The 35b profile with active DSLAMs in neighborhood cabinets "may" reach 300mbps with a fully vectored network (but just a couple of loops not being vectored greatly reduces speeds [1]). It still requires lots of infrastructure upgrades, replacing the old passive cabinets with bigger active ones containing DSLAMs, laying lots of fiber around etc. I still think its a dead end and countries like Greece that only now are installing such networks are making a big mistake.

FTTB, either through G.FAST OR Ethernet is a better use of all that money in my opinion and it's a lot easier to upgrade in the future (either the building terminal or switcing to full fiber i.e. AON/GPON etc).

[1] https://i.imgur.com/39pCRjN.png

I wonder about that. Obviously the future of internet connections is wireless. Why are we still bothering to hook up individual houses. Apartment blocks, perhaps. Houses ... why ?

> VDSL is certainly not a "dead end". 50Mbit is available in most of Germany, with 100Mbit slowly rolling out.

Using increasingly involved techniques to get more speed out of a century-old technology is pretty much a dead end.

100Mbit is decent, but it comes at the price of * the wires needing to be very short by telco standards * the use of DSL vectoring, which creates mini-monopolies in the DSLAM and allows one provider (usually Telekom) to hold all users on that DSLAM hostage

And keep in mind that if we had fiber, we could probably get a Gbit.

Make no mistake, the strong position of DSL in Germany is the result of politics, not common sense.

In the US, 100 Mbps availability hit 60% in 2014, and the next milestone is gigabit.

How often is the performance difference between 100 megabit and gigabit internet all that measurable, though? For that matter, how often do truly get even, say, 900 megabits allocated to you on that sort of connection?

I'm not doubting they actually give you a gigabit connection, but especially during peak hours, I doubt the transit links you hit would necessarily have that much to spare.

> How often is the performance difference between 100 megabit and gigabit internet all that measurable, though?

For doing backups/restores, in this age of multi-terabyte disks, fast network links (especially in the upload direction) are a necessity.

I don't mind some oversubscription, I can work around peak hours, hell, I'd love to temporarily trade away most of my downlink bandwidth in exchange for uplink bandwidth -- average upload speed is all that matters for this sort of large transfer.

It's the tail end of the PM peak here, and I'm getting 420 mbps to my web host, about 450 to Azure (on a 500/500). I think it's growing pains from Verizon's gigabit rollout, because when I was at 300/300, I consistently got over the rated speed.

In the US, if you don't have symmetric streams, the average high-speed stream is 6mbps/100mbps up/down. The next milestone is still going to discriminate against content producers, IT workers, and software engineers.

Compounding that, the latency comcast deliberately introduces into their network (I've got many ISP's to judge on here) makes the internet ugly, but only if you don't pay for their services.

What we need is not a target for faster downstream, but higher competition - I could give a shit about network neutrality if the providers offered network honesty: a 500 GB monthly limit is a 1.5mB stream - full stop.