TV Licence Super Thread

Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
It is already public knowledge that this is possible, though it isn't clear that the BBC are going to make use of this or if this is just going to be the next generation of 'TV detector van' myths where we'll never seen any evidence produced in court.

I can only see it being theoretically possible under very controlled environments.

It's akin to using the light flashes in a living room in order to determine someone is watching live BBC programming. In fact I think that would be more reliable as latency isn't a problem.
 
Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
no, it is practically possible with fairly simple equipment

How would it work when the router is feeding multiple connections at once? What if I am downloading something else? What if the router only submits packets uniformally ignoring the speed and order the BBC send the packets in? What happens with fluctuating latency and speed of an internet connection?

The signal in a BBC feed would get lost so easily.
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
How would it work when the router is feeding multiple connections at once? What if I am downloading something else? What if the router only submits packets uniformally ignoring the speed and order the BBC send the packets in? What happens with fluctuating latency and speed of an internet connection?

The signal in a BBC feed would get lost so easily.

downloading something else doesn't matter, this is just a classification problem - IIRC one PhD student demonstrated this working for detecting Skype calls using random forests, could distinguish between Skype and BitTorrent, he could detect different mobile apps etc..etc..
 
Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
downloading something else doesn't matter, this is just a classification problem - IIRC one PhD student demonstrated this working for detecting Skype calls using random forests, could distinguish between Skype and BitTorrent, he could detect different mobile apps etc..etc..

Had a feeling you were referring to academic research which doesn't reflect what is practical.

Now the BBC will be running machine learning algorithms on the fly and then present these black box algorithm findings in court as evidence beyond reasonable doubt.

Still doesn't touch the other problems. Like I said any signal will be lost and could only be caught at a probability way below 99%.

edit:

Distinguishing Skype and Bittorrent is easier than Iplayer and Netflix/Youtube/etc. as well.
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
Had a feeling you were referring to academic research which doesn't reflect what is practical.

Why do you think that? The guy was able to do this on a real network using a USB wifi dongle and a raspberry pi. This has been demonstrated to work, it isn't just theory.

Now the BBC will be running machine learning algorithms on the fly and then present these black box algorithm findings in court as evidence beyond reasonable doubt.

But I've not made that claim.
 
Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
You disputed my comment that this would only work under ideal conditions.

Let me know when iplayer can be spotted amongst youtube/netflix streams in a home broadband and router environment with greater than 99% certainty.
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
You disputed my comment that this would only work under ideal conditions.

Yes I did

Let me know when iplayer can be spotted amongst youtube/netflix streams in a home broadband and router environment with greater than 99% certainty.

Could be done now, it has been demonstrated already for say BitTorrent and Skype - 97% IIRC though some (probably a lot) of the false positives could probably be removed by considering the length of time the patter is detected for.
 
Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
Yes I did



Could be done now, it has been demonstrated already for say BitTorrent and Skype - 97% IIRC though some (probably a lot) of the false positives could probably be removed by considering the length of time the patter is detected for.

Why do you keep contradicting yourself. You are now saying this could be done in the real world because of some academic proof of concept?
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
Why do you keep contradicting yourself. You are now saying this could be done in the real world because of some academic proof of concept?

Can you point out the contradiction?

Someone has done this in the real world. They didn't teleport into some extra dimension to run this experiment.
 
Soldato
Joined
8 Nov 2006
Posts
22,979
Location
London
Can you point out the contradiction?

Someone has done this in the real world. They didn't teleport into some extra dimension to run this experiment.

First of all you gave show evidence of a different scenario in terms of packets. Bittorrent use is easy to recognise. Tell me exactly what was being downloaded and I'd be more impressed and more relevant.

Secondly was this done after going out onto the streets of residential neighbourhoods or in a controlled environment?
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
First of all you gave show evidence of a different scenario in terms of packets. Bittorrent use is easy to recognise. Tell me exactly what was being downloaded and I'd be more impressed and more relevant.

Secondly was this done after going out onto the streets of residential neighbourhoods or in a controlled environment?

