Fundamental questions on Object pooling

Hi,
I ve a class which exposes only utility methods and objects of it wont have any state at any point of time. Now which pattern should I use here...1. Can I go for Singleton pattern so that all calling classes will work on same instance in one JVM? 2. Each calling class will create one instance use it and nullify it? 3. Expose all utility methods as static methods so no physical instance in JVM as no state is needed.?
Would having only one instance degrade the performance or boost in case of having a new instance everytime?
Please share ur views on this. Pros and Cons on each pattern.
Thanks in advance
JoyBoy

Hey Chris,
For transporting objects either in IR / ID prerequisite is that you must have already configured your XI Track properly using landscape configurator as it defines the transport route for the corresponding XI objects based on the s/w component version.
As for ID doesn't have any s/w component version we define SAPINTDIR 3.0 as s/w comp. It refers to all the objects in the ID.

FOR Integration Directory transport Test(Consolidation)->Prod:
When the objects are transported from test to productive landsape unless any changes have been made to the objects in the test landscape all the objects can be transported to the id with out any complexity. CMS takes cares of conversion of the business system to that of prod if required. Usual convention that is followed is to maintain same naming schema for the objects.
cheers.
Raj.

Similar Messages

Fundamental question: disable the GUI while RMI is communicating?

Good afternoon,
I'm using RMI in a swing application and i have fundamental question:
I have a jlist with categories and another jlist which displays the items of the selected category. When a user selects a category rmi is used to get the items of the selected category from the server (internet).
Should i disable the whole application while rmi is communicating with the server? For example with a glasspane with a cancel button.
What is the best practice in this case?
Best regards,
James

I found a really nice solution, now i call my remote methods like:
taskManager.executeAction(new ITaskResult(){
      public void done(Object result) {
        Collections.sort(rpcs, new RpcComperator());
        fireContentsChanged(this, 0, getSize());
      public void failed(Exception e) {
        WindowUtils.showErrorDialog(owner, "Could not change server");
    }, new IAction(){
      public Object action() throws DBException, AccessException {
        return taskManager.getBasic().changeRpc(rpc);
    }, true);The first parameter of executeAction() is a object which gets notified when a task completes (done()) or fails (failed()). The second parameter is a object which indicates which remote functions should be called. And finally the third parameter defines if the user should wait until the remote function completes.
If the user should wait the executeAction() function shows a nice modal dialog with a progressbar and a cancel button. If the user doesn't have to wait it takes the progressbar on the statusbar to indicate the system is busy.
Much rewriting was needed, but i think it is quite simple and it is fully swing thread save. But it wonders me that such a thing wasn't available.

Fundamental questions before starting next project

Last project I put together received scathing criticism - and I have to say that I did not put any effort at all to clean the code up or optimize it. The reason is I started with small projects and within couple of months it just grew out of hand with requirements matrix gradually increasing.
So this time around I want to take advice from you guys on architecture issues before I jump into programming. The two fundamental questions:
1) I always thought that SubVi is like a method in C++. If I need to use that code multiple times, then I better put it in a SubVi. But after going through some of the posts on this forum it seems that people use SubVis to reduce # of wires,blocks,etc. on the main Vi too. Is this a standard practice?
2) If I want to have more than 1 VI in the software, what is the best way to send data from one VI to the other VI? SubVI has input and output wires. So it is easy. But VIs do not. Basically I want to have Main VI where I get all the information from the use through a front panel and then go off doing my data acquisition. I want to keep the Main VI as clean as possible. There will be some operations which will need only 1 input and will have only 1 output. But this operation will be used only once in the who operation. So can I have this operation all in a VI and then call it in the Main VI?
3) If I have 3 different VIs in my application, and I define something (say an array) in the Main VI or for that matter second_VI.vi in my application, will that be available in the third_VI.vi ?
Thanks in advance.

