Client/Server network traffic.

I don't know if here is the right forum to ask it, but let's go on.
Nowadays we a system on Forms 4.5/windows/Oracle 8.0
We have some clients machines linked to the server by a Frame-Relay link and we use Windows 2000 Terminal Server to reduce the network traffic.
Well, we are gonna update to Forms 6i/Oracle 8i.
I would like to know if Forms Server can be a good option to Windows 2000 Terminal Server.
If I haven't been clear with my question I'll can do it again. I'm new on Forms 6i and I've just installed it to test (It hasn't been working yet....)
Thanks in advance
Ronaldo.

This could provoke quite a discussion.
"If your client are on slow links you may, for example query 10,000 rows of data. "
Generally, forms shouldn't be pulling back tens of thousands of rows from the database. Your user is unlikely to page through that much, so either you want a summary (which is best calculated on the database server and the summary results dragged across the network) or you are paging through records in the tens, not tens of thousands. (Look at the 'Number of recods buffered' property in your base table blocks. Bet it's one or two digits, not five!)
"The information that is transmitted to the client is basically screen draw information - and this will be alot less than the 10,000 rows you were querying before. "
As above. In client/server you shouldn't have been bringing stuff from the database down to the client that wasn't going to be on the screen anyway, especially if you had a slow network.
Depending on how your application is written, it could well peform a lot worse on the web than in client server. Rather than having the client do a fair share of the work, it's now got to keep talking to the application server to get anything done.
For example, because navigation triggers don't allow the use of restricted built-ins, rather than putting code in a 'post-text-item/when-validate-item' trigger on the relevant item, it gets put in a form level 'when-new-item-instance' trigger. It's a bit untidy in client/server but workable. Put it on the web, and every time the user tabs between fields, the form has to go off to the application server to fire the when-new-item-intance trigger to tell it what to do next.
Another 'speed-bump' is if you have any synchronize bits (eg in a post-query trigger, or as part of a "I'm 10% complete" information messages in long running loops). These will also generate network traffic between the app server and form.
While web server does have advantages, I wouldn't be selling it on it's performance improvements.

Similar Messages

Client - Server Networking

I did not know the right place for this topic , but it may be as much of a networking problem or an applet problem.
The requirement is that the client interacts with the server through an applet (client side) , the problem being that an unsigned applet is not allowed to create connections , is there any way in which we may overcome this problem ? (Use of applets is a must)
The source is as follows
class Ping implements Runnable{
    private Thread PingThread;
    private String PingServerAddr;
    private DatagramSocket socket;
    private int Socket_no;
    Ping(String PingServerAddress){
        //Create a new ping thread
        PingThread = new Thread(this,"Ping Thread");
        PingServerAddr = PingServerAddress;
        Socket_no = 25519;
        try{
            socket = new DatagramSocket(Socket_no);
        }catch(SocketException socketexception){
            socketexception.printStackTrace();
            System.exit(1);
        PingThread.start();
    private void doPing(){
        byte PingData[] = "Alive".getBytes();
        DatagramPacket sendPacket;
        try {
            sendPacket = new DatagramPacket(PingData, PingData.length, InetAddress.getByName(PingServerAddr), Socket_no);
            socket.send(sendPacket);
        } catch (UnknownHostException ex) {
            ex.printStackTrace();
        } catch (IOException ex) {
            ex.printStackTrace();
    public void run(){
        while(true){
            doPing();
            try {
                Thread.sleep(10000);
            } catch (InterruptedException ex) {
                ex.printStackTrace();
//A Ping object has been created in the init() function of the appletThe error is as follows,
Exception in thread "Ping Thread" java.security.AccessControlException: access denied (java.net.SocketPermission ***.***.***.***:**** connect,resolve)
     at java.security.AccessControlContext.checkPermission(Unknown Source)
     at java.security.AccessController.checkPermission(Unknown Source)
     at java.lang.SecurityManager.checkPermission(Unknown Source)
     at java.lang.SecurityManager.checkConnect(Unknown Source)
     at java.net.DatagramSocket.send(Unknown Source)
     at Ping.doPing(Player.java:81)
     at Ping.run(Player.java:92)
     at java.lang.Thread.run(Unknown Source)
Any help would be appreciated.
Regards,
Flummoxed

No, signing requires the author of the applet to do
some work.
The only alternative is to only try to connect to the
host where the applet was loaded from.Yes ultimately that is what I had to revert to , we used the getCodeBase().getHost() in order to speak to the server applet.

