| 摘要 |
A method and apparatus for fault management in a network. A network architecture is disclosed including at least one network switch, several network interface cards, several processors (host) and a network manager. Fault reporting and detection logic is implemented in each switch and in each network interface card. In addition, multiple fault reporting pathways are provided for each switch and each network interface card. As a result, utilizing message exchanges such as Fibre Channel exchange messages, the switch, the network interface cards and the processors (host) are able to autonomously generate and report faults to the network manager. The combined fault reporting enables the network manager to more accurately isolate faults. In addition, because of the autonomous nature of the fault reporting, faults may be detected and corrected prior to the initiation of a communication session.
|
