IS-95 to CDMA2000-1X

An all-too-common situation for wireless operators is addressing the issue
of how to integrate CDMA2000 into their network. Many operators have
devised their own method for implementing CDMA2000 into an existing
IS-95 network. However, not all the operators have implemented
CDMA2000-1X. Therefore the following will attempt to bring to light many of the issues associated with integrating a CDMA2000 system with
that of a IS-95 system.

Migrating from IS-95 to a CDMA2000-1X platform enables the use of
packet data services along with previously stated increases in voice, circuit
switched, and carrying capacity. The migration process needs to not only
factor in the new services being offered, but also the fundamental problem
of still utilizing existing IS-95 equipment.
Figure 13-8 is meant to depict the possible paths that a wireless operator
may choose to migrate from an IS-95 system for packet data services.
The operator has the choice of waiting for 3X platforms to emerge, but the
more rational approach would be to migrate to a 1X platform and then at a
future date, when services warrant the move, migrate to a 3X platform.

For this design, both CDMA2000-1xEV-DO and CDMA2000-1xEV-DV
channel types will be available for deployment as was the situation with the
new CDMA2000-1X system design previously presented. Because the
design is a migration to a new technology platform, the system will most
likely not be coverage-limited but capacity-driven. Obviously in real life,
there are always coverage issues to address, but for the purposes of this
example, coverage issues will not be considered.

The first step in this case is to determine the desired traffic load for both
circuit switched as well as packet data. The circuit switched traffic growth
is shown for the system in Table 13-30. The growth represents the increase
in circuit switch usage that will be exhibited with CDMA2000-capable
handsets.

If the forecasting only involved circuit-switched services, such as voice,
then the design process would be straightforward in that a 1:1 replacement
of existing IS-95 radios and associated infrastructure would take place only
in the areas where capacity was of most concern. Therefore the introduction
of CDMA2000-1X would be extremely limited or highly focused into
selected areas.

But for this design example, the use of packet data services is included
with this design. Therefore utilizing the traffic loading numbers presented
earlier Tables 13-31 and 13-32 show the expected traffic load from a total of
17,500 potential users of the wireless system that sales and marketing
expect will use the system for packet services. Because the actual throughput
is undefined due to the lack of actual traffic data from the network, the
design will encompass all the possible traffic loads.

Naturally, if packet data services do not encompass all the speeds possible,
then some of the services included in the example can be eliminated.
Table 13-32 shows the expected load on the overall system in Erlangs and
Mbps. The reason for Erlangs is relative for circuit-switched data whereas
that for packet is in Mbps. In previous comments, if only an estimate from
marketing is available regarding packet data usage, given in a percentage
of voice usage, then the estimation should be done using an Erlang-C model.
Table 13-33 is a summary of the calculations derived for the system traffic
load. However some additional information is contained in the table, and
that is the relative geographic areas associated with each type of traffic. For
the purposes of this example, the areas will be considered to be contained
adjacent to each other for simplifying the example. However in real life, the
areas will be intertwined.

The next step is to determine the number of carriers required to support
the expected load. Because this is a capacity design, the number of existing
sites needs to be identified. The number and relative location within the
morphology class is shown in the Table 13-34.

All the BTS sites listed in Table 13-34 are three sector by design.
In addition, it is assumed again that for this design, a total of 8.2 Erlangs
per sector can be supported for circuit switch per sector, which is derived
from a 2-percent GoS using Erlang B with 14 trunk members. The packet
throughput is based on 2.35 trunk members at 76.8 Kbps. Both the packet
and circuit-switch traffic-handling capacities are very conservative and driven
by the link budget and process gain used.


Cell voice Erlangs
8.2 Erlangs/sector  2.64 (sector gain)

21.648 Erlangs per cell (single carrier per sector)
Packet throughput
2.35  76.8 Kbps/sector  2.64

453.15 Kbps per cell (single carrier per sector)
NCircuit Switched
Estimated traffic/cell capacity
680.15/21.648

32 cells total for the system
CDMA2000 NCircuit Switched
Estimated traffic/cell capacity

196.955/21.648
9 cells total
Packet data
(Estimated traffic/overbooking)/cell capacity

(72.15 Mbps/[10])/453.15 Kbps
16 total for the system
The radius of the particular sites is important to calculate but for this
example, the CDMA2000 is a 1:1 overlaid on top of the existing legacy
platform.

