A brief history of NTP time - VMS

This is a discussion on A brief history of NTP time - VMS ; For those interested by time management (NTP, et al), I suggest reading the paper "A brief history of NTP time" at http://www.cis.udel.edu/~mills/datab...rs/history.pdf . It has nice references to Alpha systems being used systematically as time keepers because of the stability ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: A brief history of NTP time

  1. A brief history of NTP time

    For those interested by time management (NTP, et al), I suggest
    reading the paper "A brief history of NTP time" at
    http://www.cis.udel.edu/~mills/datab...rs/history.pdf .
    It has nice references to Alpha systems being used systematically
    as time keepers because of the stability of their internal clock.
    I wonder which OS they are running and how they will be replaced
    once Alpha gets de-supported...

    --
    Marc Van Dyck



  2. Re: A brief history of NTP time

    In article , Marc Van Dyck writes:
    > For those interested by time management (NTP, et al), I suggest
    > reading the paper "A brief history of NTP time" at
    > http://www.cis.udel.edu/~mills/datab...rs/history.pdf .
    > It has nice references to Alpha systems being used systematically
    > as time keepers because of the stability of their internal clock.
    > I wonder which OS they are running and how they will be replaced
    > once Alpha gets de-supported...


    Odd, I always found VAX clocks more reliable. Most Alpha used
    a commercial clock chip that didn't run at a multiple of the
    VMS clock lsb, so Alpha had to introduce a "leap tick" to keep
    the time lsb right. I wonder which model Alpha they used and which
    clock chip it employed.

    And timekeeping in UNIX has always been hazardous below about 50
    milliseconds.


  3. Re: A brief history of NTP time

    Bob Koehler wrote:
    > In article , Marc Van Dyck writes:
    >> For those interested by time management (NTP, et al), I suggest
    >> reading the paper "A brief history of NTP time" at
    >> http://www.cis.udel.edu/~mills/datab...rs/history.pdf .
    >> It has nice references to Alpha systems being used systematically
    >> as time keepers because of the stability of their internal clock.
    >> I wonder which OS they are running and how they will be replaced
    >> once Alpha gets de-supported...

    >
    > Odd, I always found VAX clocks more reliable. Most Alpha used
    > a commercial clock chip that didn't run at a multiple of the
    > VMS clock lsb, so Alpha had to introduce a "leap tick" to keep
    > the time lsb right. I wonder which model Alpha they used and which
    > clock chip it employed.


    I have same experience. Alpha clocks needs NTP. VAX was more reliable.

    Arne

  4. Re: A brief history of NTP time

    Arne Vajh°j wrote in
    news:4788110f$0$90267$14726298@news.sunsite.dk:

    > Bob Koehler wrote:
    >> In article , Marc Van Dyck
    >> writes:
    >>> For those interested by time management (NTP, et al), I suggest
    >>> reading the paper "A brief history of NTP time" at
    >>> http://www.cis.udel.edu/~mills/datab...rs/history.pdf .
    >>> It has nice references to Alpha systems being used systematically
    >>> as time keepers because of the stability of their internal clock.
    >>> I wonder which OS they are running and how they will be replaced
    >>> once Alpha gets de-supported...

    >>
    >> Odd, I always found VAX clocks more reliable. Most Alpha used
    >> a commercial clock chip that didn't run at a multiple of the
    >> VMS clock lsb, so Alpha had to introduce a "leap tick" to keep
    >> the time lsb right. I wonder which model Alpha they used and
    >> which clock chip it employed.

    >
    > I have same experience. Alpha clocks needs NTP. VAX was more reliable.
    >
    > Arne
    >


    I seem to remember reading somewhere, there were some add-in (PCI slot)
    clocks for Alpha that were very accurate. Can anyone confirm this?

+ Reply to Thread