Print

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

[Chrysalis]

Sadly i feel underwhelmed, as the answers given are very vague and there is little there that we didnt already know

But I appreciate that you took the time to get answers on ECI even tho he was resistant to giving you an answer.

[MikeZ]

That's very helpful, kitz - thanks for making this information available 

[ktz392837]

Thanks Kitz very interesting.  I wonder if BT implements vectoring just on Huwais whether they can claim any higher speeds in adverts?  Assuming 2:1 mix if only 66% support vectoring will BT meet requirements for a mimum level to increase from the current 80/20 limit?

[GigabitEthernet]

I think BT might just roll out vectoring on the BDUK cabinets as a way to cover more properties...

[Terranova667]

Great to see the questions answered, thanks for posting that Kitz, that said a little disappointed that he seemed to avoid giving any detail for the ECI side of things, pretty much said what we already know.

It all looks great for Huawei users and i'm glad for them, ECI however for both G.INP and Vectoring it doesn't look good at all 

[kitz]

I wonder if BT implements vectoring just on Huwais whether they can claim any higher speeds in adverts?  Assuming 2:1 mix if only 66% support vectoring will BT meet requirements for a mimum level to increase from the current 80/20 limit?

I'm not sure.   If they did advertise it with the newer headline rate then,  I think it would make those (1/3rd?) of users on ECI's a tad unhappy.

There's no hard and fast rule about how much speed can be regained by the use of vectoring, but having lost 30Mbps off my headline rate and the average between users of this forum seems to be around the 20Mbps mark, then there are a lot of us impatient for the implementation of vectoring.

Its a tricky situation because if they aren't able to use vectoring on the ECI's, but rolled out to the Huaweis then it could make a two-tier situation across the UK.  Im on an ECI myself and although I would be green with envy if I couldn't get it, I don't feel that should be made an excuse as to why progress should stop for the Huawei's..  especially with some long BDUK lines it could make a difference between getting a decent speed or not.   

But then again its the shorter lines which lose the most from crosstalk and where vectoring can make massive difference.  For longer lines the potential increases arent as much.  There's a graph here produced by Alcatel. I've no doubt BT have also got some figures somewhere from the trials that they've performed on their own plant and network.

It would be good if BT were a bit more open about their plans for vectoring because there is so much myth and rumour about this particular topic.  I certainly dont feel that its a dead duck as far as BT are concerned especially because we know it could help in a lot of BDUK areas. 

But the debate on Vectoring and Crosstalk is cause for a whole separate topic

[NewtronStar]

Somebody heard my frustration as G.INP is back on my line

That's excellent your DS errored seconds are non-existent and a bloody good sync rate to boot, just keep the manual re-syncs to a minium and you won't see that terrible old interleaving wtih errored correction again 

[lf2k]

Whilst it's good to know all the Huawei cabs are done, and the ECI cabs will get done once the test equipment has been fixed, it kind of begs the question "what next"?

He was non committal re vectoring, and that appears to be a Huawei only thing.

Obviously G.FAST is next on the horizon, but that will only (probably) go to commercial areas and it's unlikely there'll be BDUK funding for that.

Given the rollout is still behind in terms of > 24Mbps coverage, and assuming G.INP could potentially be on all cabinets "soon", is there any real prospect that the target SNRM could be lowered to 3dB (or something else)?

6dB seems quite conservative...

[GigabitEthernet]

It would be interesting to see what the average SNR "swing" on FTTC lines is.

Looking at MDWS, most lines hardly change at all, so it would seem that 3dB may be justified.

[WWWombat]

Having come to this through pm, I forgot to post something about it.

I think that experience has told us how g.INP has worked, and Ian's answers are really just confirmation. In the end, I find myself mostly wondering what else kitz managed to talk about

Having said that, rereading the answers has left me with an interesting realisation...

The second point has two quotes that suggest a DLM reset now results in a different line profile:
- "Initially, low interleaving is enabled by default on all retransmission lines"
- "If a line is reset then it will initially move back to interleaved downstream and fast mode upstream." Followed by "Retransmission will be applied again after a few days."

That sounds like the new profile, post-dlm-reset, is to have "low interleaving" without retransmission.

I wonder if it means the kind of "low interleaving" we have seen alongside retransmission, with a configured delay of 0 that comes out to be 0.2ms. Or the old-style of "low interleaving" with delay set to 8ms.

Is anyone seeing either case in practice?

[GigabitEthernet]

When I had DLM reset on my line, the line went back to fast path for 12 hours or so before DLM switched on very high interleaving on the downstream (somewhere around 1000) and upstream (somewhere around 100).

It took a week or two for DLM to enable G.INP.

[WWWombat]

Hmmm. DLM intervention in less than 12 hours after a reset is definitely far from the norm. Or the old norm anyway ... so perhaps things have changed.

When interleaving was round 1000, do you have any idea what the "INP" and "delay" values were?

[Stroller]

We must remember <snip>.


Just thinking out loud . . . I wonder if there is a Kitizen who would be willing to install the OpenWrt software image onto an ECI B-FOCuS /r modem and send that modem to Kitz, our leader, for testing? 

b*kat has anybody offered this up yet or has it been 'overtaken by events'??   I have one, V-2Fub/r  Rev. B fresh from being replaced by a VMG8924 

Rgds.

[burakkucat]

b*kat has anybody offered this up yet or has it been 'overtaken by events'??   I have one, V-2Fub/r  Rev. B fresh from being replaced by a VMG8924 

As of yet, I am not aware of Kitz being offered an ECI B-FOCuS V-2Fub/r, flashed with the OpenWRT firmware, for testing on her circuit (which originates from an ECI equipped cabinet).

[Stroller]

Thanks for the update, b*cat. I'm happy to offer mine as a loan for her to test as she needs to.  I'll have a check through the guides for doing the OpenWrt flash and then I can send it on to here via whatever means is best for her.

Rgds,

Chris