[ntp:hackers] use of DHCP to assign Time Servers....

todd glassey tglassey at earthlink.net
Mon Sep 4 21:34:04 UTC 2006


Poul - damn there are so many come backs its really a bummer not to use
them... but making you look like a fool is not my aim here. I am not here to
demonstrate how childish and incompetent your tirades are... especially when
you totally missed the boat... but what can I say - the convo I started was
meant to open the door to the ISC as the NTP.ISC.ORG hosts, since they own
the code I am talking about including into the NTP package eh?
http://www.isc.org/index.pl?/sw/dhcp/  you did get that, right Poul?


As to RFC2131 and 32
----------------
And yes thanks for pointing out the obvious -  that RFC2132 is the DHCP
vendor extensions RFC and can be found http://rfc.net/rfc2132.html  for the
group's info DHCP is 2131 (http://rfc.net/rfc2131.html) and the IETF's and
ISC WG's committed to DHCP already thought about this... so why is it that
we don't use those features?.

For the record - they were defined 10 years ago as well as the methods of
creating specific vendor extensions to move larger blocks of configuration
or result data around. The idea of course was where needed that applications
on the DHCP configured stack could also get control info from a master DHCP
server. Whoa - what a concept eh?


Adding DHCP configuration
-----------------------------
As far as the rest of the list is concerned, what I was saying - is that
since  the excellent code that the ISC already has  for DHCP supposedly
already allows for the use the Vendor Extensions to facilitate  Time Server
Dissemination, then what would be the hurdle to formally incorporating these
features into NTP5 and SNTP5?

Seriously not many folks use the Vendor Extensions that I am aware of but
its not a problem per se. So lets ask what it would take on our side to
incorporate the DHCP Client Code into the package to allow for setup of the
of our applications as well.

For NTP its the daemon and the service routines which either use the daemon
or seek their own connection to a server or server set.  Also there is a
question as to how much info would want to be propagated this way - for
instance could a certified configuration be fully instantiated from a DHCP
connection?


The ISC and their services
---------------------------
hey Vix - you guys are the experts on this - how would you propose
integrating the ISC DHCP Client's functionality into NTP and SNTP?  Would
you folks also be willing to operate a master NTP DHCP Root?

Todd Glassey



From: "Poul-Henning Kamp" <phk at phk.freebsd.dk>
To: "todd glassey" <tglassey at earthlink.net>
Cc: <hackers at ntp.isc.org>
Sent: Monday, September 04, 2006 12:18 PM
Subject: Re: [ntp:hackers] use of DHCP to assign Time Servers....


> In message <040e01c6d055$cc3f35d0$010aa8c0 at home.glassey.com>, "todd
glassey" wr
> ites:
>
> >I want to standardize it Poul which means formally moving within the
group
> >instead of externally to it. Group consensus and all that.
>
> Please do your homework before wasting our time:
>
> RFC-2132 ("DHCP Options and BOOTP Vendor Extensions")
>
> Specifically section 8.3 ("Network Time Protocol Servers Option")
>
>    This option specifies a list of IP addresses indicating NTP [18]
>    servers available to the client.  Servers SHOULD be listed in order
>    of preference.
>
>    The code for this option is 42.  Its minimum length is 4, and the
>    length MUST be a multiple of 4.
>
>     Code   Len         Address 1               Address 2
>    +-----+-----+-----+-----+-----+-----+-----+-----+--
>    |  42 |  n  |  a1 |  a2 |  a3 |  a4 |  a1 |  a2 |  ...
>    +-----+-----+-----+-----+-----+-----+-----+-----+--
>
>
> The symbolism of it being code #42 should not be missed.
>
> Also, a search for "dhcp ntp" on Google would give you a first
> hit which tells you how to configure it on Gentoo Linux.
>
>
> Poul-Henning
>
> -- 
> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
> phk at FreeBSD.ORG         | TCP/IP since RFC 956
> FreeBSD committer       | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by
incompetence.



More information about the hackers mailing list