US 2008/0177741 A1 · Maintaining item-to-node mapping information in a distributed system

Abstract

A method and apparatus for maintaining an item-to-node mapping among nodes in a distributed cluster is provided. Each node maintains locally-stored system-state information indicating that node's understanding of which master nodes are alive and dead. Instead of employing a global item-to-node mapping, each node acts upon a locally determined mapping based on its locally-stored system-state information. For any two nodes with the same locally-stored system-state information, the locally determined mapping is the same. A node updates its locally-stored system-state information upon detecting a node failure or receiving a message from another node indicating different locally-stored system-state information. The new locally-stored system-state information is transmitted on a need-to-know basis, and consequently nodes with different item-to-node mappings may operate concurrently. Mechanisms to avoid nodes assuming conflicting ownership of items are employed, thus allowing node failures to propagate via asynchronous messaging instead of requiring a cluster-wide synchronization event.

Claims46

Dependent claims

A computer-implemented method for maintaining an item-to-node mapping among a plurality of nodes in a system, comprising the computer-implemented steps of:

maintaining, at a first node, first data indicating state information for each node in a set of nodes;
wherein the state information in the first data indicates, for each node in said set of nodes, the first node's understanding of whether said each node is an active master node;
determining, at the first node, a first item-to-node mapping based on said first data;
receiving, at the first node from a second node, second data indicating state information for each node in said set of nodes;
wherein the state information in the second data indicates, for each node in said set of nodes, the second node's understanding of whether said node is an active master node;
wherein the first data indicates that a particular node is an active master node;
wherein the second data indicates that said particular node is not an active master node;
in response to receiving the second data, the first node updating the first data to generate updated first data that indicates that said particular node is no longer an active master node;
and
after updating said first data, the first node determining a second item-to-node mapping based on said updated first data.

7 more elements

A computer-implemented method for maintaining an item-to-master-node mapping among a plurality of nodes in a system, comprising the computer-implemented steps of:

using a first item-to-master-node mapping at a first node based on the first node's understanding of which nodes within the system are currently active master nodes;
using a second item-to-master-node mapping at a second node based on the second node's understanding of which nodes within the system are currently active master nodes;
wherein the first item-to-master-node mapping and the first node's understanding of which nodes are currently active master nodes are out-of-date;
wherein the second item-to-master-node mapping and the second node's understanding of which nodes are currently active master nodes are not out-of-date;
and
wherein the first node uses the first item-to-master-node mapping concurrently with the second node using the second item-to-master-node mapping.

3 more elements

Dependencies are inferred from the claim text.

Maintaining item-to-node mapping information in a distributed system

Life of this patent

Maintaining item-to-node mapping information in a distributed system

Life of this patent

Abstract

Claims46