Hi Dan (et all),

you know that the customers are always unsatisfied....:-(

The first important stuff would be to have in place the metrics of consumption of the
resources some hoster "sell" to a customer. So usage metrics on
CPU / RAM /  DISK / Network (over provisioned or not is another story)
There is an aspect on actual usage versus provisioned usage.
{vm provisioned 10 GB of RAM bus os+app uses just 5 GB of them etc)
But let's keep it simple.

Should we have the usage reference then a GUI that the can define
the cost plans, would be ideal. A way to query if possible that
database, would ease existing ERPs and invoicing software to
produce / send invoices. This is for mostly post paid use cases.

For prepaid situations, there must be some regular accounting that
will stop the service through our current API. It sounds a bit scaring to
actually shutdown the provisioned vm. So at least on first step stop the
networking ....
I guess that the base is to create a separate database that will keep the
consumed metrics, so that customers could actually use it to
enforce some business logic. (as explained above).
Should there be a minimal gui that could a small hoster create
his owns plans and perhaps issue some "invoice" would be ideal.
Since there are a number of open source billing systems out there,
perhaps the integration is not that complicated. A second thought would
be to extend the data ware house gui to create the "invoice" outputs
as most of the usage info is there.
I know that ovirt is supposed to be a virtualization solution and not
a full monty billing system.

Thanks for reading so far.

Nikos


Nikos

########################################3
Zaharioudakis Nikos, RHC{A,DS,E,VA,X,I}, VCP(4,5},VCI, Mentor VCI, Zimbra Instructor
https://www.redhat.com/wapps/training/certification/verify.html?certNumber=100-001-262&isSearch=False&verify=Verify
Public  Calendar : https://www.google.com/calendar/embed?src=nzahar%40gmail.com&ctz=Europe/Athens
+30 694 720 40 63
http://zimbra.wikidot.com/zimbra-installations-in-greece

2014-11-20 15:08 GMT+02:00 Dan Kenigsberg <danken@redhat.com>:
On Thu, Nov 20, 2014 at 02:20:46PM +0200, Nikos Zaharioudakis wrote:
> Brilliant,
>
> Thanks for the info gents
> Looking forward for 3.6 then :-)

Just waiting may not be enough: I'd love to understand what are the
metrics you'd like to use for billing.

I mentioned cummulative values of network, storage, and cpu per VM. Can
you consider anything else? The latter two are begging for an RFE to be
open, as well as the report-centric RFE requested by Shirly.