The IBM POWER6™ Availability strategy is deeply rooted in an extensive history spanning multiple decades of mainframe development. By leveraging IBM’s extensive background in Predictive Failure Analysis™ and dynamic system adaptation, the Availability team has helped to create a unique processor that unleashes significant value to the client.
A cornerstone of IBM’s POWER6 Availability strategy is the ability to perform effective Predictive Failure Analysis, or PFA. IBM’s extensive First Failure Data Capture (FFDC) design, long an integral component of the POWER™ server line, enables an unprecedented level of self-awareness in POWER6 processor-based systems. The FFDC method allows the system to proactively analyze situations that indicate an impending failure, in many cases removing the questionable component from use before it can cause a threat to system stability.
IBM POWER6 processor-based systems have a number of new features which enable them to dynamically adjust when issues arise that threaten availability. Most notably, POWER6 processor-based systems introduce the POWER6 Processor Instruction Retry suite of tools, which includes Processor Instruction Retry, Alternate Processor Recovery, Partition Availability Prioritization, and Single Processor Checkstop. Taken together, in many failure scenarios these features allow a POWER6 processor-based system to recover transparently without an impact on a partition using the core.