The $1 million recall

摘要

This story began in 2003 when a 1000W power module, which just a day before was part of a huge optical-telecom network, came back from the field heavily burned, interrupting a network that transmitted many gigabits per second. The issue immediately hit the radar of the highest management. I delved into a world of telecom power modules, which "must" work forever. To achieve forever, the power distribution is made redundant: Although one unit can do the job, two power modules power the same rack in case one fails and needs replacement, which would be performed by so-called hot-plugging while the rack is under power.