That's just your assumption. It's much more likely the current DLM system was tweaked to accommodate the lower downstream SNRM target. It's a huge leap to assume the entire algorithm has been changed, the way banding is handled, and some new 2 day stepped approach added.
I agree with you here, to a reasonable degree. But I also agree with Mr Saffron:
Reading into specification documents information that is not there is a dangerous game to play and leads to urban myths perpetuating for many years potentially
My observations over the last few years suggest that BT have made rather large changes to the DLM algorithm - particularly over the last 3 years - and not just small tweaks. Those changes have then been introduced alongside each of the attempts at G.INP, but I strongly suspect that major changes have been brought in as a result of the Assia patent row.
We have certainly seen changes in the way that DLM resets occur and, alongside, the nature of banding. Banding is more likely to be introduced on a line than 5 years ago, and less likely to be removed. It is also more likely to survive a DLM reset.
But whenever DLM was going to reduce or remove banding (if ever), then it would work in a stepwise manner. In early days, it would tend to follow a small subset of speed steps, but has more recently changed to steps of 2.5Mbps (or alternate between 2.4 and 2.6Mbps). It still appears to be stepped.
With a new attempt to rollout G.INP again, and a parallel rollout of the 5-4-3dB trial, I'm expecting the pattern of the last 3 years to be repeated: that the nature of DLM changes markedly again.
Unless you know for a fact something else has changed with DLM then suggesting such changes have occurred is likely to cause confusion for others.
Absolutely.
DLM is, at best, an unknown quantity. And we can expect it to be in the midst of a change in nature again.
Making a definitive, unqualified statement about DLM is probably unwise.
Pages in this thread: 
Print Thread
