[ntp:questions] Clock time synchronization of four computers

Nazmul Islam mnislam at winlab.rutgers.edu
Tue Jul 10 16:01:13 UTC 2012


Thanks a lot for all of your responses. I really appreciate your spending
time in this matter.

It seems like that my organization's traffic controller was blocking all
NTP requests. I contacted them and was able to use the organization's NTP
server. Now, I only see a millisecond level difference between the
computers.

I plan to use Paul Kennedy's suggestions if I need more accuracy. Thanks
for the detailed instructions.

----

Nazmul

On Tue, Jul 10, 2012 at 1:37 AM, unruh <unruh at invalid.ca> wrote:

> On 2012-07-09, Dave Hart <hart at ntp.org> wrote:
> > On Mon, Jul 9, 2012 at 21:00 UTC, unruh wrote:
> >> If he is having trouble keeping many second sync with pool servers, he
> >> is going to have trouble with keeping sync with a master. And if that
> >> master goes out for some reason, he will be totally up the creek.
> >
> > There's a big difference between synchronization across a LAN you
> > control vs an internet connection.  For keeping a set of machines
> > close, a local master setup will keep them closer than each hitting
> > the pool.  You don't have to have a single point of failure.  You can
> > use just about any hierarchy you want, up to a self-healing
> > orphan+manycast mesh.
>
> While I would agree if his network error were a few ms, his claim is
> that it is a few seconds, and tht I find very hard to believe, unless he
> is on a 300 bd modem with a download going on as his network connection.
>
> And how is each computer connected to a different set of pool servers a
> "single point of failure" unless you mean the lan connection (and we do
> not know if that is single point). Anyway, a single master is also a
> "single point of failure."
>
> >
> > Cheers,
> > Dave Hart
>
> _______________________________________________
> questions mailing list
> questions at lists.ntp.org
> http://lists.ntp.org/listinfo/questions
>



-- 
Muhammad Nazmul Islam

Graduate Student
Electrical & Computer Engineering
Wireless Information & Networking Laboratory
Rutgers, USA.


More information about the questions mailing list