Architecture question

Similar Questions

• OBIEE Architecture Question?

  Hello
  
  I am brand new to OBIEE and was curious about the optimal architecture. I guess between the database and middle-tier on different physical servers is the preferred approach?
  
  Thanks in advance!

  Hi Malba,
  You can check http://www.oracle.com/us/obiee-11g-technical-overview-078853.pdf for the OBIEE 11 g architecture. With the separation of the DB and the application layer, I would say Yes and also try to make them in the same region to avoid too many jumps network and overload of netwrok cross-region routing performance.

  Kind regards
  Bharath

• Web Architecture Question - can you do this?

  I have an application that uses cookies and client variable for process/track user activity in a cluster server environment. Basically, a user is able to see a client at a time and each client has its own database.  I know that if you have multiple accouts email Hotmail, Yahoo, etc., you can access time in the same way.

  BUT I wonder if there is a way for the user to access several customers at the "same time", which means, a browser page customer is A, a browser page customer is B, a browser page customer is C, etc.  If the user has access to the right to all clients, logically he can open any two or more customers and its CRUD.

  Can this be achieved?

  Thank you

  There are two ways you could do this:

  1. Or you pass the CFID and CFTOKEN autour in the URL instead of the cookie and specify the instance of 'customer' in the CUSTOMER or SESSION scope. I wouldn't want to, recommend this method because it makes so much so much so much easier hijacking sessions.
  2. Keep the CFID and CFTOKEN in the cookie are they belong.  Pass around a variable in the URL specifying the instance of the client.  Then you should also encrypt the variable or pass around a hash with it. (hash and use the CFID and a secret key in the form of salt)

• Newbie Architecture question

  I'm trying to be a good Flex architect and put my script in external .as files using < mx:Script source = "asFile.as" / > to reference the file. The problem is, I'm referencing the different elements of the .as file command (for example, a function to clear form entries using something like: textInput.text = "";), however Flex Builder 2 shows errors (access of undefined property textInput) for all referenced controls.
  
  What Miss me?
  
  Al

  OK - thought of it. I divided the ACE in three files and called the relevant as each component file especially.

• DAQmx wait failed 100ms

  
  

  First a clarification on the issue of 100ms.  It's ONLY a problem with "Wait until the fact" behavior of meter output.  Due to the hardware architecture, question us the material every 100ms to see if the pulse is complete.  This does not change any of your output.

  With this in mind, you should NOT see any problems with the AO.  AO has a dedicated timing and engine when you do a finished we ouput only data that you ask us to.  Lets take a look what your VI.  You want to ouput 10000 (sampling frequency) *.001 * 2 (given that you add two waveforms) samples.  This is equivalent to 20 samples, or 2 ms of data.  However, you are not writing the number of samples for your VI calendar configuration.  The default value is 1000, or 980 more samples, or 98 ms.

  Just write the number of samples to your calendar VI and all should be good.

• Question of architecture / chained loops

  Hey guys,.

  So, I got a quick question on the loops of producer/consumer.

  I have a DAQ loop which I am trying to control - i.e. when to acquire data or when to go to the "idle" mode.

  At some point I want to acquire data at all times, while at another time, I want to stop and go to the "idle" mode.

  I was wondering what is the best way to to do.

  My outer loop should he send the Canadian signal and period of time later (1 or 2 seconds) will send a signal to stop, again, another signal of departure.

  1. use a notifier that sends a Notification of Cluster (option + command)--> containing some settings I need to acquire-the command 'acquire '.

  2. enter this notifier with Notification of Signal waiting

  --> This to enter the State to acquire

  3 return the same identical notification making a loop.

  OR

  1. use a notifier that sends a Notification of Cluster (option + command)--> containing some settings I need to acquire-the command 'acquire '.

  2. observe notifications with a function of State notifying get. When the declaring State changes to acquire, it will always acquire and where another notification is sent (stop) then the State will change.

  The other seems like the obvious solution, but I wonder if there are any reservations-c.-d. race conditions general/warnings warning do not use the service to get notification...