Networked client-server applications (newbie)

Hello everyone,
(Apologies if this post is irrelevant for this particular forum :)
As I understand peer-to-peer networks, separate instances of application software are installed on 'each' computer, and files are then shared by everyone logged on. My query regards client-server networks (i.e. a centralized server with less powerful client computers attached to it): Would someone please explain (and possibly point to additional resources) how applications are installed on these networks? In particular, is it true that there's only one instance of application software (e.g. MS Word) installed on the central server, and that this is then somehow shared by all the connected users? ... and if not, is it possible to have such an arrangement, so that when it comes to upgrading any application software, the update only happens at one location (instead of on each client machine)? Thanks very much for your time.

Guy: Thanks so very much :)
I see now. I was specifically wondering about networks involving WinNT / 'Win 2000 Server' as the 'central' server OS, but think that these would involve heavier clients than the mainframe example.
I assume that in a Win NT client-server network, if a client wanted to run a program (e.g. MS Word), the Win NT OS would transfer a 'copy' of the MS Word executable to the client machine where it would run and consume the resources of the client. Win NT would do this again for any other client, and so ultimately there would only be one 'stored' copy of MS Word. Also, given copies of executables are transefered, resources of the clients are used up, resulting in medium 'weight' clients with relatively less central server load.
It'd be great if you could verify this assumption,
Regards, SP.

Server admin not showing network traffic graph

This topic was posted a while ago with no solution, and it has been [archived|http://discussions.apple.com/thread.jspa?messageID=6953359]. I have just experienced the same problem with Leopard Server Admin (SA). It seems to be a problem on the server, not in the client admin, since it occurs both when I use SA on the server itself and a client, and deleting client prefs. has not helped. I'm presuming that there is a corrupted database or prefs. file for this graph.
All the other SA graphs show just fine. Fixing permissions has not helped, either.
Anyone have any idea where to look on the server? Thanks.
Message was edited by: Trevor Jacques

I'm pretty sure I have since I have a few Mac Mini's on the network with their home directory residing on the server and my Windows users are sharing a few daily use files on the server including a MYOB data file, also our intranet web site is hosting on the same server, not to mention I have 3 print queues there and all the outgoing mails from both of our all-in-one scan to email machines are using the same server. All of these services and users are using their computer as if nothing is wrong, so I have to assume that there is network traffic unless server admin interprets the term "network traffic" differently.

How can i split Client Network traffic and My exchange traffic with two differnet ip address?

Hello Everyone
sorry for my bad English and also my bad explaine
here is my network looks like
all the client on one subnet and network is 192.168.0.0
i have Dsl router that connect to Tmg server
i have Tmg with two NIC
1-192.168.0.4
2- 2 Public Ip address
i want to do this
i want to split user's traffic and my exchange traffic
i mean i want to route user's traffic with one public ip addresss
and my exchange server's traffic with another public address
but when i add additional ip address at Tmg or create new NIC card
all of my traffic route with one public ip Address what should i do?

Hi Uhan,
You need to use ENAT fuction on TMG to achive this
On the External NIC assing the Second IP as Additionl IP address (VIP)
Create a network Rule to NAT traffice From Exchange server IP address to the required Public IP which you need the E NAT.
Ensure you are creating rule only from Exchange server IP and not all INternal.
Look at the below Doc you step by step config
http://www.isaserver.org/articles-tutorials/configuration-general/Configuring-One-to-One-NAT-TMG-2010.html

Capturing SQLNET client-server traffic

Pardon the somewhat newbie-like question ...
I am trying to build a API test stub where I need to be able to capture the requests and responses between third-party two-tier applications then replay the traffic. The replay is easy. The capture, though, is giving me trouble.
For HTTP traffic, it is easy. I just change Internet Explore's HTTP port address to point to a fake proxy which records the requests the sends the requests onto the real web server.
Is there an equivalent to this in the SQLNET world? I am hoping that a proxy already exists to do this.
If not, is there a way to turn on a full trace for a specific client? I would need to capture all connections coming out of the client and no other client's connections.

