US 6,067,545 A · Resource rebalancing in networked computer systems

Abstract

Methods for load balancing a network are disclosed. Resource rebalancing includes remapping of pathways between nodes, e.g. servers, and resources, e.g. volumes/file systems. Resource rebalancing allows the network to reconfigure itself as components come on-line/off-line, as components fail, and as components fail back. In an embodiment of the invention a method for load balancing on a network is disclosed. The network includes server nodes and resources. Each of the resources are coupled to at least two of the server nodes. The method for load balancing comprises the acts of detecting a change in an availability of the server nodes; defining a first set of available server nodes and a second set of available resources and selecting for each one of the members of the second set a corresponding member of the first set to server as the administrative server for handling an administrative portion of an I/O request for the corresponding resource of the second set. In an alternative embodiment of the invention the method for load balancing comprises the act of detecting a change in an availability of the server nodes; applying a load balancing function to the network responsive to at least two attributes of each of the server nodes and the resources, responsive to the detecting act and assigning based on a result of the load balancing function each of the resources to a corresponding available server node responsive to the applying act.

Claims38

Dependent claims

A method for load balancing a network, the network including server nodes and resources, each resource coupled to at least two of the server nodes, the method for load balancing comprising:

detecting a change in an availability of the server nodes;
defining a first set of available server nodes and a second set of available resources;
allocating weights to the server nodes and to the resources;
defining an ideal capacity for each server node in terms of a total weight of resources for which each server node will handle I/O requests;
and
selecting for each of the resources in the second set a corresponding server node of the first set to serve as an administrative server for handling an administrative portion of an I/O request for the corresponding resource of the second set.

3 more elements

A method for load balancing on a network, the network including a client node, server nodes coupled to the client node, and at least a first and a second resource, each of the first and the second resources coupled to a first server node and a second server node, respectively, the resources including at least one member of the resource group consisting of storage devices and data sets, the method for load balancing comprising:

assigning weights to at least one of the first and the second server nodes, the first and the second resources, and the first and second server nodes together with the first and the second resources, the weights correspond to at least one of:
a theoretical capacity of the resource, and actual capacity of the resource, a theoretical demand for the resource, and an actual demand for the resource, a theoretical capacity of the server node, an actual capacity of the server node, a theoretical demand for the server node, and an actual demand for the server node;
and
allocating the first and the second resources to available ones of the first and the second server nodes on the basis of the assigned weights.

1 more element

A method for load balancing on a network, the network including a client node, a first and a second server node coupled to the client node, and a first and a second resource, each of the first and second resources coupled to the first server node and the second server node, respectively, the resources including at least one member of the resource group consisting of storage devices and data sets, the method for load balancing comprising:

determining weights for at least one of the first and the second server nodes, the first and the second resources, and the first and second server nodes together with the first and the second resources,
assigning for each of the first and the second resources a corresponding one of the first and second server nodes to act as preferred administrative nodes which control access by either of the first and second server nodes to the resources;
and
allocating each of the first and the second resources to available ones of the first and the second server nodes on the basis of availability, weights, and the preferred administrative server node assigned to the resources.

1 more element

Dependencies are inferred from the claim text.

Resource rebalancing in networked computer systems

Life of this patent

Resource rebalancing in networked computer systems

Life of this patent

Abstract

Claims38