[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-dev] Moving ntpd to an open VCS

From: Hal Murray
Subject: Re: [gpsd-dev] Moving ntpd to an open VCS
Date: Sun, 27 Oct 2013 23:44:23 -0700

> And they can't nail down a few IPs for the job??

If you are a big ISP, it might be reasonable to nail down a few IPs.

If you are a small organization, trying to nail down an IP Address commits 
you to working with your current ISP, and even then, they may change the 
rules with no warning unless you get the paperwork right.


address@hidden said:
> I think it's not widely understood that most distros provide their own ntp
> servers to their clients and that the dhcp setting is often ignored.

>, etc. 

I think it's more complicated than that.

host resolves to one IP Address that doesn't answer NTP 
requests.  (at least from my IP address),, and all say not found: 

There was a plan to get things like <N>.<distro> to redirect to 
The idea was to have a redirect in there.  If some distro did something 
stupid you could fix it. and such work for me.  (I haven't checked the servers.  
I assume they are part of the pool.)


In case you aren't familiar with the problems with stupid NTP clients, please 
check out:

If I was teaching computer geeks, I'd put Dave Plonka's paper on the 
must-read (and understand) list.


This is getting (has long gone?) off-topic for gpsd lists.

The key idea is that time is tangled up with cryptography.  Most crypto 
schemes assume that they know the time.

If DNS is using crypto that depends upon time, you can't use DNS to find your 
time servers.  Hard-wiring IP Addresses has a long history of problems.  (See 

If each ISP had their own NTP servers, it would be reasonable to hard-wire 
their IP Addresses into boxes like home routers/modems that were distributed 
to their customers.  (or at least more reasonable)

These are my opinions.  I hate spam.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]