This could provoke quite a discussion.
"If your client are on slow links you may, for example query 10,000 rows of data. "
Generally, forms shouldn't be pulling back tens of thousands of rows from the database. Your user is unlikely to page through that much, so either you want a summary (which is best calculated on the database server and the summary results dragged across the network) or you are paging through records in the tens, not tens of thousands. (Look at the 'Number of recods buffered' property in your base table blocks. Bet it's one or two digits, not five!)
"The information that is transmitted to the client is basically screen draw information - and this will be alot less than the 10,000 rows you were querying before. "
As above. In client/server you shouldn't have been bringing stuff from the database down to the client that wasn't going to be on the screen anyway, especially if you had a slow network.
Depending on how your application is written, it could well peform a lot worse on the web than in client server. Rather than having the client do a fair share of the work, it's now got to keep talking to the application server to get anything done.
For example, because navigation triggers don't allow the use of restricted built-ins, rather than putting code in a 'post-text-item/when-validate-item' trigger on the relevant item, it gets put in a form level 'when-new-item-instance' trigger. It's a bit untidy in client/server but workable. Put it on the web, and every time the user tabs between fields, the form has to go off to the application server to fire the when-new-item-intance trigger to tell it what to do next.
Another 'speed-bump' is if you have any synchronize bits (eg in a post-query trigger, or as part of a "I'm 10% complete" information messages in long running loops). These will also generate network traffic between the app server and form.
While web server does have advantages, I wouldn't be selling it on it's performance improvements.

Monitoring network traffic from large queries to client

Hi, is there any way to monitor the volume from network traffic from the server to the client, the statistics SQL*net message to client and SQL*net message from client don't show a full picture.
I have a query that sends several GB of data to a client but the SQL*net statistics done show anything like this, is there any way I can see the volume of traffic from the RDBMS statistics?
Regards.

Alan Wilson wrote:
Hi, is there any way to monitor the volume from network traffic from the server to the client, the statistics SQL*net message to client and SQL*net message from client don't show a full picture.
I have a query that sends several GB of data to a client but the SQL*net statistics done show anything like this, is there any way I can see the volume of traffic from the RDBMS statistics?
Sounds like you're looking at the Wait events only - you need to check the instance (or session) activity - v$sysstat/v$sesstat for the following stats:
NAME
bytes sent via SQL*Net to client
bytes received via SQL*Net from client
SQL*Net roundtrips to/from clientIt's worth pointing out, by the way, that SQL*Net can do "compression" on the fly by comparing columns across rows of data sent, so the volumes you see in these stats may not match the volume you finally see arriing in the client code. ( http://jonathanlewis.wordpress.com/2010/05/07/sqlnet-compression/ )
If you need to optimise the transfer rate, look at the SQL*Net configuration (at both ends) for the Session Data Unit (SDU) and the o/s transmit and receive buffers.
Regards
Jonathan Lewis

Client /server model of network communication

can some some tell me what is meant by the client / server model of network communication?

See this networking tutorial
http://java.sun.com/docs/books/tutorial/networking/index.html
You could have asked both questions in the same post, as they are related.

Client/server - sending ArrayList over network

Hi all
I have have written a multithreaded client/server application where multiple clients connect to the server. For each client, the server spawns a new thread to handle the connected client. When a client connects to the server, the client sends its name to the server. The server stores the name in an ArrayList and sends the whole ArrayList to all the connected clients.
It worked fine a few days ago. But now I have a problem: When I send the ArrayList to all the connected clients then only the newly connected client gets the updated ArrayList while the previously connected clients get the ArrayList that they got previously i.e. they are not updated. If instead of sending an ArrayList, I just send the name of the newly connected client to all other connected clients then it works i.e. each client gets updated with that name. It also works fine if I just send an Integer value to all the clients upon a new connection.
I know the ArrayList implements Serializable, but I dont know how it doesnt work anymore!
Plz help me with this. I have been trying all combinations, but no luck so far.
thank you
taurean

thank you jwentling
to make sure that we are on the same page, here is the relevant code.
This method updates the list of the peers' names on the Gui (that extends JFrame).
public void update(final ArrayList dataPacket)
          SwingUtilities.invokeLater(new Runnable()
               public void run()
                    setTitle(dataPacket.toString());
                    peerList.setListData(dataPacket.toArray());
                    peerList.updateUI();
     }This inner class continuously listens to server to get the arraylist that contains the names of the other peers/clients that are connected.
The con varibale is the Connection where I obtain input/output streams for this client.
Note: con.getInputStream() returns ObjectInputStream.
     private class Listener implements Runnable
          public void run()
               while(true)
                    ArrayList dataPacket = (ArrayList) Receiver.receiveObject(con.getInputStream());
                    if (dataPacket != null)
                         update(dataPacket);
     }Here is how the static receiveObject() metod looks like:
public static Object receiveObject(ObjectInputStream stream)
          Object obj = null;
          try
               obj = stream.readObject();
          catch(ClassNotFoundException cnfe)
               System.out.println(cnfe.getMessage());
          catch(IOException ioe)
               System.out.println(ioe.getMessage());
          return obj;
     }i think that the receiveObject() method should return the updated list (sent from server), but i get the previoulsy sent list fromt his method!!!
Has it to do something with the stream? I mean could it be that the list is still in the stream (as you pointed out) and instead of the updated list, i just get the one from the stream that i got initially?!
thank you for ur time

