Re: SQL Net question...

From: Joel R. Kallman <jkallman_at_us.oracle.com.no.spam.please>
Date: 1997/05/06
Message-ID: <336fb651.36663679_at_newshost.us.oracle.com>#1/1


On Tue, 06 May 1997 14:32:46 -0500, Vijay Damodaran <vijayd_at_nortel.com> wrote:

>I have a database application running on a server in
>subnet A in Dallas. Can a person in subnet B in England
>access the database via SQL*Net?

Answer: Yes. For clarity, let's say that you can have users on physically distinct LAN's access the same Oracle database using SQL*Net.
>
>We have developed an application using Developer 2000 which
>must be deployed at both sites. We have the following
>options:
>
> 1. Deploy the GUI on client machines in England
> which will access the database in Dallas via
> SQL*Net. Is this possible?

Yes, you can do this. But if this is a heavy OLTP application, and you have a large number of users sending all of their data concurrently down the same link to Dallas, you better have a pretty hefty pipe.
>
> 2. Create a similar database server in England and
> have enable replication between the 2 databases.
> But, somebody in England says that do replication
> is 6 times more costly (I do not know what she
> means by this - money or time or n/w traffic). Is
> this true?
>
>We feel that using option 1 would make the application very
>slow. But the costs may make option 2 prohibitive.

With this approach, your costs increase (now need *2* servers instead of one, and you still need the link between Dallas & England to permit replication of your data). As well, this might take another body to administer this machine, maybe someone to administer the database, etc.
>
>any input would be greatly appreciated.

This is known as compromise-city. Only you can determine what is best suited for your environment. If this is a high-latency application with very few (<5) users at the remote site, then I wouldn't feel too bad about going with option #1. I doubt this is the case, though.

If money were no problem (and unforunately it always is), I would seriously look at option #2, simply because this gives each group of users much better performance, and as well, you have some built-in data redundancy which can be put to use in disaster-recovery situations. Lastly, let's say for some reason, the Dallas site goes down. With option #1, everyone is dead. With option #2, the folks in England keep on chugging.

It's a hard question, which ultimately, only you know the details of what is viable and what is not. Hope this helps, somehow.
>
>Thanks,
>
>VJ.
>
>--
>Vijay Damodaran
>vijayd_at_nortel.com
>Work: (972) 685-8150

Thanks!

Joel

Joel R. Kallman          Enabling the Information Age through
Oracle Government                Network Computing!
Bethesda, MD                  http://govt.us.oracle.com
jkallman_at_us.oracle.com          http://www.oracle.com



The statements and opinions expressed here are my own and do not necessarily represent those of Oracle Corporation. Received on Tue May 06 1997 - 00:00:00 CEST

Original text of this message