  Am not too familiar with some of the strange scenarios and me have already been screwed by Notifications jump some orders before - it was boring hard haha

  Or actually... If you have any other suggestions on how to make this architecture works, I'm all ears.

  
  

• Question/design architecture with best practices?

  Question/design architecture with best practices?
  
  Should I have separate Web server, weblogic application and IAM?
  If yes that how this time will communicate, for example can I webgate on the server that will communicate to each other?
  Any reference to decide how to design and if I have weblogic separate one for enforcement and one for IAM that how session management will take place etc.
  
  How is general design happens IAM project?
  
  Help appreciated.

  I have a business web application deployed on weblogic1 and OHS1

  deploy webtier on this OHS1 to protect the resources on OHS1. webtier will be recorded at OAM (where will be able to communicate with OAM)

  I also IAM deployed on weblogic2 & OHS2

  IAM components have weblogic. then ok. But you don't need to ohs2 here, unless your application itself requires OHS for hosting. But here, you mention that it is fair to IAM - so no need to OSH here. IAM installing requires weblogic.

  wanted to know where should I create provider (weblogic1 or weblogic2), and install webgate (OHS1 or OHS2 or both)?

  I hope it is clear now. If this isn't the case, query post for what is not clear.

• Question of Architecture cluster

  Hi all
  
  I was curious to see what you guys think of this architecture.
  
  We have about 6 critical java applications running under Weblogic.
  
  We have divided them into three managed servers, two applications per managed server.
  
  We are now clustering, initially, I created a group on two physical computers. Each physical Machine has three managed servers for a total of 6 managed servers. Then, I have deployed applications, not for each server in the cluster managed, but for two managed servers, each on a single server.
  
  Example:
  Application 1 - ManagedServer 1 (Unix Machine 1) and Managed Server 4 (Unix Machine 2).
  2 - ManagedServer application 2 (Unix) 1 and Managed Server 5 Machine (Unix Machine 2).
  Application 3 - ManagedServer 2 (Unix) 1 and Managed Server 6 Machine (Unix Machine 2).
  
  Is this type of architecture very well, or I'm better off the coast of creating three groups instead of a big cluster?
  
  Thanks for the info!

  Hello

  See the example of your architecture it seems that applications are deployed on two physical servers, then high availability is ensured, which is one of the main goals of the use of clusters.

  Regarding your question about which is better between having 3 groups or a big cluster, which depends on your architecture. The following document can be useful for you: http://docs.oracle.com/cd/E11035_01/wls100/cluster/planning.html

  Also, I suggest you take a look at the Weblogic Clustering best practices document available at http://docs.oracle.com/cd/E13222_01/wls/docs92/cluster/best.html

  Thank you
  Cree

• Portal Oracle WebLogic with OAM - Question of Architecture

  Hi all
  
  I have the following customer need.
  
  Oracle Portal 11.1.1.6 that is deployed on WebLogic Server 10.3.6
  11.1.1.5.0 Oracle Access Manager
  Oracle Internet Directory 11.1.1.6.0
  
  The customer wants to have Single Sign On for the Oracle Portal Application. It's my understanding of the Architecture.
  
  Oracle HTTP Server (with WebLogic plugin Proxy) converses with Oracle Portal
  Oracle HTTP Server (ditto above with WebGate) talking to Oracle Access Manager
  Oracle HTTP Server has Oracle Web Cache
  Oracle Access Manager talking about OID
  Oracle Portal talk to with authenticator OID OID
  
  When I went through the Oracle Portal documents, they provide steps for Oracle Portal that is deployed on Oracle Application Server, but not with WebLogic.
  
  My questions
  
  (1) do I configure authenticator OID to myrealm of the WebLogic domain (which hosts the Oracle Portal) to connect to OID?
  (2) at - it something more, that I have to do to integrate Oracle Portal with OID (meaning running pl/sql scripts or one)?
  (3) kindly advice me on the stream Cache Web Oracle, Oracle HTTP Server and Oracle Web Gate to the OAS Portal Oracle?
  
  Thank you
  
  Kind regards
  
  Somerset

  OK - portal will always OID as prerequisite.