labview_beginner wrote:
1) I always thought that SubVi is like a method in C++. If I need to use that code multiple times, then I better put it in a SubVi. But after going through some of the posts on this forum it seems that people use SubVis to reduce # of wires,blocks,etc. on the main Vi too. Is this a standard practice?
You could call it similar to a method. I would say a subroutine or function is a better terminology since the term method is more applicable to object oriented based programming. Basically any way to encapsulate a portion of code, particularly if you use it over and over. It might have input and/or outputs, or neither.
2) If I want to have more than 1 VI in the software, what is the best way to send data from one VI to the other VI? SubVI has input and output wires. So it is easy. But VIs do not. Basically I want to have Main VI where I get all the information from the use through a front panel and then go off doing my data acquisition. I want to keep the Main VI as clean as possible. There will be some operations which will need only 1 input and will have only 1 output. But this operation will be used only once in the who operation. So can I have this operation all in a VI and then call it in the Main VI?
There is no difference between a VI and a subVI other than semantics. Both are completely independent excecutable portions of code. Your main or top-level VI can have inputs and outputs as well. You could execute it stand alone, or it could become a subVI of an even higher portion of code in your hierarchy. The only difference between the two is that a subVI is just a term to distinguish it as being a portion of a higher level of code. There are multiple ways of passing data between VI's. One is wires. When a subVI ends, the data on its indicator that is connected to the panel is available to travel down the wire to another portion of code. Another is global or shared variables which stores data to be used by multiple locations. Another is a function global variable, a LV2 style variable, or action engine. All of these use special characteristics of uninitialized shift registers to store data between calls of that subVI. Other methods of sharing data are on the Synchronization pallette such as queues and notifiers to store and pass data.
Using a subVI, even it used in only 1 location can help keep a portion of your block diagram clear. It gives you the ability to hide some details such as a complicated algorithm. You can write an algorithm as a subVI and it gives you the ability to test the inputs and outpus as a standalone. Once you have that. You can drop it in and use it even if it is only at 1 location. If you need to change it, you can modify the subVI and test it without risking the corruption of the rest of your main VI
3) If I have 3 different VIs in my application, and I define something (say an array) in the Main VI or for that matter second_VI.vi in my application, will that be available in the third_VI.vi ?
Possibly, it all depends on which of the methods listed in number 2 you would like to use.
Thanks in advance.

Fundamental question on EJB, JNDI and client jars

Hi,
This is a very fundamental question on EJBs and their clients - what
all should go into the client jar of an ejb?
I know for sure that just the remote and home interface classes of the
ejb are sufficient on the client's classpath to work with an EJB on a
totally different server, but I dont understand the logic behind it.
If the client has to pass its object parameters over the network to
the server where the ejb bean is located, should the container
generated stub not be present on the client's classpath? After the
client does the JNDI lookup of the ejb home on the server, how does it
serialize and pass the parameters over the network if the container
generated stub is not present?
Any help would be greatly appreciated. Pointers to material on the
internet which explain this/related things in detail would be a great
help.
Thank you,
Anoushka

hello,
well, the process is fairly simple actually.. The client needn't necessarily have
the container generated stub. as a client u could contact what is known as a boot-strap
service to download the stub over the wire.. infact one of the advantages of RMI
is precisely that. and since RMI is the underlying architecture of EJBs, it all
becomes rather simple. when u 'lookup' a bean, what u are in essence doing is
asking the server to send down the stub to your JVM. well.. right now this is
what i got time for..i will try to post a lengthier explanation in a couple of
days at leisure..
Vijay
"Anoushka" <[email protected]> wrote:
>
Hi,
This is a very fundamental question on EJBs and their clients - what
all should go into the client jar of an ejb?
I know for sure that just the remote and home interface classes of the
ejb are sufficient on the client's classpath to work with an EJB on a
totally different server, but I dont understand the logic behind it.
If the client has to pass its object parameters over the network to
the server where the ejb bean is located, should the container
generated stub not be present on the client's classpath? After the
client does the JNDI lookup of the ejb home on the server, how does it
serialize and pass the parameters over the network if the container
generated stub is not present?
Any help would be greatly appreciated. Pointers to material on the
internet which explain this/related things in detail would be a great
help.
Thank you,
Anoushka

