Print

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

[WWWombat]

The Openreach Fibre Index was first introduced in mid-November, with data related to Summer 2015. Some of the data caused a bit of discussion.

The index has quietly received an update for Autumn 2015, showing the average usage has increased from 190GB to 210GB (and qualifies it to mean "per month" this time).

The graph still talks of 40% year-on-year growth, yet the new figure (4x as much bandwidth over 3 years) is closer to 60%. I still can't fathom that - unless the 40% refers to core traffic, or aggregate traffic, rather than behaviour of one line or the average line.

[burakkucat]

. . . I still can't fathom that . . .

If you are unable, then I have absolutely no hope in doing so!   

I have come to rely upon your analyses to aid my understandings. 

[ejs]

That 210GB figure per month is quite a lot higher than any of the average monthly usage figures from a graph in Ofcom's Connected Nations Report 2015. Presumably the Fibre Index is using more recent data, but there still seems quite a large difference between the usage figures.

[niemand]

Given that the majority of customers are on 40Mb and that that number is way higher than VM's average I am a bit confused by that one.

[WWWombat]

Two answers:

1. I can think of one major difference with VM... BTs television offering is over IP. Same for TT, Plusnet and Sky's NowTV.

I think I'd expect to see that difference.

The Ofcom average might be explained by that too - as VM will be a part of the average.

In fact, I think you can see the effect of VM on that Ofcom graph ... Look at the extra-low values at precisely the 20, 30, 50 and 60 speeds. I think almost all VM users are encapsulated in those 4 bars - whereas most FTTC users will be spread evenly in all the bars in between.

(I expect to see VM in all 4 of those speed points because VM will have been slowly migrating users)

2. Clutching more at straws with this one...
If Ofcom's data is more oriented at internet access, or counted by the ISPs, they might not be counting multicast TV data at all.

I have no evidence that BTs figure includes multicast either, but I suspect so. Especially after seeing Doug William's presentation including analysis of their busiest 2,000 lines - and the video/non-video balance.

[niemand]

That would make sense. VM's standard VoD and TiVo IP TV services do not use the same connection as the broadband modems so won't feature.

Some of it is probably also a result of VM offering nothing but higher speeds. Those who barely use the Internet but want it there have to go superfast. Much of that is mitigated in Openreach areas as people will usually have the choice of ADSL that is 'okay' for light usage and then the SFBB options, and go for the ADSL until their usage justifies the FTTC.

They say it's an average which I take to be a mean. Median would be useful as well. If only I had all these contacts in BT I apparently have I would ask for both.

[kitz]

I too would suspect that TV services may make up the difference.

Even if it didn't - what portion of VM customers also take up their cable service. 
We dont have much VM in this area, but those people I know that do have it also have cable TV.  Yet those that do have DSL not all take a TV service, but do use methods such as iplayer.    I know my own iplayer useage can rack up a fair bit and I dont really watch that much TV.    Need some nationwide figures really rather than general assumptions :/

The graph still talks of 40% year-on-year growth,

I tried to get my head around it last time how they arrived at 40%, so shan't even attempt to figure it out again.   
 

[WWWombat]

The graph still talks of 40% year-on-year growth,

I tried to get my head around it last time how they arrived at 40%, so shan't even attempt to figure it out again. 

In today's quarterly results, there is a reference to the fibre index. It specifically says that the "40% YoY" figure comes from comparing usage to one year earlier ... rather than an average over the entire period.

This seems to be a lot closer than anything I have been able to figure.