Sorry, maybe I got carried away with the tone. But it is not sarcasm. I genuinely did not realise that the NTP service level was so low. There are two problems raised in the email: There is no on-site redundant fail-over upstream of the NTP servers. All NTP servers at the site were not automatically taken down immediately upon detection of the fault (because some were still, in some sense, within tolerance). This places all of the fault management onto downstream NTP servers. I honestly expected NIST to be running a robust cross-site timebase upstream of NTP.
Comments
Is this sarcasm? I can't tell.
Per the email:
Sorry, maybe I got carried away with the tone. But it is not sarcasm. I genuinely did not realise that the NTP service level was so low. There are two problems raised in the email: There is no on-site redundant fail-over upstream of the NTP servers. All NTP servers at the site were not automatically taken down immediately upon detection of the fault (because some were still, in some sense, within tolerance). This places all of the fault management onto downstream NTP servers. I honestly expected NIST to be running a robust cross-site timebase upstream of NTP.