Client Definition Question

I asked this question in another forum here, got some helpful feedback and a recommendation to post the question here as it might be more fitting for some in this group.
In getting ready to teach a class, I've been putting together my set of SAP terms and getting ready to explain them to the class. I've been cross referencing the SAP Glossary and other sources to try to give as correct a definition as possible. In looking at these, one element I've been wondering about is the SAP definition of a client within SAP ERP.
The SAP glossary defines client as "In commercial, organizational, and technical terms, a self-contained unit in an SAP system with separate master records and its own set of tables." I find the last phrase in that definition (which I've italicized) to be puzzling. In an instance where multiple clients are sharing the same underlying database, each client does not have its own set of tables. To my knowledge that's not a required element of a client. (That's why the mandt field is so common in tables.)
So...why is that phrase a part of the definition and what is SAP intending it to mean?  Is a distinction being drawn here between logical segregation of content (vs. actual segregation)? Or is there something here that I'm just not understanding?
As an example of what I'm talking about, if we have one instance of SAP ERP that has multiple activie clients, and If we were to take table T001 which contains company code information mapped to a particular client, wouldn't all the clients in that instance be sharing that same table?
I realize that logically if one uses transaction SE16 to view the table from within SAP ERP, only the information for the client one is working within is shown. My thinking had been that if one were to look at this table directly in the DBMS, the table would show all of the clients active within this instance. So in that sense each client is using the same set of tables with logical segregation of data based on the MANDT field. (i.e. if I had 3 clients running on one SAP instance tied to an Oracle backend, the backend would not have 3 totally independent databases for each client.)
Am I missing something here, or is the definition of client a bit misleading?

Hi Tony,
Pity that your question did not get a reply in more than two weeks.
I am happy to say that I have never come across the "its own set of tables" statement. Whoever wrote that is either ignorant or a victim of a very bad mistranslation. A client is a logical "subdivision" of the data stored in a SAP database. I'm not old enough to confirm this, but when I started at SAP I was told that the client concept came from the time when several companies shared a single database on some large central mainframe and obviously could not be allowed to see each other's business data. Today the use of a client is different, e.g. it provides for separation between customizing and sandbox clients in a development system, or between say test and training clients in a QA system.
Technically all clients share the same set of tables. A "client-dependent" table has the characteristic that its first field (and the leading field of its primary key) is always the client number. The "client-dependent" attribute is maintained in the SAP dictionary. When an ABAP program contains a statement like "SELECT * FROM TAB" and TAB is client-dependent, the DB interface in the kernel will automatically convert this to "SELECT * FROM TAB WHERE MANDT = "<client>". An SQL trace (ST05) of a simple business transaction will reveal this behaviour immediately.
Not all tables are client-dependent, some are cross-client, i.e. their data is visible to and shared by all clients in the system. As a rule, all master/transaction tables are client-dependent, while system tables containing instance-specific rather than business-specific data (e.g. the list of printers) are cross-client. Customizing tables are mostly client-dependent but some are cross-client (I'm a technical consultant so I can't really tell you why that is).
You are perfectly correct that if you look at a table at database level rather than through SE16, you would see records for all clients residing in the same table. Incidentally the same can be achieved in ABAP by adding the CLIENT SPECIFIED option to the SELECT statement but the code of business transactions must of course do no such thing.
Also, the client concept is not specific to Oracle so the advice to post the question on the SAP/Oracle forum seems a little misleading. SAP implements the client concept regardless of the underlying database.
Hope this clarifies things,
Mark

Similar Messages

  • Standard Client definition question

    In getting ready to teach a class I've been putting together my set of SAP terms and getting ready to explain them to the class.  I've been cross referencing the SAP Glossary and other sources to try to give as correct a definition as possible.  In looking at these one element I've been wondering about is the fairly universal definition of a client. 
    The SAP glossary defines client as "In commercial, organizational, and technical terms, a self-contained unit in an SAP system with separate master records and its own set of tables."  I find the last phrase in that definition to be puzzling.  In an instance where multiple clients are sharing the same database each client does not have its own set of tables.  To my knowledge that's not a required element of a client. (That's why the mandt field is so common in tables.)
    Is a distinction being drawn here between logical segregation of content (vs. actual segregation)?  Or is there something here that I'm just not understanding.
    Thanks.

