On Sat, Dec 6, 2008 at 3:40 PM, Rick Jones <[EMAIL PROTECTED]> wrote:
> 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? > By system under load, I was refering to the TCP/IP stack running on the system on which NTP is running. In particular, the available bandwidth over the wireless link. So far I've been playing with Chrony for a few hours this afternoon and it seems to be behaving pretty close to ideally for our situation. After setting the maxdelayratio, it stops incorporating measurements across the wireless when its under load and the latency goes up. Doing some simple tests with ntpdate also seems to have revealed that then the wireless link is saturated, we are almost certainly getting non-symmetric delay times. The way I am testing this is: First, I let the clocks synchronize reasonably. I confirm this when "ntpdate -q" consistently claims an offset of < 1 ms. Next, I saturate the wireless link by initiating some large data transfers. At that point, ntpdate -q starts reporting an offset up to 100ms. Once I open up the wireless link again, the offset drops back to < 1 ms, confirming the clocks have remained synchronized. --Jeremy _______________________________________________ questions mailing list [email protected] https://lists.ntp.org/mailman/listinfo/questions
