High-Level Modeling and Exploration of a Powerline Communication Network Based on System-on-Chip - Multi-objective Design Space Exploration of Multiprocessor SoC Architectures - page 156

7.2.2.5 Description of Parameters and Metrics

The use case platform has been configured in order to continuously transmit IP flows
between two of the nodes of the system (which is composed of up to 16 PLC nodes).
The configuration of this flow is the following:

￿ Simulation characteristics

- Simulation length is 0.5 s of real time

- The parameters that are not studied are kept constant and at a “normal” value

￿ Topology of the network

-

Sixteen Nodes are active during the whole simulation (all the nodes are running
and listening)

- The communication is done only between two of the nodes of the system: Node

0 and Node 1

- Channel noise can be modified by MULTICUBE tools to be between 0 and

100% probability of loss

￿ Traffic shape

- Traffic type is multimedia (MPEG) generated by the platform traffic generators

- Traffic is monitored and the quality of the link is continuously checked

- The communication is bidirectional

- All the flows are UDP-based

- Flow rates are kept constant. No peaks in the flow have been configured

- The IP flow we inject in the network from Node 0 to Node 1 is 40 Mbps

- The IP flow we inject in the network from Node 1 to Node 0 is 1 Mbps

￿ Quality of Service

- No prioritization of traffic between nodes

- No prioritization of flows within a single node

With this setup, we have defined a set of parameters that have been analyzed during
this use case and are presented in Table 7.4 .
It is interesting to note that four out of these five parameters are design parameters
while the fifth one (LOSS) is an ambient parameter. This means that while the first
four parameters have to be optimized, the fifth one is there to put the use case under
different situations for the analysis.
All these parameters are related to the main concepts we want to investigate in
the design: Memory length and QoS capabilities. None of these concepts is more
important than the other and these concepts will depend on the final application of the
integrated circuit (mainly because the optimization process was long and thus, there
was only time to optimize one of them). Till now, the procedure was to guarantee
a correct performance for all concepts. Thanks to MULTICUBE processes and the
automation of the procedure we can envision to design more adapted ICs for each
situation.
In addition to the parameters, a set of metrics have been used to evaluate the
performance of the system under different running conditions. These metrics are
presented in Table 7.5 .