Print

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

[WWWombat]

whats the rational behind stopping them putting vdsl stuff in the exchange? Seems they have a monopoly on the network regardless of where the equipment actually is.

No, BT don't have a monopoly on what is connected to the network - that ended with LLU. Sky and TalkTalk, for example, have ADSL equipment connected to the network at the exchanges ... and they could, if they chose to, put their own FTTC/VDSL2 cabinets out by the cabinets too. The original South Yorkshire project, for example, had VDSL2 cabinets, but they were not BT.

The ANFP acts as a safeguard, limiting what frequencies can be inserted into the network at different points. And the whole point of the ANFP is to limit frequencies at some locations so that they don't cause interference with "desirable" services at other locations.

A prime example is that VDSL2, at the cabinet, must limit its use of frequencies below 2MHz so as to not interfere with exchange-based ADSL services. VDSL2 ends up working at reduced power in that part of the spectrum (and so ends up 10-20% lower speed than theoretically possible).

This thinking is what prevents VDSL2 equipment being used in the exchange: IF VDSL2 was allowed to be placed in the exchange, then every VDSL2 DSLAM located at the cabinets would have to reduce power over the whole 17MHz spectrum, not just the 2MHz. The cabinet-based systems would be so degraded, they would work no better than an exchange-based one.

There are ways around this limitation. For example, if the only circuits allowed to use exchange-based VDSL2 were ones where there was no cabinet-based VDSL2. However, you create lots of problems if and when a cabinet-based DSLAM was installed: who would pay to move them? What if Sky ran the exchange-based DSLAM, and BT ran the cabinet-based DSLAM?

NICC did a study about deploying VDSL2 at the exchange in 2014, offering options. It has gone no further, suggesting there is little appetite to pursue the possibility.
http://www.niccstandards.org.uk/files/current/ND1517V1.1.1.pdf?type=pdf

One key problem they point to is that EO lines aren't always held as bundles on their own. Sometimes, they are bundled with some cabinet-based lines, making it hard to predict whether interference will happen. And, as you'd expect, records aren't always complete.

The ANFP forbids installing VDSL equipment in exchanges to prevent crosstalk with ADSL from interfering with the service of those remaining on ADSL - hence building the cabinets outside.

The key part of putting an FTTC cabinet outside the exchange is that they also feed only some of the cable bundles into it ... and specifically, non of the cables that will go on to later PCPs (with their own FTTC cabinet). The limits interference

So I guess its pretty much madness to think they would stand cabs outside the exchange to deal with (way) over capacity cabs further away?

Absolute madness. To minimise interference between two VDSL2 cabinets feeding the same coverage area, they need to be the same distance from the exchange, so they can transmit with the same power levels.

When BT introduce a new cabinet as infill for longer lines (ie 1km out from an existing PCP), such as an "all in one" cabinet, the cabling has to be designed to take over all the services from the original PCP. Anyone on those cables who had a (slow) FTTC service has to be migrated to the new infill cabinet before any new subscribers can be taken on.

[WWWombat]

I'll add that my downstream attenuation with ADSL2+ was 30db, and now on VDSL2 only using the D1 band is 30db still, I know the D1 band extends slightly higher in frequency, but shouldn't the near halving of distance caused at least a slight drop?

The attenuation figures are measured in different ways, so aren't directly comparable. Probably because the newer DSL variants use ever-higher frequencies, and attenuation is higher at those higher frequencies. Take a look at the "--pbParams" output, and you can see the attenuation increase across the spectrum.

Take a look at Hlog too. That is a form of attenuation, showing how it varies across the spectrum.

I have been over and over tracing the most circuitous routes my line could take given the visible poles from the road, and adding meters on at each one for going up and down, tracking across the road etc, the most I can come up with from house to cabinet is 1km. The exchange is at most a further 700m.

On an old line, I used to have a 30dB ADSL attenuation, with a line length of approximately 2.2km. Maybe there really is 500m missing somewhere. Or perhaps there's another explanation...

The graph shown in a post further down this section: https://images.imgbox.com/a6/a0/7JTTcZOI_o.png, would put a 1.7km line at just a hair under 20mbps predicted.

The graph is rather idealistic. It assumes that the line is made of copper, with a diameter of 0.5mm.

Unfortunately, real life isn't so homologous. There's some amount of aluminium cabling around, which doesn't work as well as copper at voice frequencies, and fares worse again at higher frequencies. And even copper isn't always 0.5mm (though most new cabling *is*).

Here, it looks like lines are made of 0.4mm copper, which behaves like 0.5mm copper that is 50% longer. That shows as slower speeds on both ADSL and VDSL2.

In some places, thicker copper is used (0.6mm and 0.9mm?), which behave like 0.5mm copper that is shorter.

It is hard to say how much of the network is affected by these issues, but that only affects nationwide statistics. All that matters for one home is what *that* line is made of.