Introduction; Ras Enhancements Of Power8 Processor-Based Servers - IBM Power Systems E870 Technical Overview And Introduction

5137ch04.fm

4.1 Introduction

The POWER8 processor-based servers are available in two different classes:
Scale-out systems: for environments consisting of multiple systems working in concert. In
such environments, application availability is enhanced by the superior availability
characteristics of each system.
Enterprise systems: for environments requiring systems with increased availability. In such
environments, mission-critical application can take full advantage of the scale-up
characteristics, increased performance, flexibility, and enterprise availability
characteristics of each system.
One key differentiator of the IBM POWER8 processor-based servers is that they leverage all
the advanced RAS characteristics of the POWER8 processor through the whole portfolio,
offering reliability and availability features that often are not seen in other scale-out servers.
Some of these features are improvements for POWER8 or features that were found
previously only in higher-end Power Systems.
The POWER8 processor modules support an enterprise level of reliability and availability. The
processor design has extensive error detection and fault isolation (ED/FI) capabilities to allow
for a precise analysis of faults, whether they are hard or soft. They use advanced technology,
including stacked latches and Silicon-on-Insulator (SOI) technology, to reduce susceptibility to
soft errors, and advanced design features within the processor for correction or try again after
soft error events. In addition, the design incorporates spare capacity that is integrated into
many elements to tolerate certain faults without requiring an outage or parts replacement.
Advanced availability techniques are used to mitigate the impact of other faults that are not
directly correctable in the hardware.
Features within the processor and throughout systems are incorporated to support design
verification. During the design and development process, subsystems go through rigorous
verification and integration testing processes by using these features. During system
manufacturing, systems go through a thorough testing process to help ensure high product
quality levels, again taking advantage of the designed ED/FI capabilities.
Fault isolation and recovery of the POWER8 processor and memory subsystems are
designed to use a dedicated service processor and are meant to be largely independent of
any operating system or application deployed.
The Power E870 and Power E880 are enterprise class systems and are designed to support
the highest levels of RAS. More details of the specific features exclusive to enterprise class
systems can be found in 4.4, "Enterprise systems availability details" on page 156.

4.1.1 RAS enhancements of POWER8 processor-based servers

Several features were included in the whole portfolio of the POWER8 processor-based
servers. Some of these features are improvements for POWER8 or features that were found
previously only in higher-end Power Systems, leveraging a higher RAS even for scale-out
equipment.
Here is a brief summary of these features:
Processor Enhancements Integration
POWER8 processor chips are implemented using 22 nm technology and integrated on to
SOI modules.
146
IBM Power Systems E870 and E880 Technical Overview and Introduction
Draft Document for Review October 14, 2014 10:19 am