Print

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

[Chrysalis]

With the recent topic I just replied to and some I seen elsewhere on the net its clear DLM will at times apply very heavy stabilisation settings to lines, my question is do we know if BT have a limit to how far they go with DLM, I mean does DLM eg. send out alerts to BT if it has to use 40% overhead on a line for FEC.  Or is this considered fine and dandy by BT to operate some lines like this?

[ColinS]

Don't know Chrysalis.

I can only give an insight based on my own line.  At some point, pretty much after DLM returned me to 80/20 non-interleaved after ~56 days of banding caused by, let's say, external factors, it decided to switch interleaving and INP on.  The sync reduction wasn't much, certainly nothing to get concerned about, so I left it.  It then seemed to fine tune it a bit, reducing the interleaving depth.

So, currently, I get about 79 out of 80 Mb/s DS.  Interleaving depth is 1237, the overhead 20%, INP 3, and the delay 8ms.  I don't play games (unintentional pun there), so I don't care about 8ms delay, others might.

Looking at it with BE1, we saw that occasionally, just very occasionally, and for no longer than say 1-2mins, FEC rates would hit the roof, e.g. 3Million/minute.  All the rest of the time, almost nothing.  But the good news is - the technology works.  Every single one of those transient 3 Million FEC/min was successfuly recovered.  No CRCs, no ES.

I have absolutely no idea why this happens occasionally, and even less chance of finding out.  Am I worried about it?  No, not at all.  If it got worse, perhaps I would.

[kitz]

Im not aware that there is anything anywhere that lists the stages DLM goes through.   If there was Im sure some of us here would love to know how it works and in what stages and what the exact triggers are.  40% seems rather a lot for overheads especially when BT have always tried to reduce bandwidth on the exchange backhauls.  I guess this will somehow be tied in to the next stage of banded profiles.  :/

[ColinS]

I guess this thread was triggered by this one http://forum.kitz.co.uk/index.php/topic,12834.0.html

The 40% overhead is just a result of the RS correction embodied in the large interleaving depth (R/N).  But, if it was needed at the time ....

If it's no longer needed, then DLM will reduce it in time.  AFAIK it's probably less than 24hrs since it last intervened, and that was probably based on the state of things before then.