FTTC speeds

My BQMs showed a brief disconnection last night in the wee small hours (confirmed by the router log), nothing of note. Until I looked more closely and saw my latency had dropped a bit... so off to the BT speedtester which showed a profile of 70.8Mbps.

The BT estimate for my number is "up to 54.8Mbps". A bit of conservatism is fine, but that's ridiculous!

More to the point, 70.8Mbps is the highest it's ever been; even when I first went to 80/20 about 18 months ago it was only about 68-69Mbps (as far as I can remember), and it was down in the low 60's for a while (around 8 months ago). I find it hard to believe that the number of users on my cabinet has dropped since then... could BT be doing some unpublicised field trials of vectoring, and could this explain the above?

(The modem is a bog-standard HG612, one of the early ones that were mostly swapped out because of over-heating, but mine escaped)


eta- Or has anyone got a simpler explanation?

Edited by billford (Thu 08-Aug-13 14:19:22)

Vectoring is about much larger boosts in speed then just 1 to 2% so probably the DLM has simply relented slightly on the line

Could be as simple as a noisy local plasma TV turned off while neighbours are off on holiday for a few weeks

In reply to a post by MrSaffron:

Vectoring is about much larger boosts in speed then just 1 to 2%

OK, that's a good point

so probably the DLM has simply relented slightly on the line

That would make a pleasant change- it usually dumps me on interleaved if someone sneezes too hard

I'll see how it behaves over the coming weeks as the holiday period tails off... I'm about 450m from the cabinet, so quite a few houses to go past!

Edited by billford (Thu 08-Aug-13 14:41:54)

Would vectoring change latency? Probably not, or if anything it might increase it marginally.

The drop could be a reboot or upgrade at a node somewhere in the network and if one node in the path is removed a small decrease could occur. Your speed, might have been sitting there waiting to improve slightly for a few days ...

In reply to a post by MHC:

Would vectoring change latency?

I've no idea

The drop could be a reboot or upgrade at a node somewhere in the network and if one node in the path is removed a small decrease could occur.

Could have been... but it looked more like a (DSLAM initiated) re-sync was involved rather than a PPP drop due to re-routing. Can't be sure because I was pounding my ear at the time.

Edited by billford (Thu 08-Aug-13 14:53:14)

In reply to a post by billford:

In reply to a post by MHC:

Would vectoring change latency?

I've no idea

I would guess it might increase it.

Simplistically, every line in the cabinet need to be sampled, that data processed and the individual calculated inverse signal for each line injected. That will take a finite time, so for it to work properly the actual modulated data stream would also need to be delayed to align with the forecast.

In reply to a post by MHC:

Simplistically, every line in the cabinet need to be sampled, that data processed and the individual calculated inverse signal for each line injected. That will take a finite time, so for it to work properly the actual modulated data stream would also need to be delayed to align with the forecast.

That's true, but it's essentially an analog signal system- one advantage of analog over digital is that it doesn't calculate, it just behaves as dictated by the component values so, in some applications, it can be faster.

But as all I know about vectoring is how to spell it, that's all purely hypothetical