| 摘要 |
A method of load distribution for a cluster of two or more nodes. The method comprises receiving an initial request packet on a network device having a virtual IP address; forwarding the request packet from the network device to a cluster of at least two nodes, wherein each of the at least two nodes has an internal dispatcher module and an unique and non-conflicting virtual IP address; establishing one of the at least two nodes as a priority dispatcher or dispatcher endpoint, wherein if any one node fails, the virtual IP address of the one node which is no longer active falls back to another node within the cluster based on cluster priorities; dispatching the request packet to one of the nodes associated with the cluster; and forwarding the request from one of the nodes to a switching device.
|
