This page is the base point to look at how DHCP and DNS (BIND) management can be integrated in a relatively common way.

There are two aspects to this. One is the configuration used in creating the data and controls used by each system. The second is the administrative and operational controls required by each server.

At a fundamental level, DNS is a wide area system with high traffic demands and strict forms. DHCP is local system that allows individual machines to get dynamic information that they need to operate.

data models

DHCP and DNS have very different object models. DNS has very large volumes of data in well organized structures that have relatively few options and few global/implicit defaults. DHCP has smaller volumes of data, but has many more options and functions that it encompasses. The data structure for DNS configuration closely follows the items defined in the large number of RFC. The data structure of DHCP configuration is much more based on operational models and has much more complexity. The ISC DHCP configuration has opted for a programming language as an interface to deal with these complexities.

At first glance, it seems quite hard to combine these two divergent object models into a common system. I think there are some ways in which each one needs to hold their unique data, but there are also unifying ideas.

It turns out that DNS configuration can benefit from programmable structures. I often refer to these as templates, but the name belies the desire for complex structures and late binding/evaluation. While the current DHCP configuration allows for arbitrary programmablity, it might benefit from a more controlled structure for all but the most challenging cases.

control models

For the operations side, the needs are somewhat closer and share a number of common goals.

Last modified 8 years ago Last modified on Oct 19, 2010, 6:06:30 PM