ACE server initiating traffic issue

Hello
I'm trying to establish a session between one of my real servers behind the ace and some external network without any NAT. According to the documentation I should only configure correct ACL on both the client and server vlan and it should works. Unfortunately, although I see hits in the ACL configured on the outside direction for client vlan the traffic is not passing the ACE.
When I configure the capture I can see traffic only in the server vlan. There is no traffic in the client vlan.
Does anybody know what else should I configure ?
Thank you in advance
Regards
Lucas

Hi,
the capture feature on the ACE only works in the input direction:
The packet capture function enables access-control lists (ACLs) to control which packets are captured by the ACE on the input interface.
http://www.cisco.com/en/US/docs/app_ntwk_services/data_center_app_services/ace_appliances/vA3_1_0/configuration/admin/guide/managesw.html#wp1035160
Is your problem resolved now or does this still not work?
If so, what do you see in the capture?
HTH,
Dario

Curious about App-V 4.6 RO Cache Network Traffic

Our environment is Server 4.5 SP2, and Client 4.6 SP3. We utilize RO cache on DFS network storage, and our clients are VDI VM's.
We recently ran into an issue when we recycled a large pool of VM's where the CPU/Network Utilization was maxed on our cache file server. Upon further investigation there appears to be about 550MB of traffic between each VM and the Cache file itself when
the VM boots (no users logged in). Can someone explain to me what this traffic could be?

In RO cache mode (ReadOnlyFSD], *nothing* should get streamed down to the client at all.
where does the amount of data show up - any specific files/folders, or monitored network traffic between the client and the DFS share?
is there anything in the App-V log file?
Did you initiate a server publishing refresh with a user that has access to all applications before you sealed the VDI Master image?
Falko
Twitter
@kirk_tn   |   Blog
kirxblog   |   Web
kirx.org   |   Fireside
appvbook.com

Network traffic

Hi,
to reduce networktraffic I read in a book: validate at a high level, f.i. recordlevel.
But, I also read that when going from one item to the next, a network round trip is necessary. When that is necessary why shouldn't I validate the item at item level?
Besides that, is there a site containing tips & tricks about reducing network traffic in an Oracle forms environment?
Thanks
Frank

Frank go to otn.oracle.com/products/forms and click on the Forms upgrade center - loads of papers which tell you how to get the best from Forms.
As for your question...The reason there is a network trip is that all trigger code is on the application server. When you tab from one field to the next, the Forms client has no idea if there is trigger code so it makes a trip to the app server to say "I have just moved from A to B. Any triggers to fire". The app server will then fire all necessary triggers (or infact say that there are not triggers to fire). It will then sent a message back to the client to say how the screen should look after doing this navigation. (And the message could be "Nothing changed").
Have a read at the papers and let me know if you have nay more questions.
Regards
Grant Ronald
Forms Product Management

Server network in a Netweaver environment

Hello there,
I have a question regarding building a server network in a Netweaver landscape.
First of all, here's our situation - Our SAP systems are hosted externally and we've had a few ORA-3113 and ORA-3114 errors recently (for those not using Oracle - this is an error that occurs when the application disconnects momentarily from Oracle. It usually happens during network problems). In investigating this problem, I discovered that our hosting partners have not build a server network and all server traffic shares the client network. The problem was caused by the daily backup being incorrectly configured so that it also ran over the network instead of over fibre to the tape library.
I did some further reading into the above problem and found the 'Adaptive Infrastructure for mySAP' book from SAP Press states that server networks are generally no longer necessary because a Gbit LAN backbone can provide enough bandwidth to handle up to 4000 users (we have only 600). However, the book does not mention how much traffic can be generated in a Netweaver environment with R/3, BW, APO, SolMan, etc... components all talking to each other. We currently only run ECC5 (with 3 app servers) and SolMan, so our server traffic isn't great. However, we plan to add BW and APO soon and I fear that the lack of server network could cause us problems.
On to my main question - can anyone provide me with real-world examples of how your landscapes are built and whether or not you use a server network to isolate your server traffic. What are your thoughts on the subject?
Kind regards,
Arwel Owen.
BASIS Manager,
Princes Ltd.,
Liverpool, UK.

