stds-80220-eval-criteria: Summary of 12/16 ECCG conference call
Attendees: Larry Alder, Todd Chauvin, Arnab Das, Farooq Khan, David Shively, Anna Tee,
Jerry Upton, Mike Youssefmir
1- Interface between link and system simulations
------------------------------------------------
The evaluation criteria would specify an acceptable interface between link and system simulations. This is needed because the link and system simulations are performed separately (the simulation complexity would be very high if joint link and system simulations are required).
The group discussed 2 potential solutions to the link-system interface:
a) use actual link curves:
In this option, link curves are produced for each of the coding, modulation and channel model combinations. However, it was pointed out that the channel models currently being specified in the channel modeling correspondence group are based on a stochastic approach and therefore results in infinite combinations making this approach impractical. Therefore, if the link curve approach is to be used, we need to have a small set (4-5) of discrete channel models.
b) specify a methodology for link-system interface:
Using this option, a single AWGN curve (or a small set of curves) is used as reference. The deviations ("fudge factors") from this reference curve for different coding, modulations and channel model realizations are determined based on some analytical formula.
It was also pointed out that the group may not need to specify any link-system interface at all. It can be left to individual proposals to justify and validate the link-system interface used in the simulations.
2- System simulation calibration
---------------------------------
The evaluation criteria would specify a system simulation calibration process. However, it is not clear, at this stage, to what level of detail different simulations need to be calibrated. The group is open to proposals to nail down the calibration specifications.
3- Simulation and evaluation of various channel bandwidths
----------------------------------------------------------
The objective of the evaluation criteria group is to have a small number of preferred channel bandwidths for simulation purpose. However, people will be free to simulate the full set of channel bandwidths. At this point, it seems too early to specify this small set of channel bandwidths for simulations.
It was also discussed that a small number of spectrum allocation (over which the results are quoted) can be specified. The individual technology proposals may then split the total spectrum into a given number of carriers w/ reuse of 1 or greater than 1 (if needed by the technology).
4- Fairness criteria
--------------------
It was pointed out that the fairness criteria based on the normalized throughput currently specified in the document may apply to only a subset of applications such as full buffers case. It was also discussed that the values used in the fairness criteria may need to be revisited.
The individual fairness criteria (or outage criteria) would be needed on a application by application basis in the absence of the fairness criteria based on normalized throughput.