Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[DeadMan]

Just curious. Does anyone know what happens if you were to unplug the modem for a day or more? Would the DLM treat it as a line fault and screw the line? Or would it just hibernate until the modem is detected again and go back to the same settings? Or maybe even go back to default settings after a long period of inactivity?

Maybe I should test the theory

[ryant704]

The DLM will monitor the connection 24/7 and apparently it makes a decisions every 15 minutes.

If is turned off for longer than 30, so the DLM has checked twice the logic behind it is that the user has turned the modem off.

At least that is what I believe... (read).

[kitz]

^ thats the theory that I read too.

It doesnt go back to the default settings though and it should pick up where it left off.

[burakkucat]

The important thing to note is that you should power down the modem before disconnecting it. The power down step causes the (Openreach supplied) modems to send a 'dying gasp' (as specified in the relevant BT SIN) to the MSAN which, in turn, recognises that modem signal as an indicator of a power loss event.

Having received a power loss signal, the MSAN sets an internal flag and stops counting the ES (and, thus, SES). If the modem is just disconnected from the line, the MSAN continues to count ES (and SES). The values stored in those counters are used by the DLM to determine what (wicked or evil) actions it may apply to the line.

[NewtronStar]

Just curious. Does anyone know what happens if you were to unplug the modem for a day or more? Would the DLM treat it as a line fault and screw the line? Or would it just hibernate until the modem is detected again and go back to the same settings? Or maybe even go back to default settings after a long period of inactivity?

Maybe I should test the theory

I remember last year when we had the wrong type of snow and eletricity went off for 3 days and when it came back on my sync reached its highest ever recorded since FTTC install so it can be good for some lines

[kitz]

I have wondered about the fact if you are going through a bad period of instability what effect it would have by turning your router off at every available opportunity - ie if you were out for the day or overnight.

The doc I read supposedly about the BToR DLM (not the BTw DLM) - suggested that the 15 min period monitoring took precedence.  The amount of errors per day were added up and divided by the online  period or something..  and if you hit 'x' then youd move into one of the monitoring brackets.  If you were online all day then you were more likely to reach target x.   I also saw something that suggested all power-downs were counted..  and exceptions were not made for the dying gasp signal (which seems a bit silly).  'y' number of disconnections would put you in the next bracket.    0 disconnections and 0 errors would move you up through the rankings and lower any previous restrictions much quicker.

Having read that about not taking into account the dying gasp, its why now if possible I will keep powered down for 30mins to ensure that at least one of the 15min periods of downtime is seen.

[burakkucat]

I think two facts should be clarified.

The Openreach supplied CPEs (either Huawei HG612 or ECI B-FOCuS devices) do send a 'dying gasp' signal. The Huawei MA5616 MSAN does recognise that signal, sets its 'CPE loss of power' flag and then stops counting error seconds.

I do not understand from where the Openreach DLM (a software utility) obtains its data, other than from the MSAN. Something just does not 'add up'. 

I agree that a minimum period of 30 minutes in the powered down state is well worth observing.

[NewtronStar]

I agree that a minimum period of 30 minutes in the powered down state is well worth observing.

But it is also worth observing when you power-up the FTTC modem, you don't want to be doing it in the evening 4pm to 11pm but your stats will give you a clue on the best time to re-sync "power-up"

[ryant704]

We still have know if the DLM uses the dying-gasp, yes it may well record it but it may not use it as any value.

[burakkucat]

We still have know if the DLM uses the dying-gasp, yes it may well record it but it may not use it as any value.

No, let me clarify. The DLM does not use the CPE's 'dying gasp'.

The CPE will output a 'dying gasp' when its power input drops below a certain threshold. The (Huawei) MSAN is known to set an internal flag to indicate 'CPE loss of power' on receipt of a 'dying gasp' and to 'turn off' the error-seconds counter.

The DLM (a software process) can, logically, only read the state of the MSAN's flags and counters.

[Chrysalis]

so is this right?

the dying gasp is still preffered to disconnecting cable as it reduces recorded errors on the DSLAM.

However it does not make DLM disregard the disconnection itself?  So the UAS, LOS, LOF counters still tally? is there a tally on the dslam for disconnections?

[Ixel]

Is this also the same for an ECI cab?

[burakkucat]

the dying gasp is still preffered to disconnecting cable as it reduces recorded errors on the DSLAM.

However it does not make DLM disregard the disconnection itself?  So the UAS, LOS, LOF counters still tally? is there a tally on the dslam for disconnections?

The DLM software looks at the ES, SES, etc, counters maintained by the MSAN and bases its decisions on how those counters have changed over a particular time-frame.

Those counters are not incremented by the MSAN, as long as the latter has received the CPE's 'dying gasp'.

We shall have to wait for Asbokid to return from his travels and then ask him to 'fill in' the answers to your other questions . . . He did offer to lend me a spare MA5616, for experimentation in The Cattery, but I declined that generous offer due to my lack of space.

Is this also the same for an ECI cab?

As far as I understand, the above process is identical for both Huawei SmartAX MA5616 and ECI Hi-FOCuS M41 MSANs. So 'yes' is the direct answer. 

[NewtronStar]

Hello burakkucat

Is the DLM a separate piece of physical hardware running the software and then its connected to control board on the MA5616 via lan cable  ?

[burakkucat]

Is the DLM a separate piece of physical hardware running the software and then its connected to control board on the MA5616 via lan cable  ?

That is one of Beattie's dark secrets, of which I am not privy. Sorry.