    Tony,
    Ah, you were talking about the underlying DBMS engine, not the application.
    When SAP is talks about tables, they are talking about the data structures within the application.  Clients do not share these tables (well, there are a few client independent tables).
    I misunderstood your original question. You are talking about the underlying DBMS, which also stores data in tabular form.  Yes, the DBMS structures do store data from multiple clients.
    When SAP passes requests for action to the underlying DBMS, such as oracle, the application 'knows' where the data is stored, and uses the keys generated within the app to perform reads and writes on the databases.  
    I am not a Basis expert, so I can't speak with authority about T001 or other specific SAP tables, and it's relationship to the underlying database.  You might want to post your question in one of the more techie forums such as
    SAP on Oracle
    Regards,
    DB49

  • Client certificate question

    /* Style Definitions */
    table.MsoNormalTable
    {mso-style-name:"Tabla normal";
    mso-tstyle-rowband-size:0;
    mso-tstyle-colband-size:0;
    mso-style-noshow:yes;
    mso-style-parent:"";
    mso-padding-alt:0cm 5.4pt 0cm 5.4pt;
    mso-para-margin:0cm;
    mso-para-margin-bottom:.0001pt;
    mso-pagination:widow-orphan;
    font-size:10.0pt;
    font-family:"Times New Roman";
    mso-ansi-language:#0400;
    mso-fareast-language:#0400;
    mso-bidi-language:#0400;}
    Hello,
    I am novice with certs and I have a question. I want to implement EAP-TLS in a WPA deployment and I have a question about the client-side certificate.
    When I install a client certificate in a machine for a specific user, is this certificate only valid for this machine and this user? Or can I export this certificate and use it in another machine but the same user?
    Thanks in advance,

    From my experience, you can copy the certificate to another computer (assuming a modern OS).  There are two problems with this, though:
    1 - You must be able to export the entire certificate, including the private key, to be able to use the certificate on another machine.  Most PKI implementations prohibit/disable this.
    2 - If you can export the certificate, including the private key, then you are risking the loss of integrity of your PKI.  Someone else can get that cert with the private key and impersonate the user.

  • JDev 11g:  Task Flows,Rich Client- some questions to the Oracle team

    Hello,
    We're currently involved (early stages) in a large project with a Jan-Mar 2008 pilot delivery timeframe - one heavily including end-user customizable workflow elements, to which the aforementioned JDev/ADF 11 features would be a great fit.
    Our plans so far were to go ADF with BC and include Shale or Spring WebFlow plus custom development to provide for the very same things that the feature-complete Task Flows will offer.
    Questions:
    1. Delivery timeframe - any unofficial pointers on the schedule for the next JDev/ADF releases?
    2. Platform compatibility - Will the framework (ADF) run on J2EE 1.3 (OAS 10.1.2.x) or does it require 1.4+?
    3. Design-time-at-runtime for Task Flows - Do we have it in the released bits? (Couldn't find it). Any pointers on how is it supposed to operate?
    4. Am I right to think that what we're seeing with Task Flows is (among other things) a piece of the replacement functionality for the decommissioned Oracle Workflow subsystem (which bit us, by the way, as it was part of our original architecture).
    5. Will Rich Client support older platforms than the currently supported IE7 and Firefox 2?
    Thanks for any and all pointers,
    Thanassis Stathopoulos
    Athens, Greece

