Optimization Techniques for Java & J2EE Applications

7/26/2019 Optimization Techniques for Java & J2EE Applications

1/10

Performance Optimization Techniques In Java & J2ee

http://marancollects.blogspot.com 1/10

PERFORMANCE OPTIMIZATION TECHNIQUES IN JAVA & J2EE

Optimizing practices for a class:

Use primitive data types instead of objects as instance variables

Keep constructors simple and inheritance hierarchies shallow

Avoid initializing instance variable more than once

Eliminate unnecessary type casts

Enumerators are faster than Iterators due to the specific implementation

Ints are the fastest datatype

Use static variables for fields that only need to be assigned once

Final classes can be faster

Use interfaces, so that the actual underlying data structure can be actually

changed without impacting the client program

Null out references when they are no longer needed so that gc can collect

them when it is running.

Avoid character processing using methods like charAt(), setCharAt() methods

Use local variables in preference to class variables

Compound operators such as n += 4; are faster than n = n + 4; because

fewer bytecodes are generated.

Shifting by powers of two is faster than multiplying.

Multiplication is faster than exponentiation.

int increments are faster than byte or short increments.

Floating point increments are much slower than any integral increment.

It can help to copy slower-access vars to fast local vars if you are going to

operate on them repeatedly, as in a loop.

For some applications that access the date a lot, it can help to set the local

timezone to be GMT, so that no conversion has to take place.

Minimize creation of short-lived objects.

Avoid excessive writing to the Java console

Minimize the number of JNI calls

Avoid unnecessary instanceof operation Avoid using long divides.

Persistency adds overheads that make persistent objects slower.

Use pipelining and parallelism. Designing applications to support lots of

parallel processes working on easily distinguished subsets of the work makes

the application faster. If there are multiple steps to processing, try to design


2/10



your application so that subsequent steps can start working on the portion of

data that any prior process has finished, instead of having to wait until the

prior process is complete.

Use bitshift instead of multiplying or dividing by powers of 2.

Scope of variables can impact performance.

Avoid repeatedly executing a parse [or other constant expression] in a loop

when the execution can be achieved once outside the loop.

Use exception only where you need them

Optimizing methods in a class:

Avoid creating temporary objects within frequently called methods

Define methods that accept reusable objects to be filled in with data, rather

than methods that return objects holding that data.

Use methods that alter objects without making copies

Eliminate repeatedly called methods where alternatives are possible [For

example, if you are processing a vector in a loop, get the size of the vector in

the beginning in a length local variable rather than calling size() method on

the vector in the condition part of the for loop]

Use private and static methods and final classes to encourage inlining by the

compiler

Inline methods manually where it is appropriate

Keep methods short and simple to make them automatic inlining candidates

Access to private member by the inner classes to the enclosing class goes by

a method call even if it is not intended to.

Keep synchronized methods out of loops if you possibly can.

Optimizing loops & conditional statements:

Reduce the number of temporary objects being used,especially in loops.

Use short-circuit Boolean operators instead of normal Boolean operators

Eliminate unnecessary repeated method calls from loops

Move loop invariants outside the loop.

Perform the loop backwards (this actually performs slightly faster thanforward loops do). [Actually it is converting the test to compare against 0 that

makes the difference].

It can help to copy slower-access vars to fast local vars if you are going to

operate on them repeatedly, as in a loop.

Use exception terminated infinite loops for long loops.


3/10



Whatever can be calculated outside of a loop should be calculated outside of

the loop.

For multidimensional arrays store a reference for the currently accessed row

in a variable.

Cache the size of the collection in a local variable to use in a loop instead of

repeatedly calling collection.size().

Optimizing Strings:

Use StringBuffer instead of String concat + operator

Formatting numbers using java.text.DecimalFormat is always slower than

Double.toString(double) method, because internally

java.text.DecimalFormat() calls Double.toString(double) then parses and

converts the results.

Convert String to char[] arrays to process characters rather than accessing

one at a time using String.charAt() method

Creating Double from string is slow

Intern() Strings to enable (==) comparisions

Use char arrays for all character processing in loops, rather than using String

or StringBuffer classes

Set the initial string buffer size to maximum if it known.

StringTokenizer is very inefficient, and can be optimized by storing the string

and delimiter in a character array instead of in String

DONT create static strings via new().

Where the compiler cannot resolve concatenated strings at compile time, the

code should be converted to StringBuffer appends, and the StringBuffer

should be appropriately sized rather than using the default size.

The compiler concatenates strings where they are fully resolvable, so don t

move these concatenations to runtime with StringBuffer.

