(OT) Database design question

ablmf has asked for the wisdom of the Perl Monks concerning the following question:

I am not very familiar with Data Base, so my question is, which design is better?
If all the servers share a single DB, I don't have to design a protocol between the them. But I think it's not as efficient as the second one.
This is a ipphone system, the servers have to manage calls from phone to phone. If you have ever heard about SIP，you may know what I am talking about.
The servers are called Proxy Server and the DB are called Location Server where Proxy looks up for telephone information such as the ip address of callee phone.
Actually I want to write the system in perl and that's why I asked question here.
I would use MySQL as DB and Linux as OS for both proxy and location servers.

+--------+   +--------+   +--------+
|        |   |        |   |        |
| Server1|-->| Data   |<--| Server2|
|        |   | Base   |   |        |
+--------+   +--------+   +--------+
                 ^
                 |
             +--------+
             |        |
             | Server3|
             |        |
             +--------+

+--------+   +--------+   +--------+   +--------+
|        |   |        |   |        |   |        |
| Server1|-->| Data   |   | Server2|-->| Data   |
|        |   | Base   |   |        |   | Base   |
+--------+   +--------+   +--------+   +--------+
    ^                         ^
    |                         |
    +-------------------------
                 |
                 V
             +--------+   +--------+
             |        |   |        |
             | Server3|-->| Data   |
             |        |   | Base   |
             +--------+   +--------+
[download]

2006-06-11 Retitled by planetscape, as per Monastery guidelines

( keep:2 edit:16 reap:0 )

Original title: 'I have a design problem'

Comment on (OT) Database design question Download Code

Replies are listed 'Best First'.
Re: (OT) Database design question by Fletch (Bishop) on Jun 11, 2006 at 16:27 UTC
About the only sensible answer you're going to get is: It depends. And more precisely, it depends on a whole lot of stuff you haven't told us (your specific application, the type of hardware you've got available, what the actual RDBMS software is, maybe the phase of the moon; the jury's still out on correlation with the price of tea in China). Provide more context. See also How (Not) To Ask A Question.	[reply]
Re: (OT) Database design question by planetscape (Chancellor) on Jun 11, 2006 at 19:43 UTC
I heartily recommend Database Design for Mere Mortals, 2nd Ed. HTH, planetscape	[reply]
Re: (OT) Database design question by marto (Cardinal) on Jun 11, 2006 at 16:28 UTC
ablmf, This is not really a Perl question. You could have given us some more information to help us to help you. Which database platform are you using? Your diagrams are little to work with. Perhaps you could describe your environment, and what exactly you are trying to achieve. I suspect you want to look into Replication (MySQL replication, MS SQL Server Replication for example). Please read the PerlMonks FAQ and How do I post a question effectively? if you have not already done so. Martin	[reply]
Re: (OT) Database design question by castaway (Parson) on Jun 12, 2006 at 08:06 UTC
More information would be nice.. I'll take a wild guess though. Are all those databases supposed to be holding the same information, or is it different sets? If its the same, then that looks like a synchronisation nightmare, and I'd suggest having one database, or a a cluster with decent replication. C.	[reply]
Re: (OT) Database design question by girarde (Hermit) on Jun 11, 2006 at 20:40 UTC
You asked "which is better?" but you have not really presented two alternatives. If you are using two instances of MySQL in the second diagram, I don't see where the need to design a protocol arises. If you are proposing a different database on the second server in the second diagram, which one is it?	[reply]