'; 
176          } 
177      } 
178  } 
179

The above solution can be improved with the following changes.

1) Producer and Consumer are in only class; and this can be separated. Run the producer in a separate thread and the consumer part in a separate thread.

2) Can leverage Fork/Join Framework

Maven2 Dependencies

HttpComponents.

<dependency>

<groupId>org.apache.httpcomponents</groupId>
<artifactId>httpclient</artifactId>
<version>4.1-alpha1</version>

</dependency>

Download code for URLChecker - Java Concurrent Completion Service

Amazon Product Advertising API

2009-09-06T18:51:00.001-05:00

Product Advertising API Signed Requests - Java SOAP using Spring Web Services

Product Advertising API - https://affiliate-program.amazon.com/gp/advertising/api/detail/main.html/178-0006257-2456255

We are using Amazon Web Services to advertise products on one of my clients websites. We used Springframework as the dependency container and Spring Web Services as web service platform as we Spring-WS favors contract-first web services. With the recent changes in the API all the requests must be signed. And the sample code (http://developer.amazonwebservices.com/connect/entry.jspa?externalID=2479&categoryID=14) helped me as a reference to migrate our code base to support the signed request requirement.

In the below sections I will walk through the steps to make signed requests using Spring Web Services.

Project Setup:

We used maven2 as our build & reporting tool we followed the maven project directory structure and is as below:

awsprocess
└───src
    ├───main
    │   ├───java
    │   │   └───com
    │   │       └───ascendant76
    │   │           └───awsdemo
    │   │               ├───client
    │   │               └───service
    │   ├───resources
    │   └───webapp
    └───test
        ├───java
        └───resources

The project object model uses few plug-ins:

maven-compiler-plugin: To compile the code that is compatible with J2SE 5.0.

maven-surefire-plugin: To execute the unit test cases. I used TestNG(http://testng.org/doc/index.html) for the test cases and their by the testng dependency is imported.

        <dependency>
           <groupId>org.testng</groupId>
           <artifactId>testng</artifactId>
           <classifier>jdk15</classifier>
           <version>5.9</version>
       </dependency>

maven-jaxb2-plugin: Creates JAXB2 binding from XSD. The plug-in scans the xjb files and generates the bindings in our case it AWSECommerceService.xjb in the resources folder.

AWSECommerceService.xjb:

<?xml version='1.0' encoding='utf-8' ?>
<jxb:bindings version="2.0" xmlns:jxb="http://java.sun.com/xml/ns/jaxb"
   xmlns:xs="http://www.w3.org/2001/XMLSchema">

   <jxb:bindings node="/xs:schema"
       schemaLocation="AWSECommerceService.xsd">
       <jxb:globalBindings fixedAttributeAsConstantProperty="false"
           collectionType="java.util.ArrayList" typesafeEnumBase="xs:NCName"
           choiceContentProperty="false" typesafeEnumMemberName="generateError"
           enableFailFastCheck="true" generateIsSetMethod="true"
           underscoreBinding="asCharInWord"
           enableJavaNamingConventions="true">
       </jxb:globalBindings>
       <jxb:schemaBindings>
           <jxb:package name="com.ascendant76.awsdemo.model" />
       </jxb:schemaBindings>
   </jxb:bindings>
</jxb:bindings>

The AWSECommerceService.xsd is extracted manually from the web service contract schema from the WSDL located at http://ecs.amazonaws.com/AWSECommerceService/2009-07-01/AWSECommerceService.wsdl

Account Setup:

Login to amazon at http://aws-portal.amazon.com/gp/aws/developer/account/index.html?ie=UTF8&action=activity-summary. Goto Access Credentials section by following this URL https://aws-portal.amazon.com/gp/aws/developer/account/index.html?ie=UTF8&action=access-key. This page allows you to create new X.509 certificates. Select the Create option to create your certificates and download them. You will have two files pk-XXXX.pem and cert-XXXX.pem

Note: You can download the private key file (pk-XXXX.pem) at the time of certificate creation only and will not be available later. If you lose it, you have to create a new certificate. So keep it safe.

Now we need to convert the certificate into the PCKS12 format. We need openssl (http://www.openssl.org) to create the certificate in pcks12 format.

C:\openssl pkcs12 -export -in cert-XXXX.pem -inkey pk-XXXX.pem -name ramesh_cert -out ramesh_cert.p12

The pkcs12 file will be generated. Execute the following command to verify the generated certificate using the user name.