A question about Object Class

I got a question about Object class in AS3 recently.
I typed some testing codes as following:
var cls:Class = Object;
var cst:* = Object.prototype.constructor;
trace( cls === cst); // true
so cls & cst are the same thing ( the Object ).
var obj:Object = new Object();
var cst2:* = obj.constructor.constructor;
var cst3:* = obj.constructor.constructor.c.constructor;
var cst5:* = Object.prototype.constructoronstructor;
var cst4:* = Object.prototype.constructor.constructor.constructor;
var cst6:* = cls.constructor;
trace(cst2 === cst3 && cst3 === cst4 && cst4 === cst5 && cst5 === cst6); //true
trace( cst == cst2) // false
I debugged into these codes and found that cst & cst2 had the same content but not the same object,
so why cst & cst2 don't point to the same object?
Also, I can create an object by
" var obj:Object = new cst();"
but
" var obj:Object = new cst2();"
throws an exception said that cst2 is not a constructor.
Anyone can help? many thanks!

I used "describeType" and found that "cst2" is actually "Class" class.
So,
trace(cst2 === Class); // true
That's what I want to know, Thank you.

I have an Adobe Indesign to Photoshop question - Smart Objects

I have an Adobe Indesign to Photoshop question - Smart Objects
I am copying and placing InDesign photo silhouettes and placing them into a Photoshop document for an online graphic.
I believe that the color holds up very well. It's when the JPG goes through a CMS ingest that the colors warp. A photo technician is able to correct the color.
I just need to get some insight on my process that (I think) I invented - since I don't see any discussions or videos online that back me up.
Any thoughts - or who could I access that might have an answer for me.
Thanks!

You haven't invented anything. Your server-side processing (based on something involving ImageMagik with PHP most likely) is simply stripping/ ignoring color profiles and converting the colors. This is very much the same issue as in your other post - your procedures just don't realyl make sense and you're basicalyl screwing yourself by using this methodology.
Mylenium

Object pool for DTO objects

Hi,
I am using BlazeDS to (de)serialize many identical small objects between a Flex client and a Java Servlet. For efficiency, I would like to create object pools on both sides. Is this possible with BlazeDS?
Thank you,
Andrej van der Zee

Hi,
Thanks for your reply.
Are there any thoughts how I can do this on the client? As far as I can see, deserialization of DTOs is hidden in the remote object. How can I tell a remote object to use a factory method for creating certain DTOs?
Cheers,
Andrej
On Sat, Jan 17, 2009 at 3:37 AM, Alex Glosband
<
[email protected]> wrote:
A new message was posted by Alex Glosband in
General Discussion --
Object pool for DTO objects
Hi Andrej,
The answer is no. There aren't any object pooling features in BlazeDS for pooling DTOs, either on the client or the server. If there are any third party object pooling solutions for DTOs on the server though you could maybe hook one of them in. . . either in your Java servlet or a RemoteObject. . . but I think anything on the client you would have to write yourself.
-Alex
View/reply at
Object pool for DTO objects
Replies by email are OK.
Use the
unsubscribe form to cancel your email subscription.
Andrej van der Zee
2-40-19 Koenji-minami
Suginami-ku, Tokyo
166-0003 JAPAN
Mobile: +81-(0)80-65251092
Phone/Fax: +81-(0)3-3318-3155

Is object pooling still necessary?

I’ve have implemented a transport server in Java, which creates many objects of the same type. In consideration of the compiler improvements does it make sense to use object pools like http://jakarta.apache.org/commons/pool/ nowadays?
What do you thing?
Thx,
Chris

Or only very rarely. Only if you have an object that is particularly
costly to create, e.g. a state machine for a parsing or
cryptographic engineIt is quite interesting though. I built a parser recently, and pooled the state-machines. However, I did all allocation for buffers (byte[]) on demand.
Performance wasn't great - so I created an allocator interface - and put my HeapAllocator head-to-head with a PooledAllocator.
Bang: 35% performance improvement with pooling.