It was Skype that they were trying to detect, which apparently can be confused with BitTorrent.

IIRC they used volunteers

They then also demonstrated detecting different mobile phone apps

point was just because a network is encrypted doesn't mean you can't make accurate inferences simply be looking at the encrypted traffic - the idea that player could be detected seems pretty plausible and this has already been demonstrated with Skype

I've not made any claims that the BBC is using this or that it will be used in court - I was skeptical about the whole thing earlier in the thread
 
Soldato
Joined
30 Nov 2007
Posts
2,989
Location
Bristol, UK
Can you point out the contradiction?

Someone has done this in the real world. They didn't teleport into some extra dimension to run this experiment.

Sorry 17 years specialising in Network Security suggests this is still rubbish. I'm sure all the nice gov people and companies would be jumping up and down if their encrypted stuff could recognised using a raspberry pie and dongle. When someone wants to post the article or explain technically some wishy washy stuff about recognising traffic types within an encrypted packet then feel free.
 
Caporegime
Joined
29 Jan 2008
Posts
58,912
Sorry 17 years specialising in Network Security suggests this is still rubbish. I'm sure all the nice gov people and companies would be jumping up and down if their encrypted stuff could recognised using a raspberry pie and dongle. When someone wants to post the article or explain technically some wishy washy stuff about recognising traffic types within an encrypted packet then feel free.

Well here is someone's thesis showing that encrypted stuff can be recognised doing exactly that:

http://discovery.ucl.ac.uk/1470734/...ivate User Information Despite Encryption.pdf
 
Soldato
Joined
30 Nov 2007
Posts
2,989
Location
Bristol, UK
Ok for the first pdf - I had to skim read 60 odd pages to get to anything worth reading. So the gist is that certain patterns can be gathered (based on the number of bytes flowing each way) for a protocol but some protocols cant be distinguished from others (i.e streaming because of their length). Packet loss had to be low otherwise the patterns didnt look right and you had to rely on statistical analysis of fixed script set up between two clients locally. Once enough data had been collected then the two endpoints could be moved to a campus network and the test rerun. Because of TCP windowing, latency, packet loss and re transmission more statistical analysis would increase the probability of recognising the skype traffic.....

So dependant on each setup if you collect enough data showing the timings between packets and packet sizes of know fixed scrips and then replay that once you have enough data - you can recognise the traffic type.... OK we can all sleep at night again now. I dont doubt this person can show this to be true in his very finite setup with enough statistical analysis done to show that the same scripted conversation in histogram form can be picked out when repeated over a campus. (Reminded me of chromotography... odd).

For the second PDF the abstract states about hiding content within existing TCP/UDP streams - stenography. What has this got to do with the thread? In this paper he is demonstrating he can hide an encrypted message within an existing stream which a user with the key can decrypt and be invisble to others

Happy to discuss the paper
 
Associate
Joined
12 May 2012
Posts
2,135
Today's the day we get our "Approved visits" ;p

Daily Mail, but.....


After watched the clip, keep it open, there's a second one.

"Ruthless and underhand tactics used by BBC licence fee agents can be exposed today.
Under an aggressive incentive scheme, hundreds of enforcement officers have orders to each catch 28 evaders a week.
Bosses promise bonuses of up to £15,000 a year, saying staff must gather evidence to take as many people to court as possible.
Homeowners who fail to pay can be fined and given criminal records.
Among the vulnerable targeted in the past seven days are a war veteran with dementia and a desperate young mother in a women’s refuge.
The revelations come from an investigation by a Daily Mail undercover reporter interviewed for an enforcement job by Capita.
The outsourcing firm is paid £58million a year to collect licence fees for the BBC, bringing in £3.74billion a year.
The reporter was told by bosses: ‘We will drive you as hard as we can to get as much as we can out of you because we’re greedy.’ He was encouraged to spy on homes and take money on the doorstep.
‘Cash, debit, credit card, we’ll take anything,’ one TV Licensing manager said. ‘I tell people I’ll take shirt buttons.’
Last night, the BBC ordered an urgent investigation into the Mail’s findings, insisting there would be ‘swift and appropriate action’." Riiiiiiiiiiiiight...