Hi,
Since each KMS key can be installed on up to six KMS hosts. You can configure another KMS host in you test environment.
Not quite sure about the KMS host upgrade you referred. If you don’t want to update the OS, some patches will be required.
For your information:
Understanding KMS
http://technet.microsoft.com/library/ff793434.aspx
An update is available for Windows 7 and Windows Server 2008 R2 KMS hosts to support Windows 8 and Windows Server 2012
http://support.microsoft.com/kb/2691586
Update adds support for Windows 8.1 and Windows Server 2012 R2 clients to Windows Server 2008, Windows 7, Windows Server 2008 R2, Windows 8, and Windows Server 2012 KMS hosts
http://support.microsoft.com/kb/2885698
Hope this helps.

New server Network & DNS Setup

I just got a New Dual G5 Server and set it up using the pdfs from apple. Mainly the Getting Started Appendix B page 149 Small Business Setup using the Gateway Setup assistant. I ran into all sorts of problems like being unable to login to the open directory and i think it was my DNS setup that was the problem i then changed the setup multiple times using various How-tos in this thread to no avail.
So i have reinstalled the the Server software and will start again using the following steps.
This is my first post so i please let me know if i have provided enough info or to much?
SETUP:
ADSL 2 connected to DLINK 4 port Wireless Router (24Mbps) 1 port plugged into Ethernet 1 on Server. Router is setup as a DHCP using 192.168.0.1 and 192.168.0.2 for xserve. subnet mask is 255.255.0.0 (dynamic ip address from ISP)
Xserve Ethernet 2 will provide DHCP and NAT via a 16 port switch to Computers and Printers. Setup as 192.168.2.1, 255.255.255.0
I want to use the xserve as a file server, filemaker database server, web server for a test site for some of our content and also possibly an email server.
INSTALL
1. Install OSX 10.4 with ethernet settings above eth1 is DHCP eth2 is manual
assigned, no services will be set to start up and i choose Standalone
Server
2. Update with Software Update
3. Setup DHCP for the Clients - Start Service
4. Setup DNS use " company.private" for the host name and the DNS server ip
will be the 192.168.2.1 address
5. Upgrade to Open Directory Master and Kerberize
6. Setup Sharepoints for Data Storage and Home directory location
7. Start Apple File Services AFP
8. Create user Accounts / Groups
9. Setup Directory Access on clients to see LDAP server and DHCP.
Login from clients to get an automounted sharepoint and Home directory,
"sounds easy when i type it up like that".
Could anyone let me know if i am doing it the right way or if there is anything i have missed wrong order or something important.
Should i change the routers subnet mask to 255.255.255.0 ?
Obviously wireless clients will have access direct to internet due to the router and the routers DHCP. Can wireless users access their Home directory in this manner via the ethernet 1 port?
Should i just let the router handle the DHCP for everyone and plug it into the switch with only 1 ethernet being used on the server?
I will eventually have to have another subnet for student computers in the future.
I would greatly appreciate any advice

Pretty long post, and I doubt anybody here will want to go through the entire setup process. However, I'd re-think your baisc hardware configuration.
I don't know why everybody gets so excited about using that second ethernet port on the Xserve. It's necessary in some setups, but if you already have a router providing DHCP on a small network it seems like overkill.
I'd setup the chain in this order:
Internet - ADSL Modem - Router - Switch - Xserve / Computers / Printers
Setup your server with a manual IP address (192.168.0.2 is fine). Point it to itself as the DNS server.
Leave the router at 192.168.0.1, and configure the router to assign DHCP above the Xserve's address (say, from 192.168.0.10 and up). Also let the router handle NAT.
Network homes work over wireless, but it's not an entirely satisfactory experience. You might consider using portable homes instead; sync times will still be slow, but overall you'll cut down on network traffic and it should at least appear faster to your users.

Client/Server to Web-Based application Conversion

Hi! Everyone,
I have couple of questions for you guys.
Our Client had recently upgraded Forms 4.5 to 6i to move from Client/Server based application to Web based application.
They are using Forms Server 6i Patch Set 1, OAS 4.0.8.1, Windows NT Service Pack 5 and Oracle 7.3. They are facing the following error every now and then, when they run the forms,
"FRM-92100: Your connection to the server was interrupted. This may be the result of a network error or a failure on the server.You will need to re-establish your session."
Please let me know what might be causing the above error. The only problem i can think about might be Oracle 7.3. If i am right only Oracle 8 and above supports Forms 6i.
Can anyone let me know some tips and/or techniques to upgrade Forms 4.5 to 6i. If there are any important settings/steps which we might have over looked during the upgrade, please list them.
Any kind of help is greatly appreciated.
Thanks,
Jeevan Kallem
[email protected]

Most of the code is use with no changes at all.
See otn.oracle.com/formsupgrade
Regards
Grant Ronald

Client/Server network traffic.

Similar Messages

Maybe you are looking for