Network Design and Optimization
By Vikram Singh
Wireless communication network consists of two parts (network infrastructures), the
radio access network (RAN) and core network (CN). 4G-BWSim and 5G-BWSim focus
on the RAN infrastructure which consists of three layers L1, L2 and L3. 4G
simulator can simulate all the three layers whereas the 5G simulate the layer 1
end to end and elements of layers 2 and layer 3 but not complete layers. Both
these simulators allows the designers/developers design the algorithms for
radio access network but also optimize the methods/algorithms based on their
customers need. Consider an example, an algorithm designed for mobile broadband
may not be the best for IoT applications because the for mobile broadband
applications the objective of algorithm design is to maximize the
data-rate/throughput whereas for IoT applications data rate requirements are
often low instead the power consumption is a major concern. Similarly, the
protocols and algorithms designed for outdoor applications might not be the
best for Indoor factory/gaming/V2X application. Hence, the terrain, scenario
and use-cases plays a crucial role while designing the algorithms. The terrain
and cell size are often captured by the channel model. In channel models, the
small-scale parameters and large-scale parameters are calculated from data
collected from a specific terrain. These parameters are characterized by the
probability distribution function, mean value and the standard deviation. To
make these channel model realistic these parameters maintain inter-parameter
correlation, spatial consistency and correlation across space and time. On the
other hand, the cell size has constraints on the base-station transmit power as
the path-loss is lower and line of sight probability is higher in small cells
(micro) compared to large (macro) cells.
To design the whole network, the designers test their algorithms, protocol
stacks and procedures implemented for different layers of the network using
simulators.
These simulators helps in identifying following:
- The efficacy of the algorithms.
- Validation of the protocol stacks and algorithms before SoC (hardware) design.
- Evaluate various QoS parameters such as error rates, complexity (power consumption), implementation time, throughput and E2E/layer-to-layer latency.
- Optimize the algorithms and protocol stacks based on the requirements.
- Cost effective evolution of the network by modelling different modules/interfaces.
GigaYasa Wireless has designed system level simulators for 4G and 5G networks,
called BWSim-4G and BWSim 5G-R15
respectively. Both our simulators are very helpful in simulating the networks
for wide varieties of use-cases, terrains and capabilities.