[ntp:questions] Re: Need ntpdc command to recreate sockets

John DeDourek dedourek at unb.ca
Mon Jan 10 22:02:01 UTC 2005

Enrique Perez-Terron wrote:
> On Mon, 2005-01-10 at 14:51, mayer at gis.net wrote:
>>Enrique Perez-Terron wrote:
>>>I am on a DSL account that regularly disconnects me for a minute and
>>>then gives me a different IP address.

> I already coded a rescan of the interfaces, and began looking at what
> else was needed.  Ouch, it seems the local interface IP address is used
> for a couple of things!
> - Enrique

The interesting question is whether a change in the
local IP address indicates a change in the characteristics
of the path to the server.  The statistics in the
"smoothing filter", e.g. the delay to the server,
presumably are designed on the assumption that the
path doesn't change.

Now, if I am on my DSL connection and the ISP changes
the IP address, I am presumably still connected to the
same DSLAM and probably have the same path to the router,
presuming the ISP is not doing selective routing based on
IP addresses.

However, if I am on my laptop attached to one network,
then detach, walk down the hall, and reattach to the
other network in the building (taking the interface
down then up) the path to the server may well change.
Two cases: I switch interfaces, e.g. from wired to
wireless; or I switch from one network to another
on the same interface, e.g. on wireless.

I won't mention the issues of Mobile IP, as that is
rather controversial as to whether it will be widely
deployed.  But this introduces the issues of the
home IP address and the "care of" IP address.

But the essential question is whether a change in the
local IP address should cause the server to go
"not synchronized" and then reaccumulate sufficient
measurements to the other servers before going
"synchronized" again.

More information about the questions mailing list