C:\jdk1.6.0_13\bin\keytool -v -list -storetype pkcs12 -keystore ramesh_cert.p12

Enter keystore password: ramesh

The output appears something like below:

Keystore type: pkcs12
Keystore provider: SunJSSE

Your keystore contains 1 entry

Alias name: ramesh_cert
Creation date: Aug 28, 2009
Entry type: keyEntry
Certificate chain length: 1
Certificate[1]:
Owner: CN=zzp69oziqnw6, OU=AWS-Developers, O=Amazon.com, C=US
Issuer: CN=AWS Limited-Assurance CA, OU=AWS, O=Amazon.com, C=US
Serial number: d61749a157
Valid from: Fri Aug 28 16:49:58 EDT 2009 until: Sat Aug 28 16:49:58 EDT 2010
Certificate fingerprints:
         MD5: 60:BF:22:AF:51:F8:5A:72:36:8F:E8:69:1D:35:8C:C0
         SHA1: D0:D9:5A:60:38:35:98:45:C3:41:CE:1A:8F:76:DC:39:80:3A:02:1B
**************************************************************************************

Once you verify the certificate generated is okay with the given user
name, copy the certificate generated above - ramesh_cert.p12 - into the
resources directory.

Codebase:

The dependency diagram of various collaborators is given below.

The AwsGateway is the sub-class of org.springframework.ws.client.core.support.WebServiceGatewaySupport is responsible for marshaling the java objects in to SOAP messages and sends to location mentioned in the service binding of the WSDL and in this case https://ecs.amazonaws.com/onca/soap?Service=AWSECommerceService. The WebServiceGatewaySupport delegates the responsibility to the Marshaller to marshal java request objects to SOAP messages and UnMarshaller to marshal the SOAP response back to Java object.

    
   <bean id="marshaller" class="org.springframework.oxm.jaxb.Jaxb2Marshaller">
       <property name="schema" value="classpath:AWSECommerceService.xsd" />
       <property name="contextPath" value="com.ascendant76.awsdemo.model" />
   </bean>

Spring-WS supports two types of messages factories to deal with SOAP messages org.springframework.ws.soap.axiom.AxiomSoapMessageFactory and org.springframework.ws.soap.saaj.SaajSoapMessageFactory. And I choose to use SAAJ (https://saaj.dev.java.net/nonav/spec-1.3/api/).

   
   <bean id="messageFactory" class="org.springframework.ws.soap.saaj.SaajSoapMessageFactory" />

Now the web service gateway definition looks as below:

    <bean id="awsGateway" class="com.ascendant76.awsdemo.service.AwsGateway">
       <constructor-arg value="${aws.accessKey}" />
       <property name="defaultUri" value="https://ecs.amazonaws.com/onca/soap?Service=AWSECommerceService" />
       <property name="marshaller" ref="marshaller" />
       <property name="unmarshaller" ref="marshaller" />
       <property name="messageFactory" ref="messageFactory" />
       <property name="interceptors">
           <list>
               <ref bean="signatureSecurityInterceptor" />
           </list>
       </property>
   </bean>

As security is an orthogonal concept, spring-ws handles it using interceptors. We use WSS4J for the authentication to produce signed requests. Amazon API rely on Timestamp and Signature.

    <bean id="signatureSecurityInterceptor" class="org.springframework.ws.soap.security.wss4j.Wss4jSecurityInterceptor">
       <property name="securementActions" value="Timestamp Signature" />
       <property name="timestampPrecisionInMilliseconds" value="true" />
       <property name="securementSignatureKeyIdentifier" value="DirectReference" />
       <property name="securementUsername" value="${aws.keystoreUser}" />
       <property name="securementPassword" value="${aws.keystorePassword}" />
       <property name="securementSignatureCrypto" ref="cryptoFactoryBean" />
       <property name="securementCallbackHandler" ref="passwordCallback" />
   </bean>

   <bean id="cryptoFactoryBean" class="org.springframework.ws.soap.security.wss4j.support.CryptoFactoryBean">
       <property name="cryptoProvider" value="org.apache.ws.security.components.crypto.Merlin" />
       <property name="keyStoreLocation" value="classpath:${aws.keystoreUser}.p12" />
       <property name="keyStorePassword" value="${aws.keystorePassword}" />
       <property name="keyStoreType" value="pkcs12" />
       <property name="keyStoreProvider" value="SunJSSE" />
   </bean>

   <bean id="passwordCallback" class="com.ascendant76.awsdemo.service.PasswordCallback">
       <constructor-arg index="0" value="${aws.keystorePassword}" />
   </bean>

To keep all the properties in one place we can use property place holder. The keystore password is being served from properties file:

    <context:property-placeholder location="classpath:app.properties" />

Change the properties files to reflect your credentials

# Amazon Associate Access Key
aws.accessKey=<<AWS ACCESSKEY comes here>>
aws.keystoreUser=<<KEY STORE comes here>>
aws.keystorePassword=<<KEY STORE password>>

Now at this stage we are ready to invoke the web service. Execute mvn clean test to execute the test cases.

