Re: [6tisch] Yatch's review on draft-ietf-6tisch-msf-08 - Re: MSF review

[Yasuyuki Tanaka <[e-mail]>]

Hi Tengfei,

Related to MAX_NUMCELLS, LIM_NUMCELLSUSED_HIGH and LIM_NUMCELLSUSED_LOW 
need to be revisited.

In Section 5.1., these values are used in the following way:

    The counters are used as follows:

    1.  Both NumCellsElapsed and NumCellsUsed are initialized to 0 when
        the node boots.
    2.  When the value of NumCellsElapsed reaches MAX_NUMCELLS:

        *  If NumCellsUsed > LIM_NUMCELLSUSED_HIGH, trigger 6P to add a
           single cell to the preferred parent
        *  If NumCellsUsed < LIM_NUMCELLSUSED_LOW, trigger 6P to remove a
           single cell to the preferred parent
        *  Reset both NumCellsElapsed and NumCellsUsed to 0 and go to
           step 2.

NumCellUsed should have an integer value which is equal to or greater 
than 0. Then, LIM_NUMCELLSUSED_HIGH and LIM_NUMCELLSUSED_LOW should be 
the same type of value. However, these recommended values are expressed 
in percentage, which cannot be compared directly with NumCellsUsed.

    Figure 2 lists MSF Constants and their RECOMMENDED values.

            +------------------------------+-------------------+
            | Name                         | RECOMMENDED value |
            +------------------------------+-------------------+
            | NUM_CH_OFFSET                |       16          |
            | KA_PERIOD                    |        1 min      |
            | LIM_NUMCELLSUSED_HIGH        |       75 %        |
            | LIM_NUMCELLSUSED_LOW         |       25 %        |

A correction would be either replacing NumCellsUsed in Section 5.1 with 
(NumCellsUsed / NumCellsElapse), or removing '%' in Figure 2 with 
changing the recommended values to something which can work with a 
recommended value of MAX_NUMCELLS.

On 11/27/2019 9:05 PM, Tengfei Chang wrote:
>     By the way, the default or recommended value of MAX_NUMCELLS is 8?
>     MAX_NUMCELLS is not listed in Figure 3.
> 
> 
> TC: 8 is just a value in the example, not recommended. For periodic 
> traffic, or traffic load changes slowly, a larger MAX_NUMCELLS is preferred.
> Some proposal for the recommended value of MAX_NUMCELLS: 64 (power of 
> 2), or 100.

Best,
Yatch