Optimizing arrays, vectors & collections:

Create copies of simple array by initializing them through loops or by using

System.arraycopy(), create copies of complex arrays by cloning them Iterator.hasNext() and Enumerator.hasMoreElements() need not be

repeatedly called when the size of the collection is known. Use

collection.size() and a loop counter instead.

ArrayList is faster than Vector


4/10



Go for a non-synchronized version of collection unless used in a threaded

application

LinkedList is faster than ArrayList for inserting elements to the front of the

array, but slower at indexed lookup

Accessing arrays is much faster than accessing vectors, String, and

StringBuffer

Vector is convenient to use, but inefficient. Ensure that elementAt() is not

used inside a loop.

Re-use Hashtables by using Hashtable.clear().

Removing elements from a Vector will necessitate copying within the Vector if

the element is removed from anywhere other than the end of the collection.

Presizing collections to the expected size is more efficient than using the

default size and letting the collection grow.

For multidimensional arrays store a reference for the currently accessed row

in a variable.

When adding multiple items to a collection, add them all in one call if

possible.

Optimizing threads:

Avoid synchronization where possible

Code a multi-thread for multi-processor machine.

Synchronizing on method rather than code blocks is slightly faster

Polling is only acceptable when waiting for outside events and should be

performed in a "side" thread. Use wait/notify instead.

Prioritize threads. Use notify instead of notifyAll. Use synchronization

sparingly.

Keep synchronized methods out of loops if you possibly can.

Maximize thread lifetimes and minimize thread creation/destruction cycles.

Use Thread pools where these improve performance.

Use Thread.sleep() instead of a for loop for measured delays.

Use a separate timer thread to timeout socket operations Use more server threads if multiple connections have high latency.

Optimizing IO?

Use Buffered IO classes

File information such as file length requires a system call and can be slow

Dont use it often. Similarly, System.currentTimeMillis() uses a native call to


5/10



get current time. Make sure your production code does not have this

statement.

Many java.io.File methods are system calls which can be slow

Use BufferedIO streams to access URLConnection sInput/Output streams.

Use the transient keyword to define fields to avoid having those fields

serialized. Examine serialized objects to determine which fields do not need to

be serialized for the application to work.

Increase server listen queues for high load or high latency servers.

Optimizing exceptions:

Be specific while handling the exception in your catch block.

Be specific while throwing exception in your throws clause.

Do not use Exception Handling to control programming flow.

Very little overhead is imposed by using exception handling mechanism

unless an exception occurs or thrown a new exception object explicitly.

Always use the finally block to release the resources to prevent resource

leaks.

Handle exceptions locally wherever possible.

Do not use Exception handling in loops.

Optimizing performance in JDBC:

Use batch transactions.

Choose right isolation level as per your requirement.

TRANSACTION_READ_UNCOMMITED gives best performance for concurrent

transaction based applications. TRANSACTION_NONE gives best performance

for non-concurrent transaction based applications.

Use PreparedStatement when you execute the same statement more than

once.

Use CallableStatement when you want result from multiple and complex

statements for a single request.

Use batch update facility available in Statements.

Use batch retrieval facility available in Statements or ResultSet. Set up proper direction for processing rows.

Use proper getXXX() methods.

Close ResultSet, Statement and Connection whenever you finish your work

with them.

Write precise SQL queries.


6/10



Cache read-only and read-mostly tables data.

Fetch small amount of data iteratively rather than whole data at once when

retrieving large amount of data like searching database etc.

Optimizing servlets:

Use init() method to cache static data

Use StringBuffer rather than using + operator when you concatenate multiple

strings

Use print() method rather than println() method

Use ServletOutputStream rather than PrintWriter to send binary data

Initialize the PrintWriter with proper size

Flush the data partly

Minimize code in the synchronized block

Set the content length

Release resources in destroy() method.

Implement getLastModified() method to use browser cache and server cache

Use application server caching facility

Use Mixed session mechanisms such as HttpSession with hidden fields

Remove HttpSession objects explicitly in your program whenever you finish

the task

Reduce session time out value as much as possible

Use 'transient' variables to reduce serialization overhead if your HttpSession

tracking mechanism uses serialization process.

Disable servlet auto reloading feature.

Use thread pool for your servlet engine and define the size as per application

requirement.

Optimizing a JSP page:

Use jspInit() method to cache static data

Use StringBuffer rather than using + operator when you concatenate multiple

strings

Use print() method rather than println() method Use ServletOutputStream instead of JSPWriter to send binary data

Initialize the 'out' object (implicit object) with proper size in the page

directive.

Flush the data partly

Minimize code in the synchronized block


7/10



Set the content length

Release resources in jspDestroy() method.

Give 'false' value to the session in the page directive to avoid session object

creation.

Use include directive instead of include action when you want to include the

child page content in the translation phase.

Avoid giving unnecessary scope in the 'useBean' action.

