Failure modes and recovery models

To create a highly available system, we need to consider the system's failure modes and how we'll maximize the MTBF and minimize the MTTR. One thing that won't be immediately obvious in these discussions is the implementation, but I've included an HA example in this book.