home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!axion!bear!jcs
- From: jcs@zoo.bt.co.uk (John C Sager)
- Newsgroups: comp.protocols.time.ntp
- Subject: Re: churchy la morte
- Message-ID: <1992Nov13.093643.7222@axion.bt.co.uk>
- Date: 13 Nov 92 09:36:43 GMT
- References: <9211061020.aa13174@huey.udel.edu>
- Sender: jcs@bear (John C Sager)
- Distribution: inet
- Organization: British Telecommunications plc
- Lines: 21
-
- It's a pity that churchy is off the Internet - I hope only temporarily -
- as it is the only other GPS-referenced stratum 1 server on the Internet
- that I know of. (Any others not in clock.txt yet?). It has been instructive
- to look at the declared offsets that my server, bear.zoo.bt.co.uk,
- quotes for churchy, knowing that the true offset is in the low microsecond
- range. I almost always got offsets at least in the tens of millisecond
- range & often above 100ms - and almost always negative. This is directly
- attributable to asymmetric propagation delay on the Internet.
- Plotting the raw samples over a period to get the 'arrowhead' diagram shown
- in rfc1128 gave me a fairly well-defined offset at the arrow point of about 1ms,
- which is probably the inherent asymmetric delay not caused by queueing in
- routers. There were far more points on the negative side of the arrowhead
- than on the positive side, which shows that there is an asymmetric bottleneck
- somewhere on the route from me to the US.
-
- Come back soon...
-
- John C Sager Mail: B67 G18, BT Labs
- Email: jcs@zoo.bt.co.uk Martlesham Heath
- Tel: +44 473 642623 IPSWICH IP5 7RE
- Fax: +44 473 637614 England
-