[Openstack] Metering question??
Hi all, I was just looking over the efficient metering stuff yesterday. Just a couple of questions, that might be dups (sorry if they are). I am noticing that there seems to be a mix of billing specifics there and metering specifics there. If say metering can just provide as much raw data as possible then is that not sufficient? Shouldn't that be the limit of openstack's job? Then there can be sets of business units or other people that figure out how to carve up that raw data into billing equivalents. This is especially since every business will have there own idears around billing. In certain situations, I can even think where one would want to use hadoop to calculate some advanced model of all that raw-data and use that for billing, but the key part that keeps on coming up in my brain is just get as much raw data as possible. Thoughts?? -Josh ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Metering question??
On Wed, May 2, 2012 at 2:31 PM, Joshua Harlow harlo...@yahoo-inc.comwrote: Hi all, I was just looking over the efficient metering stuff yesterday. Just a couple of questions, that might be dups (sorry if they are). I am noticing that there seems to be a mix of billing specifics there and metering specifics there. If say metering can just provide as much raw data as possible then is that not sufficient? Shouldn’t that be the limit of openstack’s job? I believe the rest of the group is reaching that conclusion, too. There was some discussion of that in the recent schema and counter definitions thread. Then there can be sets of business units or other people that figure out how to carve up that raw data into billing equivalents. This is especially since every business will have there own idears around billing. In certain situations, I can even think where one would want to use hadoop to calculate some advanced model of all that raw-data and use that for billing, but the key part that keeps on coming up in my brain is just get as much raw data as possible. Thoughts?? -Josh ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Metering question??
Agree with your comments, 1. Metering tasks for me are mainly 2: a) Collect relevant data b) Correlate events (that's which end event corresponds with which event start) this is very important, since only openstack knows how to correlate, but the billing system 2. Mediation process should exist, one per billing system (a.k.a driver, openstack-specific-billing-system-plugin) Map previously correlated data to a pricing model 3. Billing system tasks (obviously outscope openstack, it's does know nothing about cloud) Regards On Thu, May 3, 2012 at 12:10 AM, Doug Hellmann doug.hellm...@dreamhost.comwrote: On Wed, May 2, 2012 at 2:31 PM, Joshua Harlow harlo...@yahoo-inc.comwrote: Hi all, I was just looking over the efficient metering stuff yesterday. Just a couple of questions, that might be dups (sorry if they are). I am noticing that there seems to be a mix of billing specifics there and metering specifics there. If say metering can just provide as much raw data as possible then is that not sufficient? Shouldn’t that be the limit of openstack’s job? I believe the rest of the group is reaching that conclusion, too. There was some discussion of that in the recent schema and counter definitions thread. Then there can be sets of business units or other people that figure out how to carve up that raw data into billing equivalents. This is especially since every business will have there own idears around billing. In certain situations, I can even think where one would want to use hadoop to calculate some advanced model of all that raw-data and use that for billing, but the key part that keeps on coming up in my brain is just get as much raw data as possible. Thoughts?? -Josh ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp -- --- Luis Alberto Gervaso Martin Woorea Solutions, S.L CEO CTO mobile: (+34) 627983344 luis@ luis.gerv...@gmail.comwoorea.es ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Metering question??
I just upload a video that shows the way i manage the billing described in my previous mail in order to clarify: http://youtu.be/3A1SdZS9Iak Data is gathered and correlated in a own metering agent that i want to conform with the ceilometer spec. Data is submited to the billing system. The billing process then can be run on demand or scheduled. Regards On Thu, May 3, 2012 at 3:01 AM, Luis Gervaso l...@woorea.es wrote: Agree with your comments, 1. Metering tasks for me are mainly 2: a) Collect relevant data b) Correlate events (that's which end event corresponds with which event start) this is very important, since only openstack knows how to correlate, but the billing system 2. Mediation process should exist, one per billing system (a.k.a driver, openstack-specific-billing-system-plugin) Map previously correlated data to a pricing model 3. Billing system tasks (obviously outscope openstack, it's does know nothing about cloud) Regards On Thu, May 3, 2012 at 12:10 AM, Doug Hellmann doug.hellm...@dreamhost.com wrote: On Wed, May 2, 2012 at 2:31 PM, Joshua Harlow harlo...@yahoo-inc.comwrote: Hi all, I was just looking over the efficient metering stuff yesterday. Just a couple of questions, that might be dups (sorry if they are). I am noticing that there seems to be a mix of billing specifics there and metering specifics there. If say metering can just provide as much raw data as possible then is that not sufficient? Shouldn’t that be the limit of openstack’s job? I believe the rest of the group is reaching that conclusion, too. There was some discussion of that in the recent schema and counter definitions thread. Then there can be sets of business units or other people that figure out how to carve up that raw data into billing equivalents. This is especially since every business will have there own idears around billing. In certain situations, I can even think where one would want to use hadoop to calculate some advanced model of all that raw-data and use that for billing, but the key part that keeps on coming up in my brain is just get as much raw data as possible. Thoughts?? -Josh ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp -- --- Luis Alberto Gervaso Martin Woorea Solutions, S.L CEO CTO mobile: (+34) 627983344 luis@ luis.gerv...@gmail.comwoorea.es -- --- Luis Alberto Gervaso Martin Woorea Solutions, S.L CEO CTO mobile: (+34) 627983344 luis@ luis.gerv...@gmail.comwoorea.es ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