  Thank you
  EJ

• Java card confusing questions! Java Card architecture, advantage!

  ISSUE - 1

  As I know about architecture of any application is -
                  //normal java software                          java card                                  java card
                    //  1                                               2                                      3
  1-            application program                        host application                           host application
  2-                     JVM                                   Applets                                       Applets 
  3-              operating system                JCRE(which contain all classes, JVM ....)                 JCRE(act as a O.S too)
  4                    hardware                              Operating system                                 H/W
  5                                                                        hardware

  What you think about case 2 is right? or 3!
  I think that - case 3 is right!
  I studied from several sources, the JCRE is the complete package - it behaves like an operating system and use the functionality of the component it contains. I also think that if I buy a java from any vendor card, he would have JCRE inside (means everthing inside), I only need to develop the applet according to my requirement and install it by card setup
  
  
  QUESTION 2 - What is the advantage of java card. ?
  

  // I think.
  as i read out , i got the line that it is platform independent and support multiple application. applicaton point of
   view , I am agree but how it is platfrom independent. if we use java card for smart card development, then 
   i will purchase Java card from differnt vendor , all will provide me card with supported jcre inside it, so why
   this is advantage here?
  // in case of window based application , it really make sense  because if i will make any window application with java
  // i can run it in windows , linux , unix and most of os flavours, here i can see the beauty of java language.
   Clear me this and add some other java card advantages, which make java card good

  question 3-
  

  Now I am capaple to make simple applets, so I want to check the entire java card process atleast once before 
  going in deep of something, I am thinking to implement prepaid card concept , i made it applet for it ,it will 
  contain a page with 2 text box ( 1- for recharge 2- for money deduction ) and two button , with button click i 
  want to select the applet and want to pass the apdu.
   ( a) how i can make host window application ? (i think awt will work for me? , *
  i never make any java window application yet) * 
   ( b) still i dont have any java card , i want to make it by using simulator, is it possible?
    (c) in this scenario what framework i need to study Ex. Import javacard.framework.* for making
   applet or from where i need to start.

  With regard to:
  Rohit pathak
  
  Published by: rohit pathak on 9 February 2012 22:10

  Hello

  so why should I use applet concept here OR WE can't USE APPLET CONCEPT THERE! We will go to c code to implement...

  Yes, you can also write cmdlets in C. But in this case you also need mechanism by you that you will be able to convert your code in the .cap file and install it on the map.

  so I think that a card that does contain no JCRE, should not be called a java card. RIGHT?

  Yes, he called card native.

  or if a card is a java card must contain jcre and I need to write only applet here. RIGHT?

  Any card that can be programmed in a java programming language is called a java card. And to execute/interpret java byte code on the card, you will need something like JCRE/JVM on it.

  s This also possible case that a map is a java card, which contains no JCRE (jcre also contain JVM) java applet concept but is there? I think that it is impossible

  Already answered above.

  a card that contains your own OS (written in c) does not support the concept of the applet, here you must omit concept applet and should just your own program to manage the apdu. right?

  Either way, applet is also a program and when card java, it supports only a subset of the standard programming lauguage.
  Yes, you can name what you want :)

  Java card cocept is costly because here you use other OS and OS MEANS EVERYTHING, RIGHT so U NEED HIGH pay to OTHERS a (OWNER JAVA)?

  Are the manufacturer must also bear the cost of the license also. But in my view, it is inexpensive because of its more services and features.

  Concerning
  Umer

• Question of architecture including SAY, refinery incoming and to PDF Converter

  Hello
  
  I define a new process for our application of the Complutense University of MADRID.
  
  The process is the following:
  
  -L' user creates a new file (for example with Word). Using TELL it, he saves him from the AAU.
  -The only file format that we want in the Complutense University of MADRID is the PDF format.
  -Then, when the document is inserted in the Complutense University of MADRID, it is automatically converted into PDF format.
  
  So, to realize that I think architecture to help:
  
  -SAY, of course
  -The inbound refinery
  -The PDF Converter.
  
  As the UNIX operating system, I cannot use it, if I understood correctly, OpenOffice to PDF Converter.
  
