• Technology
  • Technology providers
  • Assessment
  • Technology roadmap
• Operations
  • Operations architecture
  • Figures and utilisation
  • Availability and reliability
• Federated cloud
  • EGI federated cloud task force

Availability and reliability

The quality of operational services delivered by Resource Centres is measured with availability and reliability metrics, computed from the results of periodic tests performed at all certified centres.

• Availability of a service or a site represents the percentage of time that the services or sites were up and running ([uptime / total time]x100).

• Reliability of a service or a site represents the percentage of time that the services or sites were supposed to be up and running, excluding scheduled downtime for maintenance and other purposes ([uptime / (total time – scheduled time)]x100).

For example: During a given 10 days, site A registered 90% availability, which means its services were up for the equivalent of 9 days. But reliability for this period was higher at 95%. This means that part of the time that the services were not available was due to scheduled maintenance work. Thus while the service was available for 90% of the time, its users could rely on it for 95% of the ten days.

Availability and reliability are monitored hourly through the Service Availability Monitoring framework (SAM), and reports are generated monthly. Certified Resource Centres guarantee 70% availability and 75% reliability for their services. The minimum availability and reliability values accepted for a Resource Centre are defined in Operational Level Agreements established with EGI.eu.