Download the code
You can download the code from here: Download the Code

References:
http://www.openssl.org/docs/apps/pkcs12.html

Nandan Nilekani's ideas for India's future

2009-08-25T07:33:00.000-05:00

Interesting video by Nandan Nilekani and his ideas for India's future. The problem domain is mapped to the following categories.

* Ideas that have arrived
* Ideas in progress
* Ideas in conflict
* Ideas in anticipation

What I like in this presentation is the approach based on the above four fundamentals.

Don't mess with Java naming convensions - Java Generics

2009-08-06T12:44:00.000-05:00

Today I found the following program over the internet. First time when I saw it I thought it won't compile.

import java.util.ArrayList;

public class ListProvider {
public static ArrayList sayHello(ArrayList ar) {
return ar;
}
public static void main(String[] args) {
System.out.println(sayHello("I am a list"));
}
}

Now I rewrote the program following java naming conventions.

public class ListProvider {
public static T sayHello(T ar) {
return ar;
}
public static void main(String[] args) {
System.out.println(sayHello("I am a list"));
}
}

is a Generic Type. It is hiding java.util.ArrayList in the first program. Second program it is bit clean using T as type.

Interesting Quote

2009-06-28T17:22:00.000-05:00

Re-use before Buy
Buy before build,
Build for Re-use

- John Zachman

Checking the file association in Windows from Command line

2009-05-19T09:34:00.002-05:00

Checking the file association in Windows:

The "assoc" will let you know the file type associated with the extension
C:\>assoc .txt
.txt=txtfile

Now the "ftype" displays or modifies file types used in file extension associations
C:\>ftype txtfile
txtfile=%SystemRoot%\system32\NOTEPAD.EXE %1

12 days of christmas - Indian Version

2008-12-22T09:22:00.000-05:00

I was stumbled to see this video from my boss. A very unusual e-mail I can say. Check this one.

When two in Tango ...

2008-10-19T17:24:00.001-05:00

When you look in to my eyes...
.. eye lids stop flittering
.. heart beats fast
.. minds freezes the frame of yours in my eye lids

Let us tango together..
.. the way no one did
.. the way no one will
.. let the eyes be in to my eyes

Let us not stop this...
.. as long as the clock ticks
.. as long as the earth rotates
.. as long as we can with stand standing

If you feel it is the last step we make ..
.. let our eyes lock together
.. let the warm hug seize the air separting us
.. let the God feel jealous of us

Let the tago continue ...
Let the eyes lock one another...
Let the time ends ...

Appropriate English Usage 1

2008-10-19T16:33:00.002-05:00

When some one acknowledges me by saying "Thank you!", my immediate reply used to be "No problem!". Today I learned that it is not traditional and sounds less polite.

Yahoo Query Language (YQL)

2008-10-14T08:38:00.001-05:00

You may be interested of this Yahoo Query Language (YQL) . This is really intereting as they are not exposing their database tables but enabling their domain model ( structured data) using SQL. Behind the scenes they can map to their tables or what ever it could be.

Yahoo! makes a lot of structured data available to developers, primarily through its web services. These services require developers to locate the right URLs and documentation to access and query them which can result in a very fragmented experience. The YQL platform provides a single endpoint service that enables developers to query, filter and combine data across Yahoo! and beyond. YQL exposes a SQL-like SELECT syntax that that is both familiar to developers and expressive enough for getting the right data. Through the SHOW and DESC commands we enable developers to discover the available data sources and structure without opening another web browser.

Yahoo Query Language (YQL)

2008-10-14T08:38:00.000-05:00

No One Needs to Know by Shania Twain

2008-10-08T05:20:00.001-05:00

No One Needs to Know by Shania Twain

2008-10-08T05:20:00.000-05:00

OMG - I Will Kill Her

2008-10-07T05:50:00.001-05:00

Is that Agony or Jealousy?

OMG - I Will Kill Her

2008-10-07T05:50:00.000-05:00

Is that Agony or Jealousy?

Nice Song

2008-10-07T05:45:00.001-05:00

Bit Girl y though!

Nice Song

2008-10-07T05:45:00.000-05:00

Bit Girl y though!

LOBO - I'D LOVE YOU TO WANT ME

2008-09-27T11:20:00.001-05:00

LOBO - I'D LOVE YOU TO WANT ME

2008-09-27T11:20:00.000-05:00

Testing Dog shit - Bull shit

2008-09-25T14:20:00.001-05:00

Officials in an Israeli city are using DNA analysis to track dog owners who allow their pets to foul the streets.

Read more here!

On the other side people are suffering with no food and some got abundant money test the shit.

Testing Dog shit - Bull shit

2008-09-25T14:20:00.000-05:00