[Openflow-dev] Re: usability of Zopedb for production?

[Paolo B. <p.b...@ic...>]

On Monday 12 March 2001 16:33, you wrote:
> Hi Paolo,
>
> I quickly scanned through Clemente's thesis and it seems that OpenFlow puts
> all data in Zopedb.  I thought Zopedb had the need of requent compacting if
> a lot of data is written (because of the undo support...).  How well is
> ZopeDB suited for a production environment of workflow?
>
> ciao

Good question.

My though is that this could be a minor problem, for three reason:

1) our objective is to put AS LITTE AS POSSIBLE in the workflow itself. We 
should be able to borrow data from our environment. The Zope choice was 
obvious for this.

2) The data itself are the tokens created and moved during the execution of a 
single workflow. This data should be limited.

3) I have really not considered performance/scalability of the whole system. 
My first objective is to have something that can do what we need in terms of 
functionalities. It could be (almost) trivial to implement things using 
something like a SQL db for token/data management. However, first we have to 
understand ho we can maximaze the usage of Zope features insideOpenFlow.

Ciao

Paolo


-- 
Paolo Bizzarri
Icube S.r.l.