Thanks for your answers tills now.

as you mentioned the data acquisition server must work 24 h/day

then the general idea is to :
- have a cache in memory (this cache must be controlled by
the application, not directly by firebird)
- using more connection to the DB to move data to the DB
in big transactions (with more seconds of data)

Do you know is could be better to use an embedded version
of firebird instead of standard (super server) one for
windows (I am using 1.5.0 on the plant) ?

Michele



--- In firebird-support@yahoogroups.com, Adomas Urbanavicius
<adomas@i...> wrote:

> Alexander Gräf wrote:
>
> >
> >
> >>-----Ursprüngliche Nachricht-----
> >>Von: Adomas Urbanavicius [mailto:adomas@i...]
> >>Gesendet: Mittwoch, 1. Dezember 2004 15:36
> >>An: firebird-support@yahoogroups.com
> >>Betreff: Re: AW: [firebird-support] Firebird in automation
> >>
> >>
> >>Well,
> >>
> >>It is interesting issue : if there are 20 connections, all
> >>are writing 50 records/sec, it would be much data safer than
> >>1000 records/sec with single connection, wouldnt it ?
> >>
> >>
> >>
> >>
> >
> >No, because if all runs quite symmetrical, for each connection it

would take exactly one second to write the 50 records. If the server
crashes, the transaction of each connection is rolled back:

> >
> >
> >
> Ok then (About forced wrties - I agree absolutely.), lets make 20
> threads to write 200 records each, and lets make writes and comits

every

> 2secs, that would be 4000 recs, meantime, data would be only

2000recs,

> what means, that eveything will be ok, que every 2 secs will become

empty.

> Adomas