It is 1:17am and Sprint appears to be routing traffic to us again. The outage which lasted from almost exactly at midnight until about 1:10am on Monday morning, Oct 12, was the result of work Sprint was doing on "route reflectors". They did not expect the work to cause an outage due to redundancy that is built in to the system that is supposed to prevent one but an outage it did cause. They did not announce the work in advance because they did not expect an outage to result. They tell me that there is sufficient redundancy in the backbone now that normally backbone work will not cause any outages and so they do not announce it. They only announce work on parts of the infrastructure expected to result in an outage. Obviously in this case their expectations were incorrect.