  My question is whether it is the right approach to achieve this functionality.
  If someone already has can the opportunity to implement this kind of concept, please you share your experience with me?
  
  Any help is welcome
  
  Concerning
  
  Stone

  Pouvez be able to use the builtin "OutsideIn" technology to create a PDF file but if you want a good quality output and you use a * NIX system for inbound refinery and then use OpenOffice

  By default, the search results link to PDF version (visible web) of the content. The content is available in format doc via GET_FILE service there is a link to this on the content information page.

  I think that you can restrict access to the version native between now & for guest users 'anonymous' easily. It would be simple hide the link but if it is extremely important that users can not access the doc file then I think you need to create a java custom filter in order to intercept GET_FILE service calls.

  You can see the example of the security in the HowToComponents.

  Tim

• Question of the Master Page Architecture

  New site needs a main page of the company, which then directs the customer to choose one of the two store restaurant locations which each will have its own pages with its options, somewhat different from the other chain store location.  What architecture will better serve this kind of situation?

  The name of your master pages are not relevant. Only, you see these names. You will need to set the page name in the page properties.

  I don't know what you mean by levels? On the plan view? If so, I would have probably a three-page senior level (hand, store 1 and 2 to store) and then direction from there. But it's really hard to say without knowing how many pages you have etc.

• Newbie question on a client-server application architecture

  I did some stuff in Java for awhile now, but I have never drawn in the network part of this one. I followed the tutorial to Oracle and I searched some other resources on the web. I don't have a very specific question, but I would be grateful if someone could help me understand all this. My goal in this learning project is to make a server application that will accept an unspecified number of independent of a client application connections. I think that I get the basics of establishing a ServerSocket on the server, put in place the socket on the client, and then open the entry and outflow on both and start a data transfer.
  
  Now, I think I'm stuck in a misunderstanding on this transfer. All of the examples I see seem to treat data transfers when the order who speaks on each side is already known, so you send something good after connection and that you set the other side of eligibility receive something after the connection, and everything works fine. If the order is not known, but I can't seem to understand the flow of the application logic.
  
  Let's say that the server gets 3 connections of 3 customers and save these Socket objects in an ArrayList. Then client B asks something, and the server responds accordingly. Later, the server needs to say something to client B, then it sends information through its support and client B responds accordingly. How should I go about setting up something like that? I think I should put listeners but I can't find examples like these, or do I define the infinite loops for each socket, try to read the data?
  
  Any help to guide me in the right direction would be much appreciated.

  Think of a very basic chat client-server application. When a client connects, it receives all messages written by other customers. It can also send messages that will go to all other customers. It is very similar to your uncertainty about who will speak in the first place, the server or the client. If users wait a bit after the connection, it is likely, that the server will talk first (because another user sent a message). If the customer/user is fast, he'll talk first ("Hello World" or something like that). There are examples of these applications of simple chat on the web, just google for them. And Yes, this is usually achieved with infinite loops. Every customer will have a such loop, try to read data from the server and the server will have as many endless loops simultaneously assets (think multi-threading) given that the connected clients.

• Question of Architecture component GoldenGate

  Reading on the architecture of the GG, I see only discussion on the components that are on the source systems target.
  What GG components reside in the FMW?
  
  Thank you
  -Art

  There is no benefit to running a GoldenGate instance on a remote server. In doing so, you introduced a point of failure (transport network to the remote server). If you lose connectivity, where are the trails? The only potential benefits are linked to the constraint of memory on the server source database, where a maxed out using 300 process instance takes about 15 GB of memory, and, possibly, a lack of disk space for the trails. Although possible, it is not probably when you consider how much it costs to license GG. If you can allow to authorize the GG, being short of RAM disk space is probably not on your list of things to worry about.

• Question of design architecture: layers of components

  I am so confused about how combine components and when I should inherit, and when things go into skins. I have a custom component tile that looks like this:

  

  Source (Simplified) looks like this:

  "< s:SkinnableContainer skinClass = 'skins. TileSkin""">

  < fx:Script >

            <! [CDATA]

  [Bindable] public var tileColor:uint = 0xFF0000;       / / Color of base tile.  The central area will be this color.

