How System Performance Can Be Compromised

System performance can be compromised by loss of orthogonality. OVSF codes, for
example, are not particularly robust in dispersive channels (e.g., large macrocells).
Degraded orthogonality increases intracell interference and is expressed in radio planning
as an orthogonality factor. Orthogonality on the downlink is influenced by how
far the users are away from the Node B. The further away they are, the more dispersive
the channel and the more delay spread there will be. Loss of orthogonality produces
code interference.
We said earlier that we can implement soft handover to improve coverage. Effectively,
soft handover gives us uplink and downlink diversity gain; however, soft
handover absorbs radio bandwidth and network bandwidth resources, long code
energy on the radio physical layer, and Node B to RNC and RNC to RNC transmission
bandwidth in the IP RAN. If lots of users are supported in soft handover, range will be
optimized, but capacity (radio and network capacity) will be reduced. If very few users
are supported on soft handover, range will be reduced, but (radio and network) capacity
will increase.
In the radio network subsystem (RNS), the controlling RNC (CRNC) looks after load
and congestion control of its own cells, admission control, and code allocation. The
drift RNC (DRNC) is any RNC, other than the serving RNC, that controls cells used by
the Node B. Node B looks after channel coding and interleaving, rate adaption, spreading,
and inner-loop power control.
The RNC looks after combining—the aggregation of multiple uplinks and downlinks.
Note that a handset can be simultaneously served by two Node Bs, each of
which sends and receives long code energy to and from the handset. In addition, each
Node B could be sending and receiving multiple OVSF code streams via two Node Bs
to either the CRNC, or if the handset is between RNCs, to both the CRNC and DRNC.
The CRNC and DRNC then have to talk to each other, and the CRNC has to decide
which long code channel stream to use on a frame-by-frame basis or to combine the
two long code channel streams together to maximize combining again. This is a
nontrivial decision-making process that will need to be optimized over time. It will
take at least 5 years for these soft handover algorithms to be optimized in IMT2000DS
networks.
The RNC also has to respond to admission priorities set/predetermined by the
admission policy, which is predetermined by individual user or group user service
level agreements. The requirements of the traffic (tolerance to delay and delay variability)
determine the admission policy and how offered traffic will be distributed
between serving cells. There are four service classes in IMT2000. The low-delay data
(LDD) services are equivalent to the constant bit rate and variable bit rate services
available in ATM and are used to support conversational and streamed video services.
The service classes are shown in Table 11.12.

LCD and UDD are equivalent to the available bit rate and unspecified bit rate services
available in ATM and are used to support interactive and best-effort services.
Low bit error rates can be achieved if data is delay-tolerant. Ahigher-layer protocol, for
example, TCP, detects packet-level error rates and requests a packet retransmission. It
is possible to reduce bit error rates. The cost is delay and delay variability.
Link budgets (coverage and capacity planning) are therefore dependent on individual
user bit rates, user QoS requirements, propagation conditions, power control, and
service class. The offered traffic statistics will determine the noise rise in the cell. For
example, a small number of high bit rate users will degrade low bit rate user’s performance.

The failure to meet these predefined service levels can be defined and
described as an outage probability.
The job of the RNC (controlling RNCs and drift RNCs) is to allocate transmission
resources to cells as the offered traffic changes. Loading can be balanced between
RNCs using the IUR interface. This is known as slow dynamic channel allocation. An RNC
balances loading across its own Node Bs over the IUB interface. RF resources (code
channels) are allocated by the Node B transceivers. This is described as fast dynamic
channel allocation, with the RNC allocating network bandwidth and radio bandwidth
resources every 10 ms. Both the network bandwidth and radio bandwidth need to be
adaptive. They have to be able to respond to significant peaks in offered traffic loading.