Domain-wide scheduling of OSPF graceful restarts for maintenance purposes

The maintenance of communication systems is a critical operation which monopolizes human and network resources. The management of multiple, parallel maintenance jobs is a complex task that can generate faults and undesirable service interruption. In the context of IP/MPLS networks, Graceful Restart mechanisms allow, under strict conditions, the maintenance of a single router without impacting its forwarding plane. Still, the network-wide coordination of the routers restarts is an unresolved problem. To solve this issue, we propose a solution for the automation and optimized orchestration of maintenance operations, based on an adaptive and fully-distributed planning process. The operator is then relieved from tedious maintenance tasks and is only responsible for setting performance objectives and assessing progress reports.