Re: [Health] A question about GNU Health scalability, logging and extensibility

[Luis Falcon]

Hi Bashar !

On Sun, Nov 27, 2011 at 12:24 PM, Bashar <address@hidden> wrote:
> Hi all, it's me again folks :D
>
>>>(for large institutions we recoommend using an OLAP GNU
>>>Health Server ), type of records (if you are storing images instead of
>>>linking them... ).
>
> Hello Mr.Luis, how are you then?
> I didn't understand this, hence, would you please provide me with some
> explanations about what you've said in these lines ?.....
>
Yes. Basically these are just examples of how to distributing the load.
* Imaging : We have two ways of dealing with images - or other documents - .
a) We can store the images as an object in the current file system (or
even as a database object)
b) We can just make a link to the filesystem where the image or set of
images (DICOMDIR for instance) resides.

I like more option "b", since you don't duplicate information and make
easier the administration tasks

* Reporting : There are different types of reporting and statistics
that you can run on GNU Health. For example, some reporting are part
of the daily processes and tasks (prescription process, patient
clinical history... ). This can be run on the same database, since
they don't really go through too many records, and so, there is little
impact on the running system.
Then, we go to epidemiology and other "BI" processes. These processes
can go through millions of records, and so, they should be run on a
different database system, through an OLAP engine. The database is
configured in a particular way for this type of process. There are
free software extractors and BI engines, and Sebastián Marró is
working now on an native module for GNU Health.
So, in this way, the daily process and the health information system
run in parallel, and offer the best scenario.

Best regards
Luis

> Best Regards
>
> Bashar
>



-- 
Luis Falcon
GNU Health
http://health.gnu.org