    Thanks Shay,
    http://www.oracle.com/technology/products/jdev/collateral/papers/11/newfeatures/index.html
    Regarding my question 3 "Design time at runtime" - quoting from "JDeveloper 11g new features" at the above URL, under "ADF Task Flow". The third entry from the paragraph's end reads:
    'Support for customization of task flows at runtime using "design time at runtime"'.
    I was wondering what this is exactly - web-deployable editor to edit the flow definition after deployment?
    Re your point 4, product replacement - with the feature set described in the paragraph above, Task Flows is much more than a Web UI controller. With "save for later", transactions, flow abstraction/nesting, BPEL outcalls, conditional logic, method calls etc etc it probably covers 50% of a typical workflow package and 75% of an embeddable workflow library (audit trail/timeouts/persistence nonwithstanding). This made me wonder whether it is part of a bigger, yet-to-be-announced OAS feature. Especially if one also considers the "design-time-at-runtime" sentence listed above.
    Now, definition-wise, I guess it' s whether this feature set is considered half-full or half-empty :-)
    Thanks again,
    Thanassis

  • Solman - Client definition in table T000

    Hello,
    I got a question from Security officer regarding Solution Manager landscape and the definition of table T000 for the productive client 800.  I am telling you that Solman functionality requires in some cases that you leave opened the client 800 to record transport requests for Test Plans and Test Packages (though you don't want to transport anyway, it asks you for a transport request),
    Is there a clear definition about what is the correct definition of table T000 for a productive Solution Manager?
    I appreciate your answers.
    Regards
    Esteban

    Already SCC4  entry for the client has old Logical system name.
    Ex:
    Client 500
    Logical System : <PRD>CLNT500
    I need this to be change to <QAS>CLNT500 after BDLS,but it is not happening after BDLs.
    Regards,
    Srinivas Chapa.

  • Client Proxy Question

    Hi all,
    Initially, when we have created a client proxy in ECC 5.0 with Oracle on Windows,
    we have chosen Local File as the WSDL source. We had the WSDL saved locally.
    This way we have generated a Client Proxy in SE80.
    Then we created a logical port using LPCONFIG.
    Since this the default logical port, we just saved it with all the settings imported
    from the WSDL definition.
    This way the Client proxy was working fine.
    The doubt that I have is, later on, if we want to just change the URL under
    call paramters in LPCONFIG to the URL of the WSDL residing on some
    other server(Non SAP), will the client proxy work?
    Regards,
    Vishnu

    Hi Gaurav,
    Looking at the error below, the binding is happening.
    However, there is a return code 503. And SAP is not able to access the
    Webservice located at the other server(Non-SAP).
    INFO 01:44:43: SOAP SESSION Protocol CL_SOAP_SESSION_PROTOCOL
    ->IF_SOAP_PROTOCOL~PRE_PROCESS() SOAP Session Protocol
    preprocessed
    INFO 01:44:43: SOAP Transport Binding CL_SOAP_HTTP_TPBND_ROOT
    ->IF_SOAP_TRANSPORT_BINDING~SEND() Try to send message ( DEST =
    ,PATH = ,URL = http://server name:portno/ ,SOAP Action = "ROI_R3
    _DF.roiSAPCreateOrder" )
    INFO 01:44:43: SOAP Transport Binding CL_SOAP_HTTP_TPBND_ROOT
    ->IF_SOAP_TRANSPORT_BINDING~SEND() Message sent
    INFO 01:44:43: SOAP Transport binding CL_SOAP_HTTP_TPBND_ROOT
    ->IF_SOAP_TRANSPORT_BINDING~RECEIVE() Try to receive message
    INFO 01:44:52: SOAP HTTP Binding CL_SOAP_HTTP_TPBND_ROOT->HANDLE
    _STATUS_CODE() Received return code 503 ( not available )
    ERROR 01:44:52: SOAP Message CL_SOAP_MESSAGE->IF_SOAP_MESSAGE
    _PART~INIT_FOR_DESERIALIZE() A SOAP Runtime Core Exception
    occurred in method XP_READ_TAG of class CL_SOAP_XP at position id
    1 with internal error id 112 and error text no extended error
    message (SOAP_ESERVER) (fault location is 1 ).
    Regards,
    Vishnu

