The High Availability subsystem executes on all nodes, detecting and reacting to hardware and software failures. Heartbeats are exchanged between nodes to evaluate membership changes and initiate appropriate recovery. Subsystem-specific monitors are used to detect software failures.

High Availability in the Scalable Video and Web Server

Upon node failures, server node IP addresses are automatically taken over, Round Robin DNS updated, and TCP Routers automatically switched over/reconfigured, so that failures are transparent to all external clients.

Recoverable VSD can be used to handle VSD server failures, and the Tiger Shark file system uses replication of disk blocks in order to continue operation of video stream playback upon disk failures. Calypso recovery can be used in the case of Calypso server failure.

Rajat Mukherjee <rajatm@watson.ibm.com>

[ IBM Research home page ]

[ IBM home page | Order | Search | Contact IBM | Help | (C) | (TM) ]