Java object pooling

Hi all,
Is there an Object pooling mechanism available in J2SE?
Does anyone know of any good open source object pooling implementation?
Thanks in advance

Unfortunately, the Jakarta's object pool is very ineffective for objects requiring a long construction time.
The pool can grow dynamically (when no idle object available for allocation and pool is not exhausted yet) and new objects might be constructed inside borrow method. Since the borrow as well as the freeObject are both synchronized (on the pool instance' lock), all other waiting clients will have to wait for that lock to be released before they can either release used object or borrow one.
Bottom line, you better not use this pooling API for pooling objects such as connections which have a very long construction time.
I built a very efficient object pool and will be glad to share it with others.
Just tell me where can i post it

Poolman Object Pooling

Hello all,
I am having problems with CodeStudio's PoolMan. I need to create a pool of custom objects, but it keeps on saying that my object can not be found. I can use the exact same code and use java.lang.String as the pool type and it works great... any ideas would be great!
Here is the log file from my poolman startup...
October 3, 2001 1:26:30 PM EST: Creating 20 initial objects in pool UserChannelPool
October 3, 2001 1:26:30 PM EST: GenericPool: Exception while initializing
EXCEPTION: edu.iu.uis.my.web.UserChannel: java.lang.ClassNotFoundException: edu.iu.uis.my.web.UserChannel
java.lang.ClassNotFoundException: edu.iu.uis.my.web.UserChannel
 at java.net.URLClassLoader$1.run(URLClassLoader.java:200)
 at java.security.AccessController.doPrivileged(Native Method)
 at java.net.URLClassLoader.findClass(URLClassLoader.java:188)
 at java.lang.ClassLoader.loadClass(ClassLoader.java:297)
 at java.lang.ClassLoader.loadClass(ClassLoader.java:253)
 at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:313)
 at java.lang.Class.forName0(Native Method)
 at java.lang.Class.forName(Class.java:120)
 at com.codestudio.util.GenericPool.create(GenericPool.java:110)
 at com.codestudio.util.ObjectPool.checkOut(ObjectPool.java:214)
 at com.codestudio.util.ObjectPool.init(ObjectPool.java:94)
 at com.codestudio.util.GenericPool.<init>(GenericPool.java:47)
 at com.codestudio.util.GenericPoolManager.createPool(GenericPoolManager.java:54)
 at com.codestudio.management.LocalPoolDeployer.startGenericPools(LocalPoolDeployer.java:137)
 at com.codestudio.management.LocalPoolDeployer.deployConfiguration(LocalPoolDeployer.java:31)
 at com.codestudio.management.PoolManBootstrap.<init>(PoolManBootstrap.java:72)
 at com.codestudio.util.GenericPoolManager.assertLoaded(GenericPoolManager.java:43)
 at com.codestudio.util.GenericPoolManager.requestObject(GenericPoolManager.java:82)
 at _0002findex_0002ejspindex_jsp_11._jspService(_0002findex_0002ejspindex_jsp_11.java:87)
 at org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:119)
 at javax.servlet.http.HttpServlet.service(HttpServlet.java:853)
 at org.apache.jasper.servlet.JspServlet$JspServletWrapper.service(JspServlet.java:177)
 at org.apache.jasper.servlet.JspServlet.serviceJspFile(JspServlet.java:318)
 at org.apache.jasper.servlet.JspServlet.service(JspServlet.java:391)
 at javax.servlet.http.HttpServlet.service(HttpServlet.java:853)
 at org.apache.tomcat.core.ServletWrapper.doService(ServletWrapper.java:404)
 at org.apache.tomcat.core.Handler.service(Handler.java:286)
 at org.apache.tomcat.core.ServletWrapper.service(ServletWrapper.java:372)
 at org.apache.tomcat.core.ContextManager.internalService(ContextManager.java:797)
 at org.apache.tomcat.core.ContextManager.service(ContextManager.java:743)
 at org.apache.tomcat.service.http.HttpConnectionHandler.processConnection(HttpConnectionHandler.java:210)
 at org.apache.tomcat.service.TcpWorkerThread.runIt(PoolTcpEndpoint.java:416)
 at org.apache.tomcat.util.ThreadPool$ControlRunnable.run(ThreadPool.java:498)
 at java.lang.Thread.run(Thread.java:484)
October 3, 2001 1:26:30 PM EST: PoolMan Local Pool Deployer: Created Object Pool Named: UserChannelPool
Thanks!
Nate

Never assume always check A lesson I had drummed into me as a youngster. In this case your assumption is wrong, this exeption is always thrown if the class stated is not in the classpath. This particular classpath is Tomcats, not your client app. Drop the jar with the class in $TOMCAT_HOME$/webapps/[your root]/WEB-INF/lib.

