Cost-effectivenesss of highly-accurate clocks for NTP

[bms at fastmail.net (Bruce Simpson)]

On 13/05/16 20:39, Eric S. Raymond wrote:
> In 2012, nearly three years before being recruited for NTPsec, I
> solved this problem as part of my work on GPSD.  The key to this
> solution is an obscure feature of USB, and a one-wire
> patch to the bog-standard design for generic USB that exploits
> it.  Technical details on request, but what it comes down to is
> that with this one weird trick(!) you can mass-produce primary time
> sources with a jitter bounded by the USB polling interval for
> about $20 a pop.
>
> The USB 1 polling interval is 1ms.

What about USB 3.1 (assuming the device is not intended to be backwards 
compatible with the polling model) ? I should point out Intel intend to 
retire EHCI/UHCI and implement only xHCI.