said by MaynardKrebs:So down from 100 points of failure to 15.
Next upgrade will be down to a single point of failure.
Too bad there most likely isn't any load-sharing/balancing in there. If there was, losing 1/15th of available capacity should have almost no impact on service.
With the much more efficient "packing" by aggregating everything on 10G links instead of individual 1Gbps virtual circuits, TSI would likely be able to get away with somewhat fewer 10G links than what their projected total of 1G links would dictate. They only need to aim for equal or more if most of their 1G links happen to frequently hit their peaks at the same time, which should be very uncommon.