Validators run in multiples, for HCC for instance and C4CW probably too, and yes multiple run on the same server, but [guess] more likely is the central scheduler the bigger constrained in the catch up process [checks such as if needing to send out an extra copy]... validations being queued so to speak. The amount of catch up was not huge [about 7/24th of 54,805 results of Friday], but it had to be fitted inside the rest of the production process with just 1 hour before midnight. For myself, 1 did not make it to the 00:00 UTC mark, so it's counted for Saturday, showing as spill in the statistics history, which was about 1 year higher than normal for the total (usually about 1.6-1.9 years):

02/11/2012 2:357:03:26:17 3,375,198 6,715
02/10/2012 402:212:03:16:18 447,969,481 863,279

--//--