  • Odbc client impact question 8.1.7 / Windows 2008 r2 64X

    howdy all , i have several questions which might have been documented , but not marked as supported so maybe someone could answer just a few of my questions here goes :
    - we have an oracle 8.1.7 (p2 ) database with support running on linux 2.4
    - the clients connecting to the database is an older 8.1.6 (2) odbc client ( either running NT / 2000 / XP ) 32bit
    - the application which uses this database and odbc connection is damgaard axapta 2.5
    i have read several documents that 10.2 oracle odbc client is supported connected to an older database, but i have not seen any documented related to my case
    ( either everybody has moved on or no-one else has experienced a similar situation )
    what im trying to accomplish , is connecting a sharepoint server (2010 ) with sql (2008 ) with powerpivot
    powerpivot reads data from the oracle server  through an odbc connection to an excel file ( works fine under a workstation rolled out as 32 bit )
    so my main concern is , what if i install the 10.2 client on my server 2008 connecting to the 8.1.7 database , what could be expected ?

    user4724664 wrote:
    howdy all , i have several questions which might have been documented , but not marked as supported so maybe someone could answer just a few of my questions here goes :
    - we have an oracle 8.1.7 (p2 ) database with support running on linux 2.4
    - the clients connecting to the database is an older 8.1.6 (2) odbc client ( either running NT / 2000 / XP ) 32bit
    - the application which uses this database and odbc connection is damgaard axapta 2.5
    i have read several documents that 10.2 oracle odbc client is supported connected to an older database, but i have not seen any documented related to my case
    ( either everybody has moved on or no-one else has experienced a similar situation )
    what im trying to accomplish , is connecting a sharepoint server (2010 ) with sql (2008 ) with powerpivot
    powerpivot reads data from the oracle server  through an odbc connection to an excel file ( works fine under a workstation rolled out as 32 bit )
    so my main concern is , what if i install the 10.2 client on my server 2008 connecting to the 8.1.7 database , what could be expected ?
    I would think you could expect a world of headaches.  Why, why, why do people insist on running software that hasn't been supported in the current millenium?  Why, why, why do they upgrade everything except the database?
    See MOS note 207303.1.  It appears that your proposed combination is not supported.

  • Client-server question

    hi there..
    I'm writing a java client that connects to a server to recieve some messages and classify them to different kinds, and then append these msgs to a database application. I have 2 questions:
    1. After connection is established to the server, how can my client program know that the server has been down for whatever reason? Note: im not the one who writes the server application, and that the server does not notify my client of its departure.
    2. Im thinking of creating a number of connections to the DB application thats equal to the no. of msg types i have( i have 4 msg typs, so i create 4 DB connections), will this improve the program performance more than having one DB onnection that carries all msg types?
    Thanks.. Mourad

    1. In your client if a connection is lost you will get an excepntion and you can deal with that...
    2. its a good idea to use a connection pool package to manage your connections. it will know when and how many open connections you will need. There are may packages out there. use an open source one

  • Client Reboot question

    When the client is set to auto reboot in like 1 hour you get a popup (2) one has a count down and then in like the last 15 mins you get another one that you can't hide, my question is there a way to change the notification message? If so how, Thanks.
    MSB

    Why not there has to be some file in sccm that I can modify.
    Why does there "have to be"?
    Are the strings displayed in the product somewhere? Absolutely.
    Did the they build the product with the ability for them to be changed? Probably as I'm sure the strings are already customized for different languages.
    Did they expose the ability to change the strings to the "consumer"? No.
    Could you find the resource files where the strings are stored? Most likely. It's probably in a resource DLL.
    Is it supported to change these by hacking the resource file? No.
    Is it possible that you screw something up in the process of hacking a resource file? Yes.
    Jason | http://blog.configmgrftw.com | @jasonsandys

  • Automating Essbase version 11.1.1.3 Client Installs Question?

    Sorry if this is not the place to post this question, but does anyone know if there is any other way to install Essbase client components individually for Essbase add-in and Essbase administration services (EAS) without using the Oracle install tool? I have a client that needs to automate the installation for 600 users and because they have no standard image the program they use needs to push the software to the desktops prior to starting the install.
    Using the installation tool required that approximately 4 GB of software be pushed to the client workstations prior to installing the software and they would like a smaller foot print. I know that SmartView has a thin installation program, but does anyone know of another option other than using the install tool provided by Oracle for the Essbase add-in and EAS?

