Re: [ntp:questions] Sub-millisecond NTP synchronization for local network
Jeremy Leibs <[EMAIL PROTECTED]> wrote: > Is there possibly a way of configuring the maximum acceptable > latency of a packet? That is, as long as you know that for some > fraction of the day (when the system is not under load) your latency > is going to be less than some threshold, say, 2 ms, configuring the > system to just throw away all packets with latency greater than 2 > ms? What precisely do you mean by "the system" in this context? The TCP/IP stack running on the system on which NTP is running, or in NTP itself? rick jones -- a wide gulf separates "what if" from "if only" these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Sub-millisecond NTP synchronization for local network
[EMAIL PROTECTED] (Jeremy Leibs) writes: >On Fri, Dec 5, 2008 at 11:35 AM, Unruh <[EMAIL PROTECTED]> wrote: >> [EMAIL PROTECTED] (Kevin Oberman) writes: >> >> >> >> >I am puzzled by the reference to "rapid changes in network >> >latency". This is bad as it would seem to indicate network route >> >instability. Worse, it might mean that the path is sometimes >> >asymmetric. This is something NTPD can't handle. >> >> In his case it was because the network link is a wireless link and >> sometimes it is swamped by data being exchanged over it. There is not >> network route instability. There is simply more or less use of the network. >> Yes, ntp does not handle such things well. Its only defence is t othrow >> away 7/8 of the packets (Ie uses that packet of the last 8 which has >> minimum round trip time. ) >> >> >Is there possibly a way of configuring the maximum acceptable latency of a >packet? That is, as long as you know that for some fraction of the day >(when the system is not under load) your latency is going to be less than >some threshold, say, 2 ms, configuring the system to just throw away all >packets with latency greater than 2 ms? In chrony, yes, in ntp, not as far as I know. ntp tries to handle it with a combination of the filter algorithm (it accepts only that packet out of the last 8 that has the minimum round trip delay) and the huff and puff filter. In the OP case, I have no idea how long the clogged network lasts or what fraction of the day it occupies. If it is say a solid 1 hour out of the day, and taht only for one hour a day, then the effect on ntp should be minimal. Will a poll of 10 (1000 sec) an hour of long delays should be ignored by the filter (well, 55 min will saturate the filter and start letting in bad packets). If the delays occur in say 5 min batches, the chances are it will not affect ntp at all. ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Sub-millisecond NTP synchronization for local network
>"Unruh" <[EMAIL PROTECTED]> wrote in message >news:[EMAIL PROTECTED] >> [...] Getting 20microseconds over a local net is easy. On a local >> net, the delay should only be about 100microseconds, certainly not >> milliseconds. ... >I must be doing something horribly, horribly wrong then. Delay is >on the order of 1 millisecond (0.6-1.3) and offsets are currently >*very* good, all between -1 and 1 millisecond. OK, strange. YOu can look at www.theory.physics.ubc.ca/chrony.html to see the delays on my network-- it is local to a university building, but all over the building. And I see essentially no difference between computers on one side or the other. The delays on the 100Mbs parts of the network are on the .15ms level with offsets at the .02ms level. However, the gigabit machines are much much worse. I had not looked recently, and an a bit perturbed now with how bad they are in comparison with previous levels ( from 6 months ago). >What should I be doing to get 20 us? Buy all new computers with >gigabit Ethernet? I suspect buying better switches. And it looks to me like you should definitely NOT go to gigabit Ethernet if you want good timing. >Groetjes, >Maarten Wiltink ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Sub-millisecond NTP synchronization for local network
"Unruh" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > [...] Getting 20microseconds over a local net is easy. On a local > net, the delay should only be about 100microseconds, certainly not > milliseconds. ... I must be doing something horribly, horribly wrong then. Delay is on the order of 1 millisecond (0.6-1.3) and offsets are currently *very* good, all between -1 and 1 millisecond. What should I be doing to get 20 us? Buy all new computers with gigabit Ethernet? Groetjes, Maarten Wiltink ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] client listening interface
Danny - Yes, that is correct. I didn't see the query-on so was yapping about listen-on so you'd have a point of reference. Thanks! Phil - Original Message - From: Danny Mayer [EMAIL PROTECTED] Sent: 12/05/2008 02:56 PM EST To: Phil Newlon Cc: questions@lists.ntp.org Subject: Re: [ntp:questions] client listening interface [EMAIL PROTECTED] wrote: > I have a Windows 2000 Server system that has two IP address configured - > one a routeable address, the other is 192.168.2.x for communication with > internal devices. My time source requires communication outside the local > environment, but it would appear that the server randomly decides to only > try to synch via the 192.168 interface, and failing. I wandered through > the support.ntp.org site, and found reference to a directive "listen-on" > being discussed. Is something resembling that functionality available in > 4.2.4p5, or is the current behavior of "all interfaces" my only current > option? > > Thanks! > > Phil What you require is the query-on option. I haven't had time to roll it into a release yet. You need outbound queries to go on the routeable address, correct? listen-on is to do with addresses/interfaces that it will accept requests on. Danny ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
Re: [ntp:questions] Sub-millisecond NTP synchronization for local network
On Fri, Dec 5, 2008 at 11:35 AM, Unruh <[EMAIL PROTECTED]> wrote: > [EMAIL PROTECTED] (Kevin Oberman) writes: > > > > >I am puzzled by the reference to "rapid changes in network > >latency". This is bad as it would seem to indicate network route > >instability. Worse, it might mean that the path is sometimes > >asymmetric. This is something NTPD can't handle. > > In his case it was because the network link is a wireless link and > sometimes it is swamped by data being exchanged over it. There is not > network route instability. There is simply more or less use of the network. > Yes, ntp does not handle such things well. Its only defence is t othrow > away 7/8 of the packets (Ie uses that packet of the last 8 which has > minimum round trip time. ) > > Is there possibly a way of configuring the maximum acceptable latency of a packet? That is, as long as you know that for some fraction of the day (when the system is not under load) your latency is going to be less than some threshold, say, 2 ms, configuring the system to just throw away all packets with latency greater than 2 ms? ___ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
