DOCSIS 3.0 moves the shared upstream from 6-28 megabits to 120 megabits. It’s already built in to all DOCSIS 3.0 modems DOCSIS is designed to deliver 50 up 50 down 95+% of the time. 4-12+ times as fast in raw speed. The real difference is likely to be higher, as congestion peaks tend to even out with more raw speed. Few really need that speed, but nobody really needs a BMW either. The evidence is overwhelming that almost all customers prefer high speeds if the price is similar, while a significant minority will pay substantially more.

Cablecos fighting DSL will have a huge advantage. Cablecos facing Verizon will need 3.0 upstream to compete with GPON designed to go up to 200 meg symmetrical, and even BPON is 155 meg shared. Ivan Seidenberg is spending $23B for FIOS. He’s told me “we need to get cable out of the house.”

He’s already selling 20 meg upstreams when few cablecos even offer 5 meg. In New York City, it will be almost exclusively GPON and he’ll rapidly offer higher speeds.

Facing AT&T, Qwest, and Bell Canada, the 50 meg of 3.0 upstream will provide an enormous advantage compared to the 1 meg of DSL from the node. 100 meg DOCSIS is selling for $25-40 in Japan and France as part of the bundle. Any cableco who prices like that should knock out a telco without fiber. Other carriers will decide to reserve the high speeds for more expensive tiers, directly improving profit.

Besides blowing away DSL 1 meg upstream, it also should essentially solve the congestion on the cable upstream. That’s a far better solution than throttling or capping.

The tech community has realized DOCSIS 3.0 is much faster than DSL since about 2003. DSL like AT&T is delivering 1 up, 10 down, 50-98% slower. The telcos sticking with DSL for the last mile will have a major technical disadvantage for the next decade. If customers desert in droves for the higher speeds (unproven, especially if they are priced high,) it would take years to move to full fiber.

source: dslprime.com