Hi All,
I am implementing uplink power control for the Ericsson RBS.
I started with a vty command to force an ms power update on a timeslot and then I implemented a very crude loop and that "works". Next, I imported power_control.c from osmo-bts into osmo-bsc. This is where I began to wonder if this code is better moved some place where it can be shared by both projects. - There would be some small modifications necessary, I think.
I also wonder if it is the right thing to do to integrate this bsc-side UL power control into the current (m|b)s-power control? Would it make sense to add a further "mode" there?
ms-power-control mode dyn-bsc
For example?
Obviously one does not want both a BSC and BTS managing a power loop. Or maybe only allow the BSC side power loop to run with bts-type that is known to not do UL power control?
There are some other issues I'm still unsure about, but this is enough by way of moving towards a patch fit for code review.
Thanks!
Keith.
I forgot to say:
This is the ticket: https://osmocom.org/issues/4922
and I pushed WIP to some private branches under keith/
Hi Keith,
On 9/23/21 18:38, Keith wrote:
This is the ticket: https://osmocom.org/issues/4922
I shared my thoughts in the ticket too.
Hi Keith,
On Thu, Sep 23, 2021 at 11:09:35AM -0500, Keith wrote:
I am implementing uplink power control for the Ericsson RBS.
congratulations!
I started with a vty command to force an ms power update on a timeslot and then I implemented a very crude loop and that "works".
excellent.
Next, I imported power_control.c from osmo-bts into osmo-bsc. This is where I began to wonder if this code is better moved some place where it can be shared by both projects. - There would be some small modifications necessary, I think.
it would have to move to libosmogsm, if it can be shared without making it awkward for either of the two programs to use that shared code (i.e. due to tight integration with application-specific data structures).
I also wonder if it is the right thing to do to integrate this bsc-side UL power control into the current (m|b)s-power control? Would it make sense to add a further "mode" there?
ms-power-control mode dyn-bsc
For example?
Obviously one does not want both a BSC and BTS managing a power loop. Or maybe only allow the BSC side power loop to run with bts-type that is known to not do UL power control?
What about hiding that from the user? OsmoBSC should know if the BTS supports power control or not. If not, then the BSC transparently implements it, based on the same set of parameters.
Basically, the location of the loop execution is an implementation detail to the user, I guess the user doesn't care where it runs, just that it runs as configured.
Regards, Harald
Hi,
On 9/23/21 23:08, Harald Welte wrote:
What about hiding that from the user? OsmoBSC should know if the BTS supports power control or not. If not, then the BSC transparently implements it, based on the same set of parameters.
Basically, the location of the loop execution is an implementation detail to the user, I guess the user doesn't care where it runs, just that it runs as configured.
I disagree here. MS Power Control loop support is not just "it supports it" or "it doesn't support it". There may be several different implementations, or similar implementation may have bugs in a given version, etc. Hence it really makes sense for the user to decide whether the BTS or the BSC should be in charge of the power control loop for each BTS, regardless of the BTS supporting some sort of MS power control loop.
Hi Pau,
On Sat, Sep 25, 2021 at 10:02:03PM +0200, Pau Espin Pedrol wrote:
[...] Hence it really makes sense for the user to decide whether the BTS or the BSC should be in charge of the power control loop for each BTS, regardless of the BTS supporting some sort of MS power control loop.
I see your point, and it makes sense.
-
Harald Welte -
Harald Welte -
Keith -
Pau Espin Pedrol