Questions regarding Connection Pooling

Hi
i have some confusion regarding how connection pooling is implemented..i have been doing some digging in to this lately but the more i read the more confused i get. i have some doubts here, will any of the posters here kindly clear them
a. Do i have to connect to my database via the DataSource method to take advantage of the connection pooling provided by my app server?..i have read conflicting information about this suggesting that Connection Pooling will not be provided to my app if i use DriverManager.getConnection()
b. if i chose to use the DataSource method do i have to implement the ConnectionPoolDataSource interface to provide the connection pooling for my app?
c. what is the best way to implement my own custom Connection pool?
Thanx

DriverManager.getConnection() literally creates, that is, builds a connection to the database using the values you previously supplied when you loaded the driver.
A connection pool is an object that contains several already made connections to the database, and simply provides you with one of those existing connections. The confusion arises because a) the method name is frequently the same and b) the connection pool object calls the driver's getConnection() method to create several connections before lending any of them out.
In other words:
DriverManager.getConnection() builds a connection to the database.
ConnectionPool.getConnection() fetches an existing connection.
So, to answer your questions...
A. This is correct. If you use DriverManager.getConnection(), you are indeed bypassing the Connection Pool entirely.
B. I'm not familiar with DataSource specifically, but in general, a third party connection pool will either give you the interface (and you implement the details) or they will give you the class file containing the DataSource object. All you would have to do in the latter case is to import that file, then create a new instance of it using the new keyword ( DataSource foo = new DataSource(); ). I suspect DataSource is indeed the class file.
C. Creating a connection pool is trivial and there are many examples of it - search on this forum or check out your favorite Java/JDBC programming book. Usually, the question "Should I use a connection pool and why?" is a more important question.

Questions about database pooling....

Hi everyone, I'm not new to java programming but I am looking to polish up my skills. I have read sooo many articles and am still confused so please be patient with me. Java has a way of adding new acronyms everytime I turn around - its very hard to keep up with...
I am using Netbeans 5 with Tomcat 5.5.9 bundled.
I am connecting to a postgresql database (version 8.1).
I would like to implement some type of jdbc connection pooling. Everything I have read, even on the tomcat website is confusing and seems to be a "best effort" type of situation - none gauranteeing correctness or reliability or replicability between what others may experience and what I might.
I looked into a pure java solution such as DbConnectionBroker but I get confused as to how this "pooling" is effected by servlet lifecycles. I have read and read and still do not understand the necessary concepts.
If I have a login servlet that makes a single query to the database, how long does that instance of the servlet stay loaded into memory? If it is dropped out of memory right away it seems there is no benefit of using the connectionBroker? (since each time someone visits the login servlet it would have to recreate the login servlet and broker object??)
Is the broker object kept in memory somewhere with connections open after the servlet stops executing? How can other servlets access that same broker object without creating a new one of their own??
I guess it comes down to really not understanding the persistence of how things are loaded and kept in memory. If this is a question for the Tomcat people I do apologize :/ It just seems these forums are by far and away the best place for any type of java help :)

A servlet has nothing to do with it.
A servlet uses a pooled instance. It is the pooled instance, not the servlet which is cached. At least in terms of discussing db connection pools.
A cache means that it remains in memory. If the cache isn't kept in memory then keeping a cache is pointless.
I guess it comes down to really not understanding
the persistence of how things are loaded and kept in memory.More of a basic java question than anything else.
The GC will collect reference which are no longer active sometime after they become not active. Thus a local reference can be collected after the method exits. But a static will not be collected until the class is (not instances of the class, but the class itself.) That won't happen until the class loader that loaded the class is collected.

