cpignata / powerbench

Benchmarking Methodology for Power
0 stars 1 forks source link

Repeatable test condition for traffic load #9

Open billwuqin opened 6 months ago

billwuqin commented 6 months ago
  1. Section 4.2 of draft-cprigf-bmwg-pwoerbench said: “ The traffic load must specify packet sizes, packet rates, and inter- packet delays, as all may affect the energy consumption of network devices. ” I believe this requirement enable repeatable test condition, which require to specify the exact repeating sequence of packet sizes from the usual set of fixed sizes and from other forms of mixed-size specification. However I am not sure the description in section 4.2, e.g., specify inter-packet delays, is consistent with ETSI or ITU-T specification, ATIS specification define IMIX traffic, which is A realistic mixture of packet sizes can be approximated by a set of packet lengths (three or more) that represent the common modal lengths, plus an even distribution of every other packet size. One example of IMIX model is: Table 2: Simple IMIX model Package Size (Bytes) Proportion of Total Bandwidth 64 7 parts 10,3 594 4 parts 54,7 1 518 1 part 35
romain-jacob commented 6 months ago

Agreed. Ultimately we should have some very clearly defined traffic profiles and appropriate tooling built to make is easy to generate those profiles.

We have started some work on this in my group, but it is still very preliminary.