Actually, on all our trading systems, our times are synced via PTP instead of ntpd for at least 50 microsecond accuracy. The stratum 1 clocks as well as NIST time are only used as comparison to verify compliance and reality. We use ntpq to determine the offset from NIST for reporting. ---- Matthew Huff | 1 Manhattanville Rd Director of Operations | Purchase, NY 10577 OTA Management LLC | Phone: 914-460-4039 -----Original Message----- From: NANOG [mailto:nanog-bounces@nanog.org] On Behalf Of Stephen Satchell Sent: Saturday, December 29, 2018 10:01 AM To: nanog@nanog.org Subject: Re: CenturyLink On 12/29/18 6:51 AM, Matthew Huff wrote:
We have two stratum-1 servers synced with GPS and a PTP feed from a provider that also provides PTP to market data systems, but we still have to monitor drift between system time and NIST time. Don't ask for the logic behind it, it's a regulation, not a technical requirement.
Having been a participant on Standards Working Groups, I understand completely. Regulations, like Standards, need to be written to apply to as wide a population as possible. Do those regulations dictate how you monitor drift? For example, can you have a system (I would use CentOS) that syncs to the appliance as well as NIST and your inside NTP sources, and use ntpq(8) to read the drift directly?