VT Inter-Usage Report, part I
This page provides information about the Virtual Team "Inter NGI Usage Report" aimed at providing a template for a regular (annual?) report giving an overview of the usage of the EGI infrastructure. The VT completed its task and provided the template at the end of January 2012. The template can be found here
General Project Information
- Leader: Torsten Antoni, NGI-DE/KIT
- Mailing List: firstname.lastname@example.org
- Status: Active
- Start Date: 10/Nov/2011
- End Date: Expected to finish before the end of February 2012
Providing a regular (annual?) report giving an overview of the usage of the EGI infrastructure, which could be used for promotional and mamagerial purposes
The expected output of this project is a template for the report. this will enable a check if all the information that is needed for such a report is readily available or if some effort has to be put into providing this information.
- Task 1: Provide input for the report template (all members)
- Task 2: Provide named individuals working on this(all members)
- Task 3: Consolidate the input (team leader)
- Task 4: Trigger checking of information availbility (team leader)
- NGIs: DE(Torsten Antoni
- HR (Emir Imamagic eimamagi at-nospam srce.ht
- ES (Isabel Campos isabel.campos at-nospam csic.es, Iván Díaz idiaz at-nospam cesga.es)
- LT, GR (Kostas Koumantaros kkoum at-nospam grnet.gr, Christos Papachristos <cpapachr@-nospam-ics.forth.gr>)
- IT (daniele.cesini at-nospam cnaf.infn.it, paolo.veronesi at-nospam cnaf.infn.it, alessandro.costantini_78 at-nospam yahoo.it)
- UK (john.gordon at-nospam stfc.ac.uk)
- EGI.eu: Sara Coelho, Tiziana Ferrari
- This wiki page for content collection
- Task 1: No input received so far
- Task 2: Initial members provided named individuals
- Task 3: depends on Task 1
- Task 4: depends on Task 3
Table of Contents for the EGI Usage Report
Why using EGI?
- Benefits for User Communities, NGIs and Ressource Providers
- Which services of EGI are used by users from a specific NGI
- Why does a country (resp. the researchers in that country) need EGI?
- Which consequences would researchers in a NGI have, if that country is no longer part of EGI – would they be cut off from the European Grid?
2. Scientific Use Cases
What is EGI used for?
- Indicators for collaboration through EGI: concrete examples for the added value for EGI users
- Examples of scientific results achieved through usage of EGI
3. Measuring usage
- Metrics to use
- Nationality of the User (to be defined by DN)
- CPU hours
- Rate of usage
- Active users (definition of active user: someone who submitted a job, or used a service (e.g. storage) at least once in the past 12 months)
- Caveats and limitations
- Chapter with tables and graphs, showing how the metrics defined are distributed by:
- Country / NGI
- Local vs. International Usage per NGI
- Local vs. International Usage per user nationality
- Percentage of Capacity used per NGI
- Usage by VO per NGI
- Usage of services per NGI
- Scientific field
- Resource usage by NGI per Scientific Field
- Resource usage by resource type (compute, storage) per Scientific Field
- Usage of services per Scientific Field
- Distribution of users by nationality per Scientific Field
- Resource usage by NGI per VO
- Resource usage by resource type (compute, storage) per VO
- Usage of services per VO
- Distribution of users by nationality per VO
- Country / NGI
(The idea is to answer the questions:
- Which VOs with users from a specific NGI are using EGI to what extent (number of jobs, cpu-hours)
- Which resources (e.g. percentage by country) were used by users from a specific NGI and how many (number of jobs, cpu-hours)
- How many users _not_ from a specific NGI are using the resources of that specific NGI and from which countries do these users come from?)
Analysis of commonalities, differences and interesting finding from the statistics.
Collection of questions that this report should answer. Please add your questions below:
- (A) which resources (e.g. percentage by country) were used by users from a specific NGI and how many (number of jobs, cpu-hours)
- T. Ferrari: The accounting portal already provides this information in the country view, however, currently the usage of international users is not broken down into the individual countries.
- (B) which VOs with users from a specific NGI are using EGI to what extent (number of jobs, cpu-hours)
- T. Ferrari: we can produce different views.
- country view/VOs: in the country view, the user can select one or more VOs of choice, and a pie chart is produced showing how many national users from VOs in the list are using resources of the country, and the same is displayed for the other international users by grouping them by CA
- EGI view/VOs: in the EGI view, the user can select one or more VOs of choice. An array of pie charts is produced, one per NGI, where for each NGI you get the distribution of national users/international users by NGI for the selected VOs
- how many users _not_ from a specific NGI are using the resources of that specific NGI and from which countries do these users come from?
- T. Ferrari: is the actual number of users relevant?
- indicators for collaboration through EGI: concrete examples for the added value for EGI users
- T. Ferrari: the difficult bit is to find good indicators... I have no ideas so far
- which consequences would researchers in a NGI have, if that country is no longer part of EGI – would they be cut off from the European Grid?
- why does a country (resp. the researchers in that country) need EGI?
- which services of EGI are used by users from a specific NGI ?
- T. Ferrari: which services do you have in mind? the EGI.eu central ones or the NGI resource centres?
Input from CH:
Input from HR:
Input from ES:
- Each country should provide a publication list and a list of projects, funded in a peer-reviewed process in each country, whose work is being carried out using the Grid infrastructure. See for example http://www.es-ngi.es/proyectos.php
Input from LT:
Input from IT:
Input from EGI.eu: