I agree that in some respects the article could lead people to perhaps misunderstand what it really means by mentioning increase in service speed... and the 1st paragraph where ECI is mentioned is also misleading as G.INP on ECI cabs is practically nil. So that is approx 1/3rd of UK lines still without G.INP.
(upstream ReTx has in the past been quite challenging to get working across their whole network but they have since solved some of those issues).
Erm.... none of those issues have been resolved and its why ECI lines don't have any retransmission at all.
What does cross my mind is what about those modems which are incapable of upstream re-tx. There are still likely quite a few ECI modems out in the wild.
I know when my daughter moved, the engineer installed an ECI modem during the period they were waiting for g.fast to be enabled. As there has not been any further f/w updates to the ECI modems, does that mean those lines could go back to losing up to 10Mbps of their downstream sync? Or perhaps they think that now the HomeHubs have been 'fixed' and that because the ECI's don't provide line stats, then no-one will notice.
This briefing is to inform all CPs about a trial Openreach will be testing the effect of higher levels of Retransmission ("ReTx") on the VDSL network
To me that would imply adding another re-tx profile by increasing the amount of impulse noise protection on any retransmitted data (DTU's). The benefit could be less (Huawei based lines) having traditional forms of error correction such as Interleave and RS encoding. Things may look good as far as stats go ie more people may see an increased sync speed, but it is highly likely to affect the throughput speed. It could also turn out that traditional interleave is preferable on some lines, as depending upon the type of noise, a higher level of G.INP could induce more jitter.
I wonder what the impact on the upstream throughput will be if all lines are to now have G.INP applied by default. Whilst there may be some coding gain, the actual overheads could be slightly more. Admittedly it's likely to be minute and I doubt the general public would notice if their sync is still showing as 20Mbps.
The main thing is that this is a
trial.. no doubt so that Openreach can assess some of those points and observe what difference it does make in the wild, before rolling it out fully.