Lesson 1: Never let me behind the camera ever again.

Lesson 2: Niall Murphy is a very knowledgeable and engaging man.

1. I think it's rather good camera work...
2. Yes he is. This 10 mins is editted from over 40 mins of coversation. Some of what had to get chopped almost made me want to cry. I may have to do a 'key quotes' follow-up.

Basically, expectations of mobile data are waaaay too optimistic.

To reiterate: the total mobile broadband capacity of all 5 networks in the UK is roughly the equivalent of that delivering fixed broadband to Slough.

The crunch has already arrived, if you ask i3GS users in London, at peak times. Next 2 years? expect to see time-of-day / location-based tariff loading. You get a bucket of credits, and it costs you more credits to download that 5MB podcast in the CBD at lunchtime than it does to do it at the park on a Sunday.

/m

Which will feel like a horrendous step back, although I accept that right now the earlyuu adopters are probably getting more than their fair share. But given how poor MNO's billing systems can be... could they handle this complexity?

Billing types have been talking about this for years - there are historical precedents in 'homezone' implementations where you were billed less or free within a certain geographic area. They never took off in Europe. Matching Time-Of-Day with Cell ID in the back end isn't rocket science. Pushing it back to the device and displaying it in a user-friendly manner however, that's another boiled kettle of fish to watch.

The data equivalent of classmarking could be applied to provide 'premium' customers with a better experience under congestion scenarios, but that's a pretty blunt instrument.

Expect things to go pear-shaped end of this year, then a period (to be known as the 'dark ages') before widespread LTE network and device availability / billing system enlightenment kicks in around 2013.

Looks like I've been vindicated: http://business.timesonline.co.uk/tol/business/...

Expect this in the UK within a year if it works.

Isn't it most lilkey to be "congestion-based" traffic management, where the user (depending on his/her service tier) is throttled at busy hour for the cell/area in question, but can then burst higher as conditions allow?

You can skin this cat a number of ways. The one most easily understood by punters will win.

[No cats were hurt in the production of this blog post. Stunt cats were supervised by professional wranglers under the inspection of the AHA]

Working this out so we (users) can understand what's going on, even notionally, is a very good point. We've done some consumer research into this (what people would prefer, etc), but even phrasing an intelligable question was tough.

Also, net neutrality concepts are involved.