3
Alcatel-Lucent | Implementing DHCP Redundancy to Sustain Critical Operations
1
Alcatel-Lucent | Implementing DHCP Redundancy to Sustain Critical Operations
Introduction
IP network administrators are understandably concerned with providing redundancy in the event
of the failure or inaccessibility of their frontline Dynamic Host Configuration Protocol (DHCP)
servers. Using a single DHCP server to manage access to an IP network runs the risk that end users
will be denied access to the network when the DHCP server is down or unreachable. When denied
access, end users will be quick to report their dissatisfaction and lost productivity to those
concerned. Consequently, IP administrators must implement some form of redundant DHCP
server configuration. This paper compares two approaches to providing redundancy—the Split-Scopes
and the Alcatel-Lucent Failover DHCP methods.
Understanding Split-Scopes
The Split-Scopes configuration generally uses two DHCP servers. Each server maintains one segment
of the entire scope of IP addresses for each managed subnet. The segment on the first DHCP server
must be separate, distinct and may not overlap the segment maintained on the other DHCP server.
Both servers receive Discover packets from clients, and both will offer those clients IP addresses
if available. In a Split-Scopes configuration, each server responding to a client’s Discover packet
competes to reach the client with an Offer. Generally, the client will accept the first Offer it receives
that contains all of its desired options. When it receives a viable Offer, the client issues a Request
to the appropriate server as a broadcast, thereby notifying each server of its choice. The appropriate
server then sends an Acknowledge packet back to the client, allowing the client to initialize TCP/IP
and access the network. Consequently, network proximity generally determines which server will
be the frontline DHCP server and which the redundant server.
The Split-Scopes configuration provides proven redundancy protection
The benefits of Split-Scopes redundancy are straightforward. First, it is a proven approach that
has been successfully implemented in many client-server environments. Second, it provides
redundancy in the event of hardware, software and partial network outages. Since the DHCP
servers in a Split-Scopes configuration operate completely independently of one another, there
is no replication or communication needed between them.
Wasted IP addresses, uncertainty about primary and redundant server assignments, and additional
network traffic all complicate DHCP management in the Split-Scopes configuration
The redundancy protection in the Split-Scopes configuration comes at a price. The configuration:
• Systematically wastes IP addresses by requiring IP network administrators to assign a
separate and distinct set of IP addresses to the redundant DHCP server that may never be
used to support services
• Prevents administrators from guaranteeing which DHCP server acts as the frontline server,
and which is the backup system
• Increases network traffic by allowing both servers to compete to offer IP addresses to clients
First, the Split-Scopes configuration requires customers to waste a portion of their scope of IP
addresses. Because the IP address segments on each server cannot overlap, IP network administrators
must assign the redundant server a segment of addresses that will only be used if and when the
frontline DHCP server fails. The intended frontline server cannot offer clients IP addresses from
the segment on the redundant server. IP network administrators will not always have to provide
a 50-50 split of addresses between the intended frontline and redundant servers. Instead, IP
network administrators can determine the appropriate split based on estimates involving the