BIND 10 and DHCP's Opportunities

David W. Hankins David_Hankins at isc.org
Wed Apr 22 23:55:38 UTC 2009


I don't frequently do this sort of thing, your time is a kind of
currency I don't want to overspend, but now that ISC has said
something about this in a press release,

  https://www.isc.org/about/pr/2009042200

I feel like I can finally talk about something very exciting, and can
not resist that lure.

ISC DHCP is mentioned in that press release, and you might be
wondering why.

The BIND 10 software package is potentially a great opportunity for
ISC DHCP in many respects.  Foremost for BIND 10 to succeed, there is
going to be a new foundation, a set of libraries, a framework upon
which to build a daemon, and we could take great advantage by using
the product of that work (to shed DHCP's cruft).

This quickly becomes an opportunity to reorganize the software.  A
time to re-architect.  Not just so you don't have to have a 2-megabyte
dhclient binary because it shares common code designed for a DHCP
server...but also so that we can divide 'centers of work' and organize
standard DHCP server work under multiple processes.

To turn the DHCP software package into a set of tools - which are in
their default 'organization' a DHCP client, relay, and server, but
could also be turned into hybrids (a server that loads config from
a client, a relay that serves some networks or some clients), or
integrated in workflow with other ISC software (where primary
authority for a DNS zone is seen as integral with a DHCP server - it
is both a DNS and DHCP server at once, and can perform ddns updates
much easier).

This work stretches the boundaries of how ISC software has classically
been configured (with vi or emacs), and pushes into more modern
territory.  I hope to see us conquer and pillage that territory.

In particular, I look forward to getting out of maintaining our own
programming language in dhcpd.conf ("executable statements"), and
rather providing for proper, maintained, embedded languages to provide
that substance.  I look forward to replacing OMAPI with a
"configuration bus", a robust mechanism that manages the running 'work
centers' and provides and updates their interconnections and run-time
configuration.

And I look forward to having a slim 'just-the-facts' dhclient and
dhcpd that can fit on a floppy, and learn to perform more robust
actions /as those services are demanded/ rather than having to have
them compiled in and using memory no matter how big or small your
system is.

So this really is a very exciting time for me, the first shovels are
being overturned in a new project.  The very air is charged with what
could be, but tinged bitterly with the smell of what might not come to
pass...without support.


The thing I would like to have more of right now is parties with
interest in DHCP as sponsors to the BIND 10 project, to help us make
sure we fully utilize this opportunity.  If there's any way you think
you can participate, please don't hesitate to contact us.

  https://www.isc.org/bind10

-- 
David W. Hankins	"If you don't do it right the first time,
Software Engineer		     you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <https://lists.isc.org/pipermail/dhcp-announce/attachments/20090422/d937440e/attachment.bin>


More information about the dhcp-announce mailing list