Difference between revisions of "24.05.2013 Accounting model and portlet requirements"
Andrea.manzi (Talk | contribs) |
Andrea.manzi (Talk | contribs) (→Resource Accounting Model) |
||
| Line 9: | Line 9: | ||
| − | * objective of the model : make possible for administrators at each level ( Infra/VO/VRE) to understand both the | + | * objective of the model : make possible for administrators at each level ( Infra/VO/VRE) to understand both the resources to account to a particular user and the usage trends |
=== Job Execution === | === Job Execution === | ||
| Line 48: | Line 48: | ||
it should be generalized ( serviceclass/servicename => consumerId, targetFile=> targetResource) | it should be generalized ( serviceclass/servicename => consumerId, targetFile=> targetResource) | ||
| − | |||
== Actions == | == Actions == | ||
Revision as of 11:27, 24 May 2013
Details
- Description: Meeting notes on Resource Accounting
- Date: 24-May-2013 from 10:00 to 12:15
- Topics: Resource accounting model
- Participants: Andrea Manzi (CERN), Ermanno Travaglino (ENG), Pasquale Pagano (CNR), Luigi Fortunati (CNR).
Resource Accounting Model
- objective of the model : make possible for administrators at each level ( Infra/VOVirtual Organization;/VREVirtual Research Environment.) to understand both the resources to account to a particular user and the usage trends
Job Execution
we should include the requirements of other services running jobs on the infrastructure (Aquamaps/Statistical Manager) and in case adapt the model
instead of plan/ExecutionEngine we should put Job/task
we have to remove any reference to specific Services
to add jobId in the Execution Engine.( task)
move the n core, n cpu to Task level
we need to have a new record to be published by the portal in order to map the user with the reference to the operation/job he performed.
Service
the model is fine as it is now.
Data Access
we should include also SPD ( not only Tree Manager), but we should also think about the Geonetwork access ( trough our Geonetwork libraries)
we should calculate the quantity of data accessed in terms of kb.
we should include the type of data accessed ( geo/biodiversity)
we don't need a record for data transfer (it's a combination of data access and Storage records)
Storage
we should use the same model also for timeseries/services using DBs backend
we have to include also the UPDATE as type of operation
there is also the need to have a summary of the storage used for each user/service, to be filled service side periodically. a new record operation should be used for this (summary)
it should be generalized ( serviceclass/servicename => consumerId, targetFile=> targetResource)
Actions
Ermanno to update the wiki based on the discussion
Andrea to contact CNR to understand possible new accounting requirements for Statistical Manager
Lino/Luigi to understand accounting possible new requirement for SPD/Storage/timeseries
Ermanno to check with Fabio the inpact of the model changes to tree manager
next phone call for Accounting portlet use case definition is scheduled for Tuesday 28th at 14:30