At one point in the clip, he says things said at the persons property is a guaranteed conviction. No, it isn't. Only the courts can decide on that.

He said that those without a tv, get a gentler visit. Mkay, all those dozens of letters threatening me (& others) of court action, harassing for money for something I don't (& have already proven) need. I guess I'm just imagining all of that.

"Failure to pay can result in a criminal conviction and a £1,000 penalty. Thirty-eight people – mostly women – were jailed last year for not paying the fine."

Jail time, for this!?

"The firm’s 330 field officers are told they must hit a target – the ‘magic 28’ – the number of evaders an officer must catch a week, signing them up to pay."

So around £70million in payments expected to be caught, at least. That's if they can actually get that many.

"Capita collects taxes and assesses benefit claimants for the Government. It is also entrusted by the BBC – and paid £59million a year – to collect the £145.50 licence fee."

Doesn't seem like it's worth it at that rate.

"Last year alone, they turned up at three million properties – and caught 298,000 evaders."

Over 2.7 million innocent people harassed via letters/visits...

"The salary is £20,000 a year, but, as Mr Doyle explains, ‘You can earn another thousand, fifteen hundred pounds a month." £38k a year? No wonder they're so pushy.
"Mr Doyle brags: ‘One of my officers – my highest performing officer – generally comes in around about 38 to 42 sales a week. They’re obviously after more money. My guy really goes for it.’"

More ridiculous:
CHURCH HOUNDED FOR PAYMENTS
A church in Nottinghamshire has been hounded for payments by TV Licensing. East Leake Methodist Church has received about four letters per year, threatening fines of up to £1,000 for fee evasion.
Each time, treasurer Roger Latham has replied saying there are no TVs in the building.
Mr Latham, 66, wrote in one letter: ‘Only God lives here and he knows everything, so doesn’t need to watch the television.’
Mr Latham, former chief executive of Nottinghamshire County Council, said the letters began arriving in 2009. He said: ‘I sent all the letters back saying: “This is a church”.
They only stopped when Mr Latham told TV Licensing he feared the church’s address was being used by others trying to evade the fee.
He added: ‘People who don’t need the licence are treated like those who are being fraudulent.
A spokesman for TV Licensing said: ‘We updated our database last May to ensure mailings to the Church stopped.’

P.S. https://www.reddit.com/r/ukpolitics..._tv_licence_bullies_are_exposed_how_ruthless/
 
Last edited:
Caporegime
Joined
20 Oct 2004
Posts
26,508
Location
....
Can i ask why? you seem to have forgotten your argument to back up your statement. (Not trying to cause an argument, I'm just interested to know why they can't query a database.)

To get any sort of MAC table they'd need at least to connect to your route, they'd need some sort of warrent (i'd imagine) to do this illegally. Anyway, the BBC isn't doing this.
 
Soldato
Joined
15 Feb 2012
Posts
3,290
Location
2
For the second PDF the abstract states about hiding content within existing TCP/UDP streams - stenography. What has this got to do with the thread? In this paper he is demonstrating he can hide an encrypted message within an existing stream which a user with the key can decrypt and be invisble to others

Happy to discuss the paper

Jim, I'm not trying to prove anything to anybody. Earlier in this thread I didn't think it was possible for TVL/BBC/whoever to see if a wifi user is using BBC iPlayer content without illegally entering the network and/or breaking the WPA2 encryption. I went away and researched it and now believe it is possible through steganography. That paper was one of a few that convinced me to change my position. As far as I can tell you don't believe it is possible, even after reading that paper. We disagree; we have different opinions. It is allowed, in fact I would say it is healthy. I respect your opinion (after all you have a wealth of experience in this field) and, contrary to normal GD behaviour, I do not seek to defend my position or ridicule yours... it's science. Plus it's an interesting discussion. :)

I die a little in the GD threads where people adopt a position and refuse to shift (or in fact entertain other positions at all) while simultaneously trying to discredit and/or convert the (non-)believers. :(
 
Back
Top Bottom