DNS Amplification Attacks... and a trivial proposal

[John Levine]

>So, may I infer that rather than being put off until the end of the
>century, which seemed to be the previous implementation timeline,
>pervasive implementation of BCP 38 may now be expected at around the
>time that 32-bit UNIX clocks are anticipated to wrap-around to negative?

Perhaps, but I think that's still a lot sooner than a yet-to-be-designed
hack to DNS servers will be widely used.