Superhub 200Mbit coming soon? (According to a survey) - Page 8

Kushan · 05-01-2015, 20:17

Quote:

Originally Posted by qasdfdsaq

That's about as likely as 10Gb on D3.1.

If they ever go 10Gb it'll involve such a big rebuild they may as well go FTTB or FTTP.

yeah that's pretty much my point, if D3 is capable of 1.5Gbit and we're sitting on 150Mbit (maybe 300 or even 600 in some years), then all the hoohah about D3.1's theoretical speeds is worth nothing.

Ignitionnet · 05-01-2015, 20:50

I think there's some confusion here.

Pierre wasn't saying that VM will deliver 10Gb via DOCSIS 3.1 this year or next. He was I believe saying that VM will be delivering 300Mb via DOCSIS 3, which they will, and that they are working towards 3.1, which they are.

The network is certainly going to have work done on it, however this work will not be anywhere near as extensive as an FTTP overlay, it will be closer in cost per premises passed to BT's FTTC build as it will entail conversion to an FTTLA HFC architecture.

Shorter term the main plan will be upgrading the network so that return is split at 85MHz or higher and downstream is at 1GHz which along with further resegmentation will push fibre deeper into the access network anyway. Between those, CMTS, and CPE upgrades delivering 1Gb/100Mb via HFC using nothing more than DOCSIS 3 becomes a possibility.

DOCSIS 3.1-capable kit is there right now with more being deployed.

Pierre · 05-01-2015, 20:52

Quote:

Originally Posted by Ignitionnet

I think there's some confusion here.

Pierre wasn't saying that VM will deliver 10Gb via DOCSIS 3.1 this year or next. He was I believe saying that VM will be delivering 300Mb via DOCSIS 3, which they will, and that they are working towards 3.1, which they are.

The network is certainly going to have work done on it, however this work will not be anywhere near as extensive as an FTTP overlay, it will be closer in cost per premises passed to BT's FTTC build as it will entail conversion to an FTTLA HFC architecture.

Shorter term the main plan will be upgrading the network so that return is split at 85MHz or higher and downstream is at 1GHz which along with further resegmentation will push fibre deeper into the access network anyway. Between those, CMTS, and CPE upgrades delivering 1Gb/100Mb via HFC using nothing more than DOCSIS 3 becomes a possibility.

DOCSIS 3.1-capable kit is there right now with more being deployed.

Ta,

I was just about to type out a reply to clarify where I was coming from.

pip08456 · 05-01-2015, 21:19

I would think they will as they know BT have trialled G.FAST and have opened a lab at Adastral Park to develop it further.

Source http://www.btplc.com/Innovation/Inno...fast/index.htm

Ignitionnet · 05-01-2015, 23:16

Quote:

Originally Posted by pip08456

I would think they will as they know BT have trialled G.FAST and have opened a lab at Adastral Park to develop it further.

Source http://www.btplc.com/Innovation/Inno...fast/index.htm

I doubt they'll be too scared of a product in production exactly nowhere with no finished final hardware and not on BT's roadmap for commercial deployment.

BT Labs are pretty busy with lots of new tech, such as VDSL and PON, still a way behind in deploying them commercially.

muppetman11 · 05-01-2015, 23:34

Quote:

Originally Posted by pip08456

I would think they will as they know BT have trialled G.FAST and have opened a lab at Adastral Park to develop it further.

Source http://www.btplc.com/Innovation/Inno...fast/index.htm

They haven't even managed vectoring yet , things continue to move at a glacial pace.

Ignitionnet · 05-01-2015, 23:37

Quote:

Originally Posted by muppetman11

They haven't even managed vectoring yet , things continue to move at a glacial pace.

Openreach and BT generally have never had any interest in competing with Virgin on speed and this isn't going to change.

Vectoring will happen as and when BT need it to in order to adhere to the Superfast Cymru contract and any others that name 100Mb as a requirement.

qasdfdsaq · 05-01-2015, 23:41

Quote:

Originally Posted by Ignitionnet

I think there's some confusion here.

Pierre wasn't saying that VM will deliver 10Gb via DOCSIS 3.1 this year or next. He was I believe saying that VM will be delivering 300Mb via DOCSIS 3, which they will, and that they are working towards 3.1, which they are.

Quote:

Originally Posted by Pierre

Ta,

I was just about to type out a reply to clarify where I was coming from.

Sorry. I was specifically referring to the first sentence that said "could eventually... 10Gb" so I assumed the response was directed at that.

pip08456 · 06-01-2015, 01:12

I would expect the roll out of vectoring to start in Q3 this year.

Pierre · 07-01-2015, 09:10

http://www.pcworld.com/article/28641...n-in-2015.html

Kushan · 07-01-2015, 09:54

Quote:

downstream speeds will still probably be about four times that of upstream speeds

I wish

qasdfdsaq · 07-01-2015, 14:09

I also like how 802.11n has now been inflated up to 1Gb/s in the 2.4Ghz band.

Ignitionnet · 08-01-2015, 13:49

Quote:

Originally Posted by Kushan

I wish

The ratios are way better this time around

392MHz of spectrum for downloads, 192MHz for uploads.

Currently on US cable networks it's more like 35MHz for uploads, not all of which is usable that's closer to 27MHz, and over 900MHz for download, of which modern CPE can use 192MHz.

Sightly better on Euro networks but even on the best of those the ratio is 60MHz, of which perhaps 45MHz is usable, to 256MHz via modern 32 channel CPE out of, again, over 900MHz.

Fully DOCSIS 3.1-ready networks have been in planning for a while as it's always been expected that at some point customers would be wanting much higher upstream speeds.

No reason why VM's plant can't be mid-split before my 40th birthday.

qasdfdsaq · 08-01-2015, 14:15

Quote:

Originally Posted by Ignitionnet

The ratios are way better this time around

392MHz of spectrum for downloads, 192MHz for uploads.

Well, that's the hardware capabilities of the CPE, but as you say the network itself has to be physically wired to support it... and that's not a small task.

The maximum hardware capabilities haven't ever been particularly relevant to the actual marketed end-user speeds. After all, previous (DOCSIS 1.0, DOCSIS 2.0, and debatably the VMNG300 D3.0) hardware have all been capable of much better ratios than current CPEs that actual products never came close to offering.

Quote:

Sightly better on Euro networks but even on the best of those the ratio is 60MHz, of which perhaps 45MHz is usable, to 256MHz via modern 32 channel CPE out of, again, over 900MHz.

And if a highly capable D3.1 CPE were to be deployed on the current networks within the next year or so, it would actually make it worse, not better, since the CPE would allow more downstream to be utilised yet the upstream spectrum would still be limited by the physical network.

Ignitionnet · 08-01-2015, 16:17

3.1 can offer an immediate improvement in upstream speeds over the current set up even without physical network changes. 3.1 also needs work on the downstream; there's a fair amount of linear TV and VOD there and finding the equivalent of another 33 free QAMs of downstream spectrum over the current band plan isn't trivial.

Regardless the CPE and standards support a far better practical upstream ratio than the predecessors.

It's also worth mentioning that the ratios deployed elsewhere in the world are quite different from those here which have historically been conservative on the upstream.

The USA come to mind as some who pushed upstream hard in the past. A 101/35 tier from Cablevision on DOCSIS 3.0 and 30/5 on US DOCSIS 2.0 come to mind.

VM and the rest of the Liberty Global group will be upgrading networks accordingly in any event.

It's a non-issue though as VM aren't going to be deploying a DOCSIS 3.1 product imminently, there's plenty of life left in DOCSIS 3.0.