Requested revision
Standard: | IEEE Std 802.1Qcc-2018 | Clause: | 46.2.3.5.2 and others |
Clause title: | MaxFramesPerInterval |
Rationale for revision
Section 46.2.3.5.2 states that MaxFramesPerInterval specifies the maximum number of frames that the Talker can transmit in one Interval. This description is repeated a number of places, including 35.2.2.8.4, for the variable MaxIntervalFrames. However, there is no indication exactly what these variables are counting. Is it:
1. The number of frame transmissions initiated?
2. The number of frame transmissions completed?
3. A fractional number of frames rounded up? Down?
4. The number of whole frames both initiated and completed?
In particular, if you slide the moving Interval checking window so that both ends of the window are in the middle of two different (or even the same!) frame, how many frames do you count? (What would answer 4 mean, then?)
This is an important question, because there are examples of transmission timing such that every pair of interpretations, above, will give a different count. (Yes, even counting packet starts vs. packet ends can give different results.)
Are there two different answers, depending on whether you are transmitting or checking for conformance?
My tentative conclusion, after consultation with others, is that interpretation 1 makes the most sense. It seems to answer all of these questions.
Proposed text
Change 46.2.3.5.2 MaxFramesPerInterval from:
MaxFramesPerInterval specifies the maximum number of frames that the Talker can transmit in one Interval.
to:
MaxFramesPerInterval specifies the maximum number frame transmissions that the Talker can initiate in one Interval.
Impact on existing networks
Unknown. To the extent that the current text is open to any interpretation, there could be no impact or a large impact.
Originator
Name: | Norman Finn | Email: | nfinn@nfinnconsulting.com <mailto:nfinn@nfinnconsulting.com> |
Affiliation: | Huawei Technologies Co. Ltd | ||
Submitted: | 2019-11-14 |