No matter how you slice it, as I see it, it boils down to: Microsoft's software did insufficient input validation on untrusted data (specifically in this case configuration changes for one customer, it seems), and this led to widespread outages affecting many, many, *many* customers in many different ways.Security done properly should not rely on "blocking erroneous" stuff, but should verify that everything is good and only *if* that succeeds allow any of it through.@spytfyre @rysiek