The EGEE project officially ended on the 31 March 2006
EGEE II started on 1 April 2006 and the new EGEE website can be found at:
http://www.eu-egee.org
What is a Regional Operations Centre (ROC)?
The Regional Operations Centres are the heart of the operational support for the Grid infrastructure. They have a key role as sources of expert advice and technical support in the process of building and operating the infrastructure. The ROCs will have the expertise to assist the growing number of Resource Centres in joining the infrastructure, through the deployment of middleware releases and the development of procedures and capabilities to operate those resources. The development of deployment procedures and documentation will require some effort, and each initial deployment at a new Resource Centre will require close support, including on-site assistance where required. This support is provided by the ROC team.
The middleware engineering activity will periodically deliver new releases for deployment across the infrastructure at both Resource Centres and in the CICs. The ROCs have the responsibility for acceptance testing of the new middleware on a variety of platforms before deployment. This activity is done in conjunction with the certification team in the OMC at CERN. Once validated against test suites produced in collaboration with the certification team and the applications groups, the Resource Centres and Core Infrastructure services will be upgraded in a co-ordinated way to produce minimal disruption to operational Grid services. The ROC is the coordination for the deployment of middleware releases within its region, providing the point of contact for the Resource Centre and its main line of support. Timing of installation at the Resource Centres is the ROC responsibility in concertation with the OMC.
Once a Resource Centre is up and running, the relevant Regional Operations Centre will provide first-line support in resolving any operational problems originating at the Resource Centres. The ROCs will have sufficient expertise and information on the Grid operational state to diagnose problems as originating with operation of a resource, of an infrastructure service, or with the Grid middleware itself. In many cases, the problem will either be one already encountered and for which a known solution is available, or resolvable by the ROCs themselves. If problems cannot be easily solved, ROCs will refer problems onwards to the Core Infrastructure Centres. Once a problem is referred to an ROC, ownership rests with the ROC to ensure that it is resolved where possible and that the originator of the problem is kept informed of its status.
The Regional Operations Centres will also provide first-line support to Grid users to resolve infrastructure problems; i.e. those arising from operations and middleware. The same problem ownership and referral to CICs (if appropriate) applies as with resource-oriented problems. User problems originating with applications per se must be identified and referred to the relevant user community and/or applications development; larger and more sophisticated user organizations are likely to have their own support structures which can make the initial problem triage. On a less immediate level, applications issues may be pursued with the assistance of the application development support described in NA4.
Under the coordination of the Operations Management Centre the ROC is responsible for negotiating service level agreements (SLAs) with each Resource Centre in the region and will include monitoring the Resource Centres to ensure delivery of the agreed services.
Each ROC will have a Centre Manager reporting to the OMC, a small deployment team and a larger team that provides 24-hour support on a shift basis. Support activities will be co-ordinated through a problem-tracking system operating across all centres.
There will be a coordinator of all the nine ROCs who will report directly to the Operations Manager, and who is responsible for organizing and coordinating the activities and operations of the ROCs.