Obviously from the example, the system is coverage-limited and not
capacity-limited. However in briefly looking at the traffic data, the treatment
of one section of the system, building, needs a higher throughput than
the vehicular areas, which is obvious. Therefore the deployment recommendation
is to have two carriers deployed F1 being an IS-95 channel or 1x
and F2 is 1xEV-DO, or 1x which is assigned for data transport only.
From the previous calculations, a total of 9 sites out of the total 32 are
required involved with growth. Because this is an overlay design, the core
of the network will be focused on for CDMA2000-1X carrier deployment
because the bulk of the growth is coming from the building and pedestrian
morphology where in the past, not previously mentioned, the design was for
vehicular only.
Figure 13-9 represents a possible channel deployment scheme that
applies to a PCS system having operating with 15 MHz of duplexed spectrum.
The inclusion of 1x,DO, and DV channels is listed but is really left up
to the traffic mix as well as true availability for the technology. However in
examining the diagram, the inclusion of a legacy channel is left in place for
F1. A 3X deployment is also included from which to see later channels being
deployed are positioned correctly with the channel bit map that does
require the migration from a legacy channel to that capable of 1X or 3X.
There is of course the cellular band that has many unique issues associated
with it when trying to deploy any new technology platform. Figure
13-10 highlights the channel deployment scheme for the A- and B-band
cellular operators. The deployment scheme is meant to help transition the
new technology but also to address the legacy issues.

If a cellular operator, or even a PCS operator has more than one IS-95
channel in operation, the channel deployment schemes can be easily modified
by selecting the next channel on the list as the CDMA2000-1x channel
of choice. The channel deployment sequence is shown in Table 13-35.
The next step is to deploy the channels in a logical fashion, meeting the
individual capacity requirements and maximizing the use of the legacy
equipment. The initial system layout is shown in Figure 13-11 and assumes
that the BSC 1 is collocated with the MSC. However, the remaining BSCs
may be located remotely or also collocated with the MSC. In real life, a system
of this size would expect to have more than one MSC or concentration
nodes to reduce the leased-line costs.
It is recommended that the core of the network consisting of the building
environment utilize two CDMA-2000 carriers while the pedestrian and
vehicular zones use only one carrier and that those carriers be a mix
between CDMA2000 and IS-95 carriers.A hard handoff, of course, will need
to take place between the F2 and F1 zone. However, it is recommended that
in a situation like this that the BTS-F1 carriers process primarily voice
traffic while the F2 is more of a data only situation. As mentioned earlier,
this can be done via software and user-definable parameters.
While poorly represented in the diagram, primarily due to size limitations,
Table 13-36 is the breakdown of the carriers by BSC type. The distribution
should be based on the individual site loading. The distribution
example assumes that packet data services will not be offered throughout
the entire footprint of the system. If a true 1:1 overlay was desired, then the existing legacy equipment should be redeployed or sold on the secondary
markets for a more rural application where voice services would
only be utilized.
As a brief reminder when handing off from a CDMA2000 channel to a
IS-95 system, the loss of packet data services will occur.
Next, the various pipe sizes were estimated for the initial concept. From
the diagram it would be advantageous to collocate BSC 1 with the MSC provided
the MSC is located near a tandem. The other BSCs, however, due to
their initial traffic load, should be considered to be remotely located provided
the operational and support issues can be met. In addition, BSC 6 and
12 are considered to be IS-95 only and therefore are not connected to the
packet network as depicted in the diagram. While it is possible and advisable
to mix the legacy equipment within a BSC, it is not shown in Figure
13-13.
Continuing the facilities between the BTS and BSC are assumed to be
unstructured TDM when the BTSs have CDMA2000 channels. The connectivity
to the off-net data networks assumes a 80/20 mix of public verse private
networks. The assumption used is that 100 percent of the packet traffic is off-net and that mobile to mobile packet sessions will not have a high
enough penetration to consider in the design aspect presently.
Looking at the BTS’s two different configurations are proposed to help
facilitate different areas of the network. The first shown in Figure 13-14 is
for the core area of the network and involves using STD for the transmit
diversity scheme because two carriers are initially needed. One could also
install more antennas if feasible or utilize cross pole antennas.
Regarding the antenna systems, there are some different considerations
to take into account when migrating from a IS-95 system to a CDMA2000
system if it is an AMPS or PCS spectrum. Thediversity, and this will be
achieved either by a STD or OTD method. However, the STD method is the
preferred version. Figure 13-14 (a) shows a STD transmit diversity scheme
whereas Figure 13-14 (b) shows an OTD transmit diversity scheme.
Figure 13-14 shows a typical situation where there are two or three
antennas per sector available for use. Sometimes there is only one antenna
but it is a cross pole antenna, which can be treated as two separate antennas.
With an AMPS system as the underlying legacy system, the use of a
STD transmit diversity scheme is possible with a configuration shown in
Figure 13-14 (a) with the exception that only one carrier is used for CDMA.
If a second carrier is added, then OTD diversity is utilized and the configuration
shown in (a) is used. Now if the operator has been able to secure
more antennas per sector, that is, 5, then the configuration shown in (b) is
the desired method where the AMPS and CDMA systems are bifurcated.
The use of STD or OTD is again dependant upon the number of carriers
required at the site.
The PN offset assignment scheme that is presented in the earlier part of
the chapter should be used for the system design following a N
19 reuse
pattern for the PN offsets.
Just as with the design example used for a new CDMA2000-1X system,
there is a plethora of issues not covered in the example. However, it is
believed that the preceding material should help in the construction of the
thought process to achieve the desired goal of supporting the customer
requirements for service delivery and transport.