Globally optimal distributed synchronous batch reconfiguration for efficient hazard-free dynamic provisioning: How an entire network can #x201C;think globally and act locally#x201D;

摘要

This paper proposes an alternative concept for handling dynamic provisioning changes and ongoing network reconfiguration and reoptimization in general. A prime motivation is concern about the dependency of existing concepts for dynamic provisioning on the real-time coherence of databases of network state at diverse geographic locations. Not only is the continual updating of such global state everywhere in the network an intensive real-time load, but inevitable incoherencies pose hazards in network operation. The alternative proposal is a framework that makes use of precise time synchronization and the computational power of network nodes to solve identical local instances of incremental reoptimization problems in situ. The scheme removes the database coherency hazard, reduces signaling volumes, and increases resource efficiencies in service provisioning. It also provides a framework in general for continually ongoing distributed reoptimization of network configuration. The proposal also points directions towards some new research questions such as implementation of optimization models for identical solution on disparate platforms and definition of a variety of incremental batch reoptimization problems. It is thought that this may be one of the first application concepts for transport network management that would involve on-line embedded use of operations research methods in communication networks.