Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[snadge]

I've been told that ISPs (or maybe just PlusNet as this was on the PN forum) that they (ISPs) dont use Profile 8a etc for 40Mb connections, they use 17a and "band" it down to 40Mb...

is this correct?

thanks

[gt94sss2]

Openreach, who own the FTTC DSLAMs, use Profile 17a for FTTC connections - they then apply a software cap if the service ordered is lower than the sync speed achieved.

The ISP doesn't control this itself.

Openreach did initially use Profile 8a for all connections but switched some years ago to 17a instead.

[j0hn]

Doesn't matter if it's 40/2, 40/10, 55/10 or 80/20, they all use 17a.  There's a newer 30a, but OR won't be deploying it and are focusing on G.Fast instead.

[ejs]

I don't think profile 30a is any newer, it's just not generally used. Profile 35b is the newer one.

[Chrysalis]

yeah check the 2nd slide on here, 35b looks to have no commercial future.

http://www.slideshare.net/Sckipio/gfast-vs-vdsl-35

also a pic of the newer jdsu type unit black sheep will be carrying?

http://www.prnewswire.com/news-releases/intec-presents-argus-tester-for-vdsl2-profile-35b-and-gfast-at-the-broadband-world-forum-in-london-596279141.html

[ejs]

I think Deutsche Telekom have a different opinion on 35b a.k.a. SuperVectoring.
http://4gbb.eu/images/GOLD_Publications/TNO_UFBB_seminar_June_2015/2015_06_Marko_Loeffelholz.pdf

Sckipio of course only make G.fast stuff.

[Chrysalis]

true, I read a sales man's slide

g.fast one can argue is more future proof, but it does require shorter loop length's to shine.

I suppose the difference here is what the german telco has already commited to a proper FTTC rollout (aka vectoring), whilst BT did their's with only half commitment essentially deciding they dont like the tech enough to put further money into it and instead move onto what allows them to market higher speeds, even if it benefits less lines.  Thats the nature of the market we have in the UK.

[burakkucat]

also a pic of the newer jdsu type unit black sheep will be carrying?

http://www.prnewswire.com/news-releases/intec-presents-argus-tester-for-vdsl2-profile-35b-and-gfast-at-the-broadband-world-forum-in-london-596279141.html

I remember checking those Intec "Argus" HHTs, about two years ago, to see if they would be a worthwhile addition to Walter's wheel-barrow. The available features were good but the purchase price was prohibitive! 

[PhilipD]

Hi

true, I read a sales man's slide

g.fast one can argue is more future proof, but it does require shorter loop length's to shine.

I suppose the difference here is what the german telco has already commited to a proper FTTC rollout (aka vectoring), whilst BT did their's with only half commitment essentially deciding they dont like the tech enough to put further money into it and instead move onto what allows them to market higher speeds, even if it benefits less lines.  Thats the nature of the market we have in the UK.

It's interesting to note that VDSL can match speeds of G.Fast, the new 35b version of VDSL certainly can.  The bit-rates are mostly if not almost entirely decided by the frequency range used, so they don't tend to differ too much.

G.Fast was designed to be used close to the user premises, which is why it uses time division multiplexing (this works best when lines are a similar length and short), and time division multiplexing allows for easier power saving options (at the cost of latency) which was important if powering small G.Fast nodes by sending power back down the telephone line.  G.Fast time division multiplexing also allows symmetrical download and upload speeds, like 100/100, but that is harder to achieve when you have 100 G.Fast connections all terminating back in the same cabinet as noise and crosstalk is at it's worst, a problem that was largely avoided using smaller nodes.

BT using G.Fast from the same VDSL locations so line lengths are not reduced means G.Fast is not the better technology to use, but VDSL 35b would be better.

Basically the money was spent looking at G.Fast with nodes closer to premises, they then decided they didn't want that investment so BT went for a cheaper option of using G.Fast from existing FTTC cabinets (or close by), so have ended up with the wrong technology for that.  We now have the additional problem of co-existence of VDSL using frequency division and G.Fast using time division emanating from the same location at the cabinet.

It's a Frankenstein monster of a decision.

Regards

Phil

[Chrysalis]

Think most people agree with you Phil.

I think whats caused BT to go down this route is they have a mindset of needing to sweat assets for as long as possible, the thought of replacing cabinet equipment which is less than 5 years old probably terrifies them.

We will never know if closer g.fast nodes were ever planned in the early stages, my opinion is they wasnt, but it looks like the decision is set in stone now with contracts signed with g.fast vendors.

[j0hn]

I don't think the current DSLAM's could use VDSL2 35b anyway. They can be made to run 30a, but 35b is much newer (I mixed them up earlier). If I recall correctly there was also only Broadcom doing 35b, nobody else. BT always use the dual vendor approach.

[burakkucat]

I have a vague memory that the BT Group plc planned to use the FTTC concept over a 25 year lifetime.

[ejs]

In the photos of ECI line cards, the stickers on them have a date of manufacture in 2012 and a warranty exp. date in 2017, so a 5 year life.

Edit: Add link

[Mark07]

Interesting, does that mean they all get replaced every 5 years then?

[Chrysalis]

The most alarming thing about that is that if the g.inp rollout fails this year then the warranty will be expired and ECI absolved of responsibility for a new firmware.

Mark07 more likely that BT will consider a decision to extend the support or just let it expire and pay per incident instead.