Clustering is a technique of linking multiple machines to work as single logical entity. It is a basic way of scaling execution environment.

As clustered machines act as single logical entity, cluster allows to introduce:

  • high availability(HA)
  • fault tolerance(FT)

High availability is a system design and procedures which ensure continual service at desired platform. In cluster environment node failure or addition does not change availability of cluster to user.

Fault tolerance is term closely tied with HA. Fault tolerant design(or fail-safe) is design that enables system(cluster) to continue operation at reduced level, in favor of complete failure when part of system(cluster) fails. Operation may be understood as system operation or user specific operation.

Mobicents Cluster aims to provide HA/FT framework which can be reused in EEs as base for building HA/FT system. Mobicents FT/HA framework introduces following:

  • Core framework
  • FT/HA Timers
Core framework

Core framework is a base for whole solution. It defines procedures to ensure integrity of clustered data and cluster resources.

Core ensures that operations performed on clustered data have means of achieving fault tolerance.

FT/HA Timers

Mobicents Cluster framework comes with low cost HA/FT timers. Framework timer implementation ensures that timers scheduled on one cluster node can be recreated on any other node and run on it.

Mobicents Cluster