    Hi,
    Have a read of this post for details on another option for installing the excel addin :- EssBase excel add-in automated install
    For EAS you can always use the web client so you wouldn't need to install anything on the client machine.
    Cheers
    John
    http://john-goodwin.blogspot.com/

  • WSUS - Forefront Client definition update not yet been downloaded

    Hi! I need help about 
    Definition Update for Microsoft Forefront Client Security - KB977939 (Definition 1.167.1978.0) 
    I have WSUS 3 SP2 and this definition appears like approved but the file for this update  have not yet downloaded 
    i downloaded the definition on Microsoft Catalog but i don't now how deploy in my clients with wsus automatically 
    where save the package ? 
    thanks !

    Hi,
    Mostly, update will only be downloaded after approval. So please confirm if you have approved it.
    Since you have downloaded it from Microsoft catalog, you can import it into WSUS server.
    Step 3 Import and Export: Copying Metadata from Database
    http://technet.microsoft.com/en-us/library/cc720437(WS.10).aspx
    Hope this helps.

  • Bluetooth simple client example question

    Dear all,
     I am not sure whether I am close to make it. I update my bluetooth driver {http://forums.ni.com/ni/board/message?board.id=170&message.id=301780#M301780}, so it seem the labview (simple bluetooth server example) can work with my bluetooth (please read the first picture). However, the error message comes up when I run the simple client program. I have a question. Do I have to make the client and the server program in two computer?
     I need your help
    Thanks
    ====================
    =Labview 7.0 & 8.0 & 8.5=
    =====================
    Attachments:
    labview5.JPG ‏203 KB
    labview6.JPG ‏214 KB

    Just want to make a couple things clear
    1. my bluetooth device is working well, after I remove dell driver and using XP bluetooth stack <-----I think
    2. I can run the server without no problem, the problem comes from the client part.
    I have attached the pic from the device manager
    Please help, failure will lose my job
    ====================
    =Labview 7.0 & 8.0 & 8.5=
    =====================
    Attachments:
    labview7.JPG ‏41 KB

  • Client Server question:Getting client responses in 2 locations, avoiding

    I have a client that holds all the connections. Then a GUI running inside the same VM can do specific on demand actions on some selected clients.
    I keep all the connections inside a Vector, then in the GUI when I want to send for example to a client a "ver" command the client should respond with his version.
    I implemented the behavior but my question is as follows: when I click the ver button on the GUI form, I get the Vector connection from the server class, then I get the associated BufferedReader and PrintWriter from that Vector<Connections>, where Connections is the class that holds socket, bufferedreader and printwriter, I send to selected clients the "ver" string and when a client responds with his version the string is first caught in the ThreadedServer while loop
    while ((temp = br.readLine()) != null)instead of the GUI form class or any other class that implements the behavior.
    The client should send twice the response. Why is that? And why is firstly caught in the server then in the loop that sended the "ver" command.
    Is there a way to avoid this? I mean can I get the response in the method that asked for that version first not in the threaded server that holds all the connections?

    1) What will happen ?Depends on what your code does.
    2) Will both clients get their own objects seperately ?Depends on what your code does. Remote objects can have an 'id'. If the ids are the same then the same object is produced If not then different objects.
    It is likely that your code is using different objects, but that is only a guess.
    3) How do they track the use of a method in the remote object ?How does "who" track it? And why would it matter?
    A client doesn't 'track' code in a remote object. Basically a RMI proxy sends a message to the RMI server. The RMI server gets the message and processes it. The act of processing causes the object method on the server to be called. When that method exits a message is sent back to the client. At that time the client returns from the method that it called. Notice that the client doesn't do anything with the server code all it does is wait for the message to come back. Does that answer your question?

