Which of the following would be the BEST approach to keep a single controller storage subsystem on a single fabric from being a single point of failure?