A question about objects and queries

I just started using COM objects in ColdFusion. I am mostly
using CreateObject, but occasionaly use <CFOBJECT>
My question is; Do I have to destroy these object variables
when I am done? I know in a lot of other languages I work in, you
have to "clean up'" after yourself and destroy some variables when
you are done to avoid memory leaks. How does CF handle this?
While I am asking, what about queries? Do hey have to be
destroyed as well? What about queries made with QueryNew?
Thank you for the help.

Cold fusion cleans up for you.
You do not need to explicitly destroy anything unless threads
are spawned in a non standard way by Java or C code that you are
using.

Solution Manager - v3.2 - Fundamental Question

Hi
I have a fundamental SAP / SM question which I cant seem to actually clarify....
Do ALL SAP servers in a SAP Infrastructure have to have a servername of 13 Characters?...
When revieing the OSS it mentioned HOSTNAME...and it seems to use this interchangably with Computer Name...
So, for example...
Can I have a Solutions Manager server with a Netbios name of "SOLUTIONSMANAGE" - 15 CHARACTERS and a SAP ECC Server with a server name also with 15 Characters......?
The OSS and all the documentation refers to SAP Prereq of 13 Characters for the Host...but what is the host defininition in this case?..What does the 13 Character limit apply to? How do companies with Naming conventions longer than 13 Characters implement the landscape?
Many Thanks!
J

Hi Oliver,
Check out these URL's
Solution manager
http://help.sap.com/saphelp_sm32/helpdata/en/index.htm
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.sdn.logon.redirect
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/library/biw/how to integrate bex web applications into sap solution manager.pdf
http://www.sap.com/services/pdf/BWP_SB_SAP_Solution_Manager.pdf
http://www.sapinfo.net/index.php4?ACTION=noframe&url=http://www.sapinfo.net/public/en/print.php4/article/Article-220903dbfa95a95bd1/en
https://service.sap.com/solutionmanager
http://www.sapinsideronline.com/spijsp/article.jsp?article_id=41229&volume_id=5577
http://service.sap.com/roasmaps
Regards,
RK

Questions on Connection Pooling

I already read the documentations on data-sources.xml, but I still have questions. In the following data-sources.xml (for OC4J 9.0.3):
<data-source
class="com.evermind.sql.ConnectionDataSource"
name="ccf"
location="jdbc/ccf"
xa-location="jdbc/xa/ccf"
ejb-location="jdbc/ccfPool"
url="jdbc:oracle:thin:@erb:1521:prod"
connection-driver="oracle.jdbc.driver.OracleDriver"
username="scott"
password="tiger"
max-connections="300"
min-connections="5"
max-connect-attempts="10"
connection-retry-interval="1"
inactivity-timeout="30"
wait-timeout="30"
/>
I know that if we want to use a connection pool, we should use "ejb-location" and it is the only one used for emulated data sources. I also know that both "location" and "xa-connection" are used for non-emulated data sources.
Then my questions are:
1) Does non-emulated data source mean that it can not used for connection pooling? In other words, both "location" and "xa-location" can not be used for connection pooling? The only one that can be used for connection pooling is "ejb-location"?
2) How can I use a non-pooled connection? right now, I am using "location" for non-connnection pooling, is this the right way to do it? The documentation says that it will be deprecated in future release, then what should I use if I want a non-pooled connection? can I still use "ejb-location" by setting "min-connections" and "max-connections" to 1? This is assuming that we can NOT use "location" to get the data source.
3) Putting the username and password in the data-sources.xml seems not be good from the security point of view. But, if I am using CMP, I must supply the username and password in the data-sources.xml. Is there another way around this?
Thank you very much for the help in advance.
Jingzhi

Liu
I may not answer all ur questions.But would like to share my experiences....
3) U can set the username and password dynamically also...we use it in orion..so u should be able to use it in oc4j..as a matter of fact, we can set all the data-sources parameters dynamically(programmatically).
1) I may not be correct 100 percent with this :but for pooling only ejb-location is used and for non pooling location is used.
For non pooled connections also, use ejb-location and dont give any max and min connections..let it take the default values...
I am not aware that Oracle is going to deprecate the location ..well it is good to hear this which will reduce some confusion..
seems there is lot being raised on these topics..
yugandhar

Fundamental questions on Object pooling

Similar Messages

Maybe you are looking for