  • Endpoint Protection Client - definitions couldn't be updated

    Am on SCCM 2012 SP2 and have EndPoint protection client deployed to computers during the Task Sequence, and they get the update.
    The next day I will try an update from the client's GUI and will get the error "Virus and spyware definitions couldn't be updated". 
    In the antimalway policy applied to the collection the device is a member of, I indeed have its definition update source set to "Updates from UNC file shares", then in the server path for the UNC, it is set to "\\server.domain.com\D$\sources\Packages\Apps\Microsoft\EP_Definitions\Updates\x86"
    which is where the "mpam-fe.exe" and "nis_full/exe" files are. 
    I have no maintenance windows set on the device collection that this antimalware policy is applied to. 

    Hi, I just wanted to clarify, I only have the "UNC" path as the option for the source of updates. 
    I have also verified that in the registry in hklm policies Microsoft AntiMalware that the UNC path is indeed there and I can manually access the path from Start > Run. 
    I've tried running the Endpoint definitions update manually as an Administrator, and with the Windows Update service in every combo of state I could try, and still nothing (not that I want Endpoint getting updates from the internet anyways). 
    Not sure what we're looking for in the windowsupdate.log but here are the last few lines before the time I tried running the update. The error from definition update doesn't appear to add anything to this log file.
    2014-11-19 18:50:01:854
    1012 10a0
    Service *************
    2014-11-19 18:54:12:693
    2068 1398
    Misc ===========  Logging initialized (build: 7.5.7601.17514, tz: -0600)  ===========
    2014-11-19 18:54:12:693
    2068 1398
    Misc  = Process: C:\WINDOWS\CCM\CcmExec.exe
    2014-11-19 18:54:12:693
    2068 1398
    Misc  = Module: c:\Windows\system32\wuapi.dll
    2014-11-19 18:54:12:693
    2068 1398
    COMAPI FATAL: Unable to connect to the service (hr=80070422)
    2014-11-19 18:54:12:693
    2068 1398
    COMAPI WARNING: Unable to establish connection to the service. (hr=80070422)
    2014-11-19 18:54:33:507
    2068 1098
    COMAPI FATAL: Unable to connect to the service (hr=80070422)
    2014-11-19 18:54:33:507
    2068 1098
    COMAPI WARNING: Unable to establish connection to the service. (hr=80070422)

  • Forms 6i client/server questions

    Hi, I have three questions:
    1) Is Patch 17 for Forms6i cumulative? (we have patched up to patch 13)
    2) After applying patch 17 do we need to recompile all of our forms?
    3) We are connecting from 6i client/server forms to a 10.1.4 database, is there a recommended Oracle client? (we are very behind and still running a very old version and I'm trying to get information to move forward)
    Thanks.

    Patch 17 is designed for the 10g database. There is no hint, that a 9i-db is better than a 10g. Use 10g.
    I always recompile all (!) libs, menues and forms after installing a new patch. It's better to do that once, than to search for failures, that result from patching.
    try it
    Gerd

Maybe you are looking for

  • How can I convert an iMovie to a DVD?

    I have a movie created in iMovie and I need to put it into DVD format so that I can add menus and such. Help? I have a new Mac that is not supported by iDVD so what do I do?

  • How to Resend OR Cancel Exactly Once Messages

    Dear all, I have nearly 20 messages with status "Holding" under adapter engine message monitor. These interfaces are Exactly Once type. I am trying to Resend or cancel them. Nothing worked here. When i check the error reason for the Holding status me

  • Won't sleep. Instead logs out?

    When I click Sleep on MacbookPro, it logs off, instead of sleeps.  So each time use have to provide password.

  • How do I know the correct burn speed?

    My Pioneer/Apple DVD burner is recognized in the iTunes preferences, I have DVD-R media, but when I try to back up my iTunes, I get an error message saying that the burn speed is too fast for my media. I selected 1x speed in preferences, and I still

  • Change SMB settings (sharepoint, status, etc.) with Server Admin

    Hi. Big problems with Server Admin to administer SMB service with Mac OS X Server 10.5.2. When you share a sharepoint, you set the enable oplocks and enable strict locking as you want and when you save, all your settings disappear. Fix with modifying