Is U2 being higher than the D2 and D3 indicative of anything or just considered a general part of this profiling?
It is just a result of the applied DSM profile.
I'm I right to presume that the -140bDM/Hz noise floor because of POTS services mean we are at the upper frequency range with VDSL2 (17a) so POTS on the same line makea 30a a no go?
The QLN plot shows how quiet the circuit can be across its defined bandwidth.
Consider the y-axis of the plot. The higher any one point, the noisier the circuit; the lower any one point, the quieter the circuit.
An absolutely perfect QLN plot would be a horizontal, straight, line. With a POTS service "below" the xDSL service that theoretical horizontal line would intersect the y-axis at no lower than -140 dBm/Hz. When there is no POTS service sharing the metallic pathway, that theoretical horizontal line could intersect the y-axis < -140 dBm/Hz. (I.e. even quieter.)
The above has no implication for the profile of the xDSL service. (This profile, as in 8a, 17a or 30a, is not "that" profile, the DSM profile.) When VDSL2 services were first made available in the U.K. (back in 2011), they were provisioned as Profile 8a. Once sufficient real-time data had become available (the BTagent, in the original white Openreach modems, "called home" with snapshots of the circuit's statistics) and was analysed, all VDSL2 service were then upgraded to Profile 17a. To subsequently change to Profile 30a would give all VDSL2 circuits even greater bandwidth . . . but at the expense of the bandwidth for G.Fast circuits.
A decision had to be made and that decision was to not increase the VDSL2 bandwidth.