Do not use custom tags if you do not have reusability.

Use application server caching facility

Use Mixed session mechanisms such as 'session' with hidden fields

Use 'session' and 'application' as cache.

Use caching tags provided by different organizations like openSymphony.com

Remove 'session' objects explicitly in your program whenever you finish the

task

Reduce session time out value as much as possible

Use 'transient' variables to reduce serialization overhead if your session

tracking mechanism uses serialization process.

Disable JSP auto reloading feature.

Use thread pool for your JSP engine and define the size of thread pool as per

application requirement.

Optimizing practices for EJB:

Use Local interfaces that are available in EJB2.0 if you deploy both EJB Client

and EJB in the same EJB Server.

Use Vendor specific pass-by-reference implementation to make EJB1.1

remote EJBs as Local EJBs if you deploy both EJB Client and EJB in the same

EJB Server.

Wrap entity beans with session beans to reduce network calls and to promote

declarative transactions. Optionally, make entity beans as local beans where

appropriate.

Make coarse grained session and entity beans to reduce network calls. Control serialization by modifying unnecessary data variables with 'transient'

key word to avoid unnecessary data transfer over network.

Cache EJBHome references to avoid JNDI lookup overhead.


8/10



Avoid transaction overhead for non-transactional methods of session beans by

declaring 'NotSupported' or 'Never' transaction attributes that avoid further

propagation of transactions.

Set proper transaction age(time-out) to avoid large transaction.

Use clustering for scalability.

Tune thread count for EJB Server to increase EJB Server capacity.

Choose servlet's HttpSession object rather than Stateful session bean to

maintain client state if you don't require component architecture and services

of Stateful session bean.

Choose best EJB Server by testing with ECperf tool kit.

Choose normal java object over EJB if you don't want built-in services such as

RMI/IIOP, transactions, security, persistence, resource pooling, thread safe,

client state etc..

Optimizing stateless session beans:

Tune the Stateless session beans pool size to avoid overhead of creation and

destruction of beans.

Use setSessionContext() or ejbCreate() method to cache bean specific

resources.

Release acquired resources in ejbRemove() method

Optimizing stateful session beans:

Tune Stateful session beans cache size to avoid overhead of activation and

passivation process.

Set optimal Stateful session bean age(time-out) to avoid resource congestion.

Use 'transient' key word for unnecessary variables of Stateful session bean to

avoid serialization overhead.

Remove Stateful session beans explicitly from client using remove() method.

Optimizing entity beans:

Tune the entity beans pool size to avoid overhead of creation and destruction

of beans.

Tune the entity beans cache size to avoid overhead of activation, passivationand database calls.

Use setEntityContext() method to cache bean specific resources.

Release acquired resources in unSetEntityContext() method

Use Lazy loading to avoid unnecessary pre-loading of child data.


9/10



Choose optimal transaction isolation level to avoid blocking of other

transactional clients.

Use proper locking strategy.

Make read-only entity beans for read only operations.

Use dirty flag to avoid unchanged buffer data updation.

Commit the data after the transaction completes to reduce database calls in

between transaction.

Do bulk updates to reduce database calls.

Use CMP rather than BMP to utilize built-in performance optimization facilities

of CMP.

Use ejbHome() methods for global operations.

Tune connection pool size to reduce overhead of creation and destruction of

database connections.

Use JDBC tuning techniques in BMP.

Use direct JDBC rather than using entity beans when dealing with huge data

such as searching a large database.

Use business logic that is specific to entity bean data.

Optimizing message driven beans:

Tune the Message driven beans pool size to promote concurrent processing of

messages.

Use setMesssageDrivenContext() or ejbCreate() method to cache bean

specific resources.

Release acquired resources in ejbRemove() method.

Use JMS tuning techniques in Message driven beans.

Optimizing Java Message Service (JMS):

Start producer connection after you start consumer.

Use concurrent processing of messages.

Close the Connection when finished.

Choose either DUPS_OK_ACKNOWLEDGE or AUTO_ACKNOWLEDGE rather

than CLIENT_ACKNOWLEDGE. Control transactions by using separate transactional session for transactional

messages and non-transactional session for non-transactional messages.

Close session object when finished.

Make Destination with less capacity and send messages accordingly.

Set high Redelivery delay time to reduce network traffic.


10/10



Set less Redelivery limit for reducing number of message hits.

Choose non-durable messages wherever appropriate to avoid persistence

overhead.

Set optimal message age (TimeToLive value).

Receive messages asynchronously.

Close Producer/Consumer when finished.

Choose message type carefully to avoid unnecessary memory overhead.

Use 'transient' key word for variables of ObjectMessage which need not be

transferred.

Documents

Optimization Techniques for Java & J2EE Applications