Danish Vectoring trials & deployment

There seems to be some news that TDC in Denmark have been given approval to rollout vectoring from the start of 2015. The restriction seems to have been in getting approval for VULA unbundling, rather than a technical holdup.

When looking at these, I found a report on the result of trials that took place in 2012 and early 2013 in a village called Folby. The results are in English:

https://marketingwholesale.tdc.dk/assets/tdcws/14113...

The trial shows speeds of around 90-100/35 at 500 metres, and 45/12 at 1km.
There is also one result of 22/2 at 1500m.

Is it comparable to UK?

The tests used the 17a profile with 998ADE17 bandplan, and used PSD masks to prevent interference with ADSL2+ frequencies. That's good for comparisons, but I think the ADSL2+ modems are in the same DSLAM as the VDSL modems, so the PSD masks won't be identical (in particular, I think it makes the upstream result for the 1500m property very low)

The result at 1500m used the 8b profile, so might be less comparable.

Unfortunately, there is no indication of what gauge copper is used.

Interesting find, it also mentions the superiority of G.INP is demonstrated over FEC/Interleaving.

Not just limited to Broadcom chipsets. The 7390 uses Ikanos

http://www.avm.co.uk/en/press/announcements/2013/201...

Edited by lexden16 (Fri 25-Apr-14 19:38:48)

nice find, come on openreach stop been slowmo's. Seems they want to wait for everyone else to try it first and be ultra conservative.

In reply to a post by WWWombat:

Unfortunately, there is no indication of what gauge copper is used.

I've found the answer to this part...

The graphs in the report actual use a measure of "attenuation at 150kHz, in dB" rather than distance, but the text gives a conversion factor into metres.

It turns out that the conversion factor they use, where 1500 metres = 12 dB (and 400m = 3dB), means that the copper is 0.5mm

That makes the trial fairly equivalent to the UK access network.

Useful to know where we may be if/when Openreach deploy the technology commercially here. Thank you.

Very interesting - especially as it notes in slide3 that there was no unbundling present - I wonder how you cope with that!

An update here...

A further trial is ongoing in Denmark (and reading some of the problems makes me understand why BT continue their trials).

Status report:
https://wholesale.tdc.dk/Produkter/Vectoring_OEBR5_s...
It is in Danish, but a couple of good rate-vs-reach diagrams (conversion factor: 12dB == 150m)

Google can run a translation: put the URL into the text box here, then hit "translate".

From the first trial in early 2013, it now looks like they'll go live with their first locations in July 2015. They look to be converting 400 "nodes" per year up to 2019, but I think some of those nodes are nearer exchange-size, rather than just cabinet-sized.

Implementing vectoring was never going to be easy. Among other things, to do it well, I believe it needs reliable cable configuration information and this has to be integrated into the vectoring configurations. One of those tedious bits of detail that turns out to be an enormous logistical exercise in data cleansing and systems integration.

Whoops.

That should read 12dB == 1500m, as stated earlier in the thread.