| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
After some talking with Brett McClellan, Eve Huang, Lance Wang, and Tim Baggett, I realize that the hard and soft aging cycles ARE actually BEACON cycles, so we don’t have to increase the soft aging cycle
default to match the number of nodes on the network. George Zimmerman, Ph.D. President & Principal CME Consulting, Inc. Experts in Advanced PHYsical Communications 310-920-3860 From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Trying to get more minds on this subject: I’m looking at the new descriptive language that describes the aging of claims: “Soft claims are removed from the claim table, txop_claim_table, if they are older than soft_aging_cycles. Similarly, stale hard claims are removed every hard_aging_cycles.” This appears to be the function desired. When I look at the D-PLCA Aging state diagram, this appears to occur in the state TXOP_END. However, if I am right, this isn’t what TXOP_END does. What it looks like TXOP_END does is clear the soft claims from the claims table every soft_aging_cycles transmit opportunities (irrespective of when during the past soft_aging_cycles Tos the soft claim was
set). Hard claims are more difficult to describe. If I look at the state diagram, the txop_claim_table and txop_claim_table_new are kept synchronized in that they are initialized clear at the same time (DISABLED state), updated each at the same time (in the
UPDATE_SOFT and UPDATE_HARD states, and have soft claims cleared out at the same time (TXOP_END state). However, the clearing that happens at hard_aging cycles seems the relevant part: First the txop_claim_table_new is moved into the txop_claim_table (this is the one that gets used in the D-PLCA Control state diagram).
THEN, the txop_claim_table_new gets completely cleared (both hard and soft claims). At this point it is out of sync with the txop_claim_table being used – in terms of soft claims, until the next soft claiming
cycle, but that doesn’t seem to matter… In terms of hard claims, what I think this means is that the claims transferred to txop_claim_table (and therefore used by D-PLCA Control) are those that have been hard claimed during the previous hard_aging_cycle
transmit opportunities. First, IS THIS THE FUNCTION WE WANT? (or do we want something more like each claim, hard or soft, having it’s own age – something I don’t think is in the state diagram as written). IF MY UNDERSTANDING IS CORRECT, I would suggest replacing “Soft claims are removed from the claim table, txop_claim_table, if they are older than soft_aging_cycles. Similarly, stale hard claims are removed
every hard_aging_cycles.” With “All soft claims are cleared from the txop_claim_table every soft_aging_cycles. Similarly, only those hard claims detected within the previous hard_aging_cycles transmit opportunities are loaded into the
txop_claim_table used by the D-PLCA Control state diagram.” Second, for this to work properly, it is a requirement that hard_aging_cycles be greater than soft_aging_cycles – or the algorithm won’t work… Ideally it is greater than several cycles of soft aging cycles
(but not necessarily a multiple of soft aging cycles). If my understanding above is correct, I would go further and replace the final sentence of the paragraph: “The value of hard_aging_cycles should be sufficiently greater than the value of soft_aging_cycles
to maintain stability of the D-PLCA process as well as interoperability with statically configured PLCA nodes.” With “The values of both hard_aging_cycles and soft_aging_cycles should be greater than several than the maximum number of transmit opportunities expected to be on a mixing segment. Additionally, the value of
hard_aging_cycles should be sufficiently greater than soft_aging_cycles, preferably several times greater, to maintain stability of the D-PLCA process as well as interoperability with statically configured PLCA nodes.” This seems to suggest that our default for soft_aging_cycles might be too low… 50 nodes seems to be within the scenarios many have considered. We may wish to consider changing this to 100. I do NOT plan
to submit that comment this round, but want people to think about it. The default for hard_aging_cycles (1000) seems perfectly fine, even if we change soft_aging_cycles to 100…. I haven’t submitted comments yet, so input is appreciated. George Zimmerman, Ph.D. President & Principal CME Consulting, Inc. Experts in Advanced PHYsical Communications 310-920-3860 To unsubscribe from the STDS-802-3-SPMD list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPMD&A=1 To unsubscribe from the STDS-802-3-SPMD list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPMD&A=1 |