Triple redundancy marks new PACS controller

January 17, 2001

A portable, scalable, and fault-tolerant PACS controller was the highlight of an infoRAD exhibit at November's RSNA show. Known as a resilient continuous availability server, RCAS features three identical Unix machines running copies of the same

A portable, scalable, and fault-tolerant PACS controller was the highlight of an infoRAD exhibit at November's RSNA show. Known as a resilient continuous availability server, RCAS features three identical Unix machines running copies of the same operating system and application software. Its novel design could allow mission-critical image vendors to advertise near 100% availability, according to Brent Lui, who developed the machine at Saint John's Health Center in Santa Monica, CA.

To ensure nearly continuous operation, data transfers are sent to all three machines, which perform special "voting" routines that determine the operational condition of each machine. Operation is then executed step-by-step at the machine with the majority vote. This guarantees that no hardware failure can stop the execution of the operation. In a demonstration at the RSNA meeting, unplugging one of the three servers during image transmission caused less than a two-second interruption.