cheers for that, I knew there was different types and some was better than others. Sky seems to have got their's right as i know a few people on Sky ADSl and got longer lines than I have and got a far better sync speed than I did with BT.

I still think that we are better off most of the time without DLM, my connection with Cable and wireless was great, the only time it disconnected was when I pushed the SNR right down. Far better sync speed than I ever had with BT.

In reply to a post by lexden16:

By way of balance, apart from a blip in Feb, I get 75.3 Down, 16.9 Up with a Latency of 21 at all times of the day so I am very happy with FTTC. That said, DLM does have a marked effect when line conditions go awry. My ISP Line Data shows that it took 16 days from interleaving to drop off after a fault before I got back the full 80/20 profile. It was many days later before interleaving fell away fully and I got back my current max attainable rates of 93.5 Down and 28 Up. I am fortunate as I live c.300M from FTTC cabinet. That said, when people live much further way from their cabinet and the max attainable line speed is less than 80, I can well see why the effects of DLM can be so frustrating.

This is the problem and it seems like the ISPs don't care as long as you got a ok speed. Mate complained to Plusnet and yes they did send a Bt open reach engineer over, who said there was nothing wrong and he should be getting around 70Mb/s as that is the speed the modem is reporting or his equipment,
So i take it that plusnet is traffic managing way too much, they offer these faster speeds and can't cope with the amount of data.

System I am on don't have any DLM, it is 10Mb/s and that is it, ok it does go slightly higher most of the time, but it never drops unless they have a problem and that have not happened for months now.

May only be ten, but I know it is ten.

In reply to a post by zyborg47:

So i take it that plusnet is traffic managing way too much, they offer these faster speeds and can't cope with the amount of data.

I don't understand that comment at all.

Talking of DLM, my modem resynced at 4am today, as you can see from my BQM, and added some interleaving, increasing my latency and decreasing my sync speed by a bit. I'm not complaining, I'm still well above my estimate. This was after the longest period I have ever had without interleaving, 4½ months at the same sync speed. The PPP session didn't drop so I still have the same IP profile, now 99.08% of my sync speed

does this PHYR dynamically apply the latency overhead or just the speed overhead?

to me the latency penalty is bigger than the sync speed penalty as many people dont understand the impact of latency on speed.

eg. load a site hosted in asia and one might think asia has some tiny backbone or something and that why its slow, when the real reason is the high latency means the tcp ramps up so slow that the effective speed is slower.

another way I can put it.

a 10mbit connection with 10ms latency to say london will load most websites faster than a 100mbit with 30ms latency. The 10mbit connection will be faster than the 100mbit connection for pretty much all activity that deals with small packets and will hit its max speed quicker on downloads. The 100mbit would win on large file downloads. By small I dont mean tiny as in 1kB, even eg. images that are 250kB will load faster on the first connection as the speed will ramp up quicker.

this is especially apparent when uploading as many applications still use a hard set static low upload buffer size, meaning latency is king on throughput in those situations.

The reasons I consider BTw DLM better than openreach is simple, increasing the snrm still allows the user to overide, (better for user, worse for operator) whilst line banding is a hard cap on sync speed that cannot be overidden (better for operator, worse for user). BTw DLM prefers to raise snrm first ahead of applying interleaving, this to me is a better first choice than applying interleaving. Openreach prefers to apply interleaving ahead of banding. Finally BTw DLM can be reset by the isp, Openreach DLM cannot. BTw DLM also (as I remembered it on adsl max, not sure about 21CN adsl) doesnt forcefully drop line sync to apply new settings, they only applied when the line drops for some other reason, whilst Openreach DLM is disruptive it will drop a line to apply new settings. Ironic since DLM is supposed to help stability yet it causes instability to apply settings.

To me these are all logical reasons why the BTw DLM is better, yet many on here consider openreach's better, I would like to know why.

Edited by Chrysalis (Sat 31-Aug-13 22:24:48)

In reply to a post by billford:

The ADSL one is owned and operated by BT Wholesale, FTTC is Openreach's and they have very little in common.

But don't let me confuse you with facts if your mind is made up.

What about ADSL2+ DLM run by BTw or BTo?

BTW. But the IP Profile part of it is also applied on FTTC, at 96.79% instead of the ADSL2+ 88.2%

Edited by RobertoS (Sat 27-Sep-14 19:40:12)

ADSL2+ 88.2% is pretty low. Should be 96.79%. Would be nice if both are at 98%.

Edited by adslmax (Sat 27-Sep-14 19:55:20)

ADSL2+ has inherent overheads from using ATM. The IP ADSL2+ IP Profile largely absorbs that and also the TCP/IP overhead.

On O2/Be that didn't have an IP Profile, the highest possible data throughput was about 85%. As the best ISP can put data through BT Wholesale at nearly the IP Profile it isn't an issue.

FTTC doesn't have ATM. Only TCP/IP.

In reply to a post by adslmax:

ADSL2+ 88.2% is pretty low. Should be 96.79%. Would be nice if both are at 98%.

When your out of contract I'll happily give you service and access so you can monitor and sort your own line.

As you know

There is a good A&A article http://aa.net.uk/kb-broadband-how-atm.html which describes why there is such a seemingly large difference between line sync rate and IP throughput for ADSL services.

VDSL2 (aka FTTC) uses PTM which has a significantly lower overhead than ATM so the maximum attainable IP throughput is much closer to the line sync rate.