  [Bindable] public var tileThickness:uint = 4;            / / How high the tile seems to have his Center Party raised.

            ]] >

  < / fx:Script >

  < / s:SkinnableContainer >

  All drawing is found in the skin.  I have a subclass this component to have a LabeledTile:

  

  Source (Simplified) looks like this:

  < components: tile >

  < fx:Script >

            <! [CDATA]

  protected static const int PADDING = 3;

  public function get text1 (): String { return Label1.text ;}

  public function set text1 (str:String):void {Label1.text = str ;}

            ]] >

  < / fx:Script >

  " < s:Label id ="Label1"x =" tileThickness{}"y =" tileThickness{}"

  {" height =" {}height - 2 * tileThickness}"width =" {} {width - 2 * tileThickness}"

  "" " maxDisplayedLines ="1"textAlign ="center"verticalAlign ="middle"

  {' {' paddingLeft = "{}PADDING}" paddingTop ="{} {}PADDING" paddingRight = "{}PADDING}" paddingBottom ="{} {PADDING}"/ >

  < / components: tile >

  I want to have a subclass of LabeledTile called LabeledTileWithGizmo that looks like this:

  

  I thought this might work:

  < components: "" LabeledTile width ="130" height ="32" tileColor = "0x0077EE"

  contentCreationComplete = "ContentCreated ()"> "

  < fx:Script >

            <! [CDATA]

  private function ContentCreated():void

                 {

  Label1.setStyle ( "fontSize", 14);

  Label1.setStyle ( "color", 0xFFFFFF);

  Label1.setStyle ( 'fontFamily', '"Trebuchet MS" BOLD, Arial, Helvetica, _without' );

                 }

            ]] >

  < / fx:Script >

  "" < components: Gizmo x = "100" y ="4" height = "24"/ > "

  < / components:LabeledTile>

  The gizmo appears on the tile, but the label disappears.  If I put the component Gizmo just next to the Label in the LabeldedTile element, they both get drawn.  But I want to have labeled the tiles that have no gadgets, as well as those who do.  I also have different types of gadgets.

  So, should I make a subclass of tiles with optional subcomponents?  And how could I do?

  Should I put more than this in TileSkin?  And the other components are States in the skin?

  I should have a different skin for each tile, LabeledTile and LabeledTileWithGizmo?  The second skin must inherit from the first skin and the third to the second?

  Can I get a skin for a component to set styles on a subcomponent (for example to have a LabeledTile set the fontSize, etc on the subcomponent Label)?  Or is it still possible.

  I am so confused on how all these elements must adjust.  Any idea would be appreciated.

  Your MXML for LabeledTile has Label1 as a child element in MXML.  When you subclass that class using MXML)

  Looks like what you want to do is subclass of SkinnableContainer to have a label, part of the skin and move your label in the skin.  In this way, when you add child elements to your container it won't replace the label with your child items.

  This article is a good start to learn more about the counting of spark: http://www.adobe.com/devnet/flex/articles/flex4_skinning.html

  If you follow the process above, so here are some answers to your specific questions:

  So, should I make a subclass of tiles with optional subcomponents?  And how could I do?

  > You could label an optional part of the skin, so if someone does not want the label will then appear, they would create a custom look that does not include it.  An alternative approach is to expose a property on your component showLabel who would control the visibility of the part of the Label.

  Should I put more than this in TileSkin?  And the other components are States in the skin?

  > Yes move the label on the skin.  You could use States or expose a showLabel property as I mentioned above.

  I should have a different skin for each tile, LabeledTile and LabeledTileWithGizmo?  The second skin must inherit from the first skin and the third to the second?

  > Looks like you could do all this with a single component and a skin by adding another part of skin customized for the gizmo. Legacy via MXML skins is not trivial to implement.

  Can I get a skin for a component to set styles on a subcomponent (for example to have a LabeledTile set the fontSize, etc on the subcomponent Label)?  Or is it still possible.

  > Yes, you should be able to do this, simply call setStyle on the part of the skin