Definition of Failover in Network Encyclopedia.
What is Failover?
Failover is any technology that allows one device to take over for a similar device that has failed. An example of a system that takes advantage of failover technology is clustering. Failover is also often used to ensure that a break in a communication line doesn’t cause a break in communication between networked systems.
How it works
Suppose you have a cluster that has two nodes, each containing different resources. If Node B experiences failure, failover occurs and the workload of Node B (its set of resources) is transferred to Node A. In Microsoft’s clustering services (Microsoft Cluster Server for Microsoft Windows NT 4 Enterprise Edition and the Cluster service for Windows 2000 Advanced Server), the cluster resources (network applications, data files, and other tools installed on the nodes of the cluster) provide services to clients on the network. A resource can be hosted on only one node at any given time, but by using the Cluster Administrator program you can configure the resource to fail over to the second node if the first node fails. This causes the resource and its operation to move from one node to the other if a failure of one node occurs.
Failover is initiated automatically by the Cluster service when a failure is detected on one of the nodes. This process can take up to 10 seconds to initiate. Failover is transparent to the users if they are accessing cluster resources using stateless protocols such as Hypertext Transfer Protocol (HTTP), and does not require any special client software to be installed on users’ machines. If a client is connected using a tool such as Windows Explorer, it will be notified that the connection is unavailable. The user should abort, retry, or cancel the connection attempt. (To connect to the resource on the failover node, retry the connection attempt.) For other Cluster service applications, users might have to log on again to the resource.
NOTE
Microsoft’s clustering services support failover at the level of virtual servers, which means items such as Web sites, print queues, file shares, and applications can be protected from system failure.
See also:
