🚀 go-pugleaf

Hi!

I'm considering using replication to solve a performance problem I have. 
  Currently, there are 5 or 6 machines working with a Linux database 
server on a LAN, which is fine. However, there are also 3 clients 
connecting to the same server using a 128 bps ADSL internet connection. 
I've already done quite a lot of optimisation, and the program is 
usable; still, it is _very_ slow. So I was wondering if replication 
would be a good solution.

I've had a look at the standard IB replicator, as well as a few others, 
but it seems that it might be hard to make it work just the way I want 
it. I would like to synchronize updates in both directions, over all 
tables of the database, as frequently as possible. The clear advantage 
of doing things this way, is first of all, that select queries aren't 
sent to the server, only udates inserts and deletes, and secondly, that 
the user doesn't have to realise that the updates are being applied; 
everything seems instantaneous. All 4 databases must be identical. So my 
needs are very simple. Do you think it would be a good idea to implement 
it by myself (simply using triggers and a log table, with a process 
running in the backgroud, sending and getting changes), or would I be 
better off using the standard one?

Another factor to consider is the (crushing) cost of licences in such an 
installation...

Any advice, help or examples would be greatly appreciated!

Thanks a lot!

Jonathan Neve.