Principles

• ôMission-criticalö attitude

• Assume many HW & SW failures

  • Graceful degradation (not perfection)
  • Decide on ôfield-replaceable unitö
  • Assume server ôpollutionö over time

• Minimize human involvement

• Focus on ôhot-swap everythingö

Previous slide

Next slide

Back to first slide

View graphic version