[AusNOG] "i want a pony!" (was Re: Long live the NBN. The NBN is dead?! [personal])

[Mark Smith]

Hi Tom,

(Not specifically singling you out, as others have made broad
sweeping statements, although your association with the NBN might mean
you can provide more definite answers or examples)

On Wed, 11 Aug 2010 12:53:24 +1000
Tom Sykes <TomSykes at nbnco.com.au> wrote:

> David Connors writes.....
> > Another thing lost on the one-eyed-NBN-at-any-cost nerds arguing in this 
> > debate is that there are entire sectors of the population who don't give 
> > a toss about quality of service on the Internet- loading a web page is 
> > FINE for a lot of people who just want to tool around on Facebook or whatever
> 
> 
> 
> If you freeze the application capability at a point in time (e.g. Text Webpages) then we would never have required DSL.
> 
> If you are building this network for "Facebook", then I mostly agree, you probably won't see much difference between your Internode 25M service and a Fibre 25M Service. However, making the "internet" faster is clearly not the only goal here. 
> 
> There are clearly other applications (and more to emerge) which either (a) simply cannot be provided using current infrastructure, or (b) may work, but suffer from the lack of determinism in the copper loop and can't be offered on a consistent basis.
> 
>  

Which is the sort of justification that politicians have been
making (in particular one of them), and is in effect "i want a pony".
For $43BN, I want a Horse!

For me to be accepting of the need for 100Mbps figure, I'm looking for
use cases, backed up by numbers. This might be hard for politicians to
do, but shouldn't be hard for that many people here - yet we have seen
a number of people not providing these examples, using
"emergent uses" or "because they've got it in Korea or Japan" to
justify it. I know they've Kittys in Japan, but do they have ponies? Do
they have any Joneses in Korea, because we seem to be wanting to
keep up with them?

Here is an example of what I'd be looking for and expecting. Call it
a component of a business case, if you like.

If the typical family in Australia has 2.4 children, rounding that up
to 3, that means in the slightly above average case there are 5 people
living in a residence. If each of those people wants to conduct a high
definition video conference at the same time, that is approximately 5 x
8 Mbps symmetric bandwidth [0], or 40Mbps. That is of course peak
bandwidth, and worst case. 3 children is not that common, and I think 5
concurrent HD video conferences is even less likely to happen. However,
it is a feasible and possible use case.

So what is the other 60Mbps for?

So where are the use cases with that sort of detail? My question from
this morning was to see if people on this list, who know what people
use networks for, could provide cases for or against 100Mbps, or 1Gbps.
Some people have come close, by pointing out applications that may use
up to 20Mbps - but that's a long way from 100Mbps, and, as I said
earlier, lets borrow another $5BN to get 1Gbps if we can if there are
cases for it. *Are there cases for it*? The ephemeral cases quoted for
100Mbps are equally applyable to 1Gbps! If we're trying to keep up with
the International Joneses, at nearly any cost (which from my financial
perspective is what $43BN is), then I want to keep up with the Swedish
Joneses, pronounced "Lothberg".

I'm on a Horse!

Regards,
Mark.



[0] 
"HOW MUCH BANDWIDTH?"
http://www.icf.at/en/6000/how_much_bandwidth.html

(Who new there was an organisation to promote the use of cables! Note
the figures they quoted were provided by the ITU)