IP Addressing

The issue of IP addressing is important to a CDMA2000 system design. The
introduction of simple IP and mobile IP with and without Virtual Private
Network (VPN) requires the use of multiple IP addresses for successful
transport of the packet services envisioned to be offered. It is therefore
imperative that the IP addresses used for the network be approached from
the initial design phase to ensure a uniform growth that is logical and easy
to maintain over the life cycle of the system.
Not only does the introduction of packet data require an IP address
scheme for the mobility portion of the system, each of the new platforms
introduced needs to have its own IP address or range of IP addresses. Some
of the platforms requiring IP addresses involve
■ PDSN
■ FA
■ HA
■ Routers
Some of these new devices require the use of private addresses as well as
some public addresses. However, because the range of perturbations for IP
address schemes is so vast and requires a specific look at how the existing
network is set up and factoring into the mix the desires for the future, a
generic discussion on IP address schemes will follow.

The use of IPv4 format is shown in the following. IPv6 or IPng is the next
generation, it enables for QoS functionality to be incorporated into the IP
offering. However, the discussion will focus on IPv4 because it is the protocol
today and has legacy transparency for IPv6.
Every device that wants to communicate using IP needs to have an IP
address associated with it. The addresses used for IP communication have
the following general format:
Network number Host number
Network prefix Host number
There are, of course, public and private IP addresses. The public IP
addresses enable devices to communicate using the Internet, although private
addresses are used for communication in a LAN/WAN intranet environment.
The CDMA2000 system will utilize both public and private
addresses. However, the bulk of the IP addresses will be private in nature
and depending on the service offering, will be dynamically allocated or static
in nature.
Table 13-15 represents the valid range of public and private IP addresses
that can be used. The private addresses will not be recognized on the public
Internet system and that is why they are used. Also it will be necessary
to reuse private addresses within sections of its network, profound as this
may sound. Because the packet system is segregated based on the PDSN,
each PDSN can be assigned the same range of IP addresses. Additionally
based on the port involved with the PDSN, the system can be segregated
into localized nodes, and the segregation enables for the reusing of private
IP addresses ensuring a large supply of a seemingly limited resource.
The public addresses are broken down into A, B, and C addresses with
their ranges shown in the following.
The private addresses that should be used are shown in Table 13-16.

To facilitate the use IP addressing, the use of a subnet further helps
refine the addressing by extending the effective range of the IP address
itself. The various subnets are defined in Table 13-17. The IP address and
its subnet directly affect the number of subnets that can exist and from
those subnets, the amount of hosts that can also be assigned to that subnet.
It is important to note that the IP addresses assigned to a particular
subnet include not only the host IP addresses but also the network and
broadcast address. For example, the 255.255.255.252 subnet that has two
hosts requires a total of four IP addresses to be allocated to the subnet:
two for the hosts, one for the network, and the other for the broadcast
address. Obviously, as the amount of hosts increases with a valid subnet
range, the more efficient the use of IP addresses becomes. For instance,
the 255.255.255.192 subnet enables for 62 hosts and utilizes a total of 64
IP addresses.
Therefore you might say, why not use the 255.255.255.255.192 subnet for
everything? However, this would not be efficient either, so an IP-address plan needs to be worked out in advance because it is extremely difficult to
change once the system is being or has been implemented.
Just what is the procedure for defining the IP addresses and its associated
subnet? The following rules apply when developing the IP plan for the
system; the same rules are used for any LAN or ISP that is designed. There
are four basic questions that help define the requirements:
1. How many subnets are needed presently?
2. How many are needed in the future?
3. How many hosts are on the largest subnet presently?
4. How many hosts are on the largest subnet in the future?
You might be wondering why the use of multiple hosts should be factored
into the design phase for CDMA2000. The reason is that it is possible to
have several terminals for a fixed application using a single CDMA2000
subscriber unit or fixed unit.
Therefore using the previous methods, an IP plan can be formulated for
the wireless company’s packet-data platforms. It is important to note that
the IP plan is should not only factor into the design the end customers’
needs but also the wireless operators’ needs.
Specifically, the CDMA2000 operators’ needs will involve IP addresses
for the following platforms as a minimum. The platforms requiring IP
addresses are constantly growing as more and more functionality for the
devices is done through SNMP.
■ Base stations
■ Radio elements
■ MicroWave point to point
■ Subscriber units
■ Routers
■ ATM switches
■ Work stations
■ Servers (AAA, HA, FA, and PDSN)
The list can and will grow when you tally up all the devices within the
network both from a hardware and network management aspect. Many of
the devices listed previously require multiple IP addresses in order to
ensure their functionality of providing connectivity from point A to point B.
It is extremely important that the plan follows a logical method. Some CDMA2000 network equipment may also require an IP plan that incorporates
the entire system and not just pieces.
A suggested methodology is to
■ List out all the major components that are, will be, or could be used in
the network over a 5 to 10 year period.
■ Determine the maximum amount of these devices that could be added
to the system over 5 to 10 years.
■ Determine the maximum amount of packet data users per BSC.
■ Determine the maximum amount of packetdata users per PDSN.
■ Determine the maximum amount of mobile IP users with and without
VPN.
■ Determine the maximum amount of simple IP users with and without
VPN.
The reason for the focus on the amount of simple- and mobile-IP users
lies in the fact that these devices will have the greatest demand for IP
addresses due to their sheer volume in the network.
Naturally, each wireless system is unique and will require a different IP
address scheme to be implemented. However, the concept presented has
been beneficial and should prove useful. If more information is sought on IP
address schemes, an excellent source for information is available on the
Web at www.cisco.com.