Each software component submitted by the technology providers is assessed against a set of defined criteria before being deployed into the production infrastructure. The criteria are in public domain and they are made available to the technology providers, fostering proactive development of new software components.
New software components delivered by technology providers are assessed by an international team distributed across the several National Grid Infrastructures (NGIs) that make up EGI.
The assessment team uses the public criteria previously deemed appropriate for this specific piece of software. This first assessment phase is designed to test functionality: does the software work or not?
The team executes the quality criteria and verifies the software against them. If the software passes the functionality test, then it’s ready for the next step.
The staged rollout broadens the scope of previous testing to see how the new software fares in a production environment. This phase is managed by operations teams across the NGIs.
During the staged rollout, the new software runs in a limited area of the production infrastructure under close monitoring. The idea is to assess how well the new software behaves in reality. For example: Does it scale well? Is it reliable under load? Is its resource consumption reasonable?
If the operations teams are satisfied with the software’s performance, then the new software is published in common software repository and made available for the EGI community. This is usually broadcast to the sites in the infrastructure.
The assessment procedure aims to be transparent and auditable. Every step leaves a paper trail available to the public in the documents database.