This page describes the configuration files for the DHCP server.

It is based on XML files which is the default configuration format. An alternative way is to put all data is the database when in high-availability mode (to be described later).

The dhcpd-j server configuration philosophy is IP-centric, compared to other products such as ISC's dhcpd which are more host-centric. This choice is driven by the dynamic nature of DHCP which is the natural way to use DHCP servers in ISP solutions. Static host by host configuration is supported as exceptions.

The central notion is the server is the Subnet (or LAN). Subnets or LANs represent computer networks defined as broadcast domains: every broadcasted packet is seen by all devices on the same subnet. Subnets are connected to each others through routeurs.

Normally, every client on a subnet can use any avalaible address in the subnet range, but cannot use addresses from other subnets (or at least will not be able to communicate with subnet's clients).

Subnets hold the following data:

• a policy composed of the following data:
  • a filter pattern to reject non-eligible hosts, composed of rich boolean expression
  • a lease time policy precising the range of acceptable lease times
  • a list of DHCP options to send back to the client
• a list of pools of addresses to allocate to clients
• a list of statically allocated addresses to specific hosts