Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors

Traffic Balancing

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
  • #46821
    Mohammaad Shan

    Dear all,

    could you please explain me how can i solve this task..

    i have 2 sites cover the same area and i want to balace traffic between them knowing that the 2 sites are 1800 and they are layer 2 and reach that area at the same power level.

    thanks in advance


    use the HO_MARGIN
    for instance, to balance traffic towards cell 1 :

    HO MARGIN (cell1, cell2) = 10 (instead of 5)

    HO MARGIN (cell 2, cell 1) = 0 (instead of 5)


    Mohammad Shan

    Hi Pix,

    Could you please tell me exactly what will happen when i set these parameters?

    Many thanks



    Hi all,
    it is good idea. I’ m using this solution for traffic offloading between two or more cells. This solution move border of cell (but it work only if there is traffic on the border). But my opinion is, that it is necessary also to change idle behavior – to set C2 in correct way.
    BR Tomar


    hi Shan and Tomar,

    Thanks tomar for your valuable inputs, it’s very interesting 🙂

    Shan :
    HO MARGIN is the margin used for the power budget handover :
    cell 1 = serving cell
    cell 2 = neighbour cell

    if the difference of rxlev between cell1 and cell2 is greater than HO MARGIN (cell1, cell2) –> MS will handover from cell 1 to cell 2

    To prevent the MS to go back to cell 1, you have to adapt the margin “HO MARGIN (cell 2, cell 1)”, in order to always verify the rule :

    HO MARGIN (cell 1, cell2) + HO MARGIN (cell 2, cell 1) = 10 dB

    yes, of course, the idle mode behaviour should be modified, otherwise HO will be triggered right after the call setup. The Cell Reselect Offsets should be equal to:

    CRO(cell 1) = (HOMARGIN(cell 1,cell 2) / 2) + (HOMARGIN(cell 2,cell 1) / 2) = 10/2 + 0/2 = 5 dB

    CRO(cell 2) = 0 dB

    (this formula gives a CRO that makes the MS in idle mode to camp in the middle of a zone where there won’t be a PBGT HO right after the call)

    it reminds me : I used to work with negative HO MARGIN (homargin(1,2) = +15dB and homargin(2,1) = -5dB), in case of GSM900 cells and DCS1800 cells + activation of PBGT handovers…

    Mohammad Shan


    i think you mean Cell Load shearing.

    i tried this feature and it did not work as you think.

    Pix, many thanks for you.


    Mohammad Shan

    Dear Pix,

    yoy could apply this feature ( HO margin ) if you have 2 cells cover the same area with diffirent levels, i mean one cell is more strength than the onther, but in my case the 2 cells cover that rea in same level.

    an example: if i have site cover a staduim, and i wanna put Mobile BTS in that staduim at important match to load off traffics from that site.

    knowing that the sites and the MBTS cover staduim in same level, and the 2 sites are 1800 and layer 2.


    what do you mean, same rxlev ?
    the rxlev(cell 1) = rxlev(cell 2) only in some areas of your stadium. Most of the time, in one place of the stadium, one cell will have greater RXLEV than the other cell…

    UNLESS… both antennas are located at the same point 🙂

    In this latter case (colocated sites, or almost colocated), you must use a “capture condition”, based solely on the rx level of the neighbor cell. It is not defined in the 3GPP standard, but i’m certain your equipement vendor has implemented it in its system.

    i.e., while in cell s :

    if rxlev (cell n) > threshold –> HO from cell s to cell n.

    of course, to prevent going back to cell s with a better cell condition, you have to deactivate the PBGT HO from cell N to cell S.

    Good luck !


    Hi Dear Pix,

    First regarding the rule you mention,
    “HO MARGIN (cell 1, cell2) + HO MARGIN (cell 2, cell 1) = 10 dB”

    Is this mandatory to folloew this rule? and 10dB is vendor specific or its your recommendation, because we are using 12dB?

    and if we follow this rule then according to your second rule
    “CRO(cell 1) = (HOMARGIN(cell 1,cell 2) / 2) + (HOMARGIN(cell 2,cell 1) / 2) = 10/2 + 0/2 = 5 dB

    CRO(cell 2) = 0 dB”

    value of CRO(cell 1) will always be same 1.e. 5dB and how did you calculated CRO(cell 2)? e.g. if we use, HO MARGIN 1–>2 = 6dB
    and HO MARGIN 2–>1 = 4 dB
    (following your first rule), what will be CRO (cell 2)?

    another point: if we use different values of “rxLevelAccessMin” and “msTxPowerMaxCCH” for different cells, will the crieteria for calculating CRO be same?

    Thanks in advance.


    Hello Abdul,

    big questions for a thursday morning 🙂

    The 10dB “hysteresis” is something I use, it is not mandatory. You can use 12dB if you prefer.

    The CRO of cell 2 is always equal to 0, if you want to apply my “CRO” formula. Basically, what i’m saying is that we only modify the CRO of cell 1, based on the fact that the CRO of cell 2 is = 0.
    Keep in mind that the CRO is “relative”, it only add more or less weight to a cell, comared to other cells.
    i.e. if all cells have a CRO = 25dB, it’s the same as if they were all havgin CRO = 0dB.
    i.e. if CRO (cell1) = 6 and CRO(cell 2) = 4, it is equivalent to:
    CRO(1) = 2 and CRO(2) = 0

    In the case HOMARGIN(1,2) = 6 and HOMARGIN(2,1) = 4, I would use CRO(1) = 1dB and CRO(2) = 0dB.
    Since CRO=1dB is not possible (CRO has 2dB step-increments), I will use 0dB on both cells.

    Regarding RXLEV ACCESS MIN and MS POWER MAX, it will impact the C1 and C2 computation. So, YES, you’ll have to take it into account in your CRO, if you want to balance the traffic properly.
    For instance.
    cell 1:
    cell 2:
    RXLEV ACCES MIN = -110dB

    A MS will measure a “bonus” of 10dB on cell 2, compared to cell 1. So you should put CRO(1) = 10dB and CRO(2) = 0dB, to counteract this “bonus”.



    Hi Dear Pix,

    Sorry for the late reply.

    Thanks for your answers, but i would like to have more 🙂

    Now i got your formula it is:

    “CRO(cell 1) = (HOMARGIN(cell 1,cell 2) / 2) – (HOMARGIN(cell 2,cell 1) / 2)”

    you misstyped the “-” sign.

    I have a confusion, regarding this, why in idle mode the CRO difference is half the HO Margin difference in dedicated mode? will it not cause HO just after call set up? I mean if we use CRO = the difference of HO margin, would it not better?

    Secondly if we enable Cell reselection PI only in cells where we r using non zero CRO(e.g 1800 cells), and not enabled in cells where we r using CRO=0 (e.g 900 cells) will there be any disadvantage? as in this case change of CRO value in (1800 cell) effects the traffic. But recently i have enabled Cell reselection PI in both layers(900 and 1800) in some cells and i can see the change in traffic trend (more in 1800), can you explain?

    Another point: it is recommended to have same BCC and TSC, can you tell me of any advantages/disadvantages and effects?

    I have tried to be clear, but if iam not kindly let me know.

    Thanks a lot once again.

Viewing 11 posts - 1 through 11 (of 11 total)
  • The forum ‘Telecom Design’ is closed to new topics and replies.