Maths Problems

Open discussion regarding technological or telecommunication issues
datax1969
Posts: 21
Joined: Sat May 02, 2009 8:53 am
Location: Sydney

Re: Maths Problems

Post by datax1969 » Mon May 04, 2009 11:01 pm

CoreyPlover wrote:You should not blame others for "encouraging" something when they are simply providing the service.
Alright, your point. "encouraging" was a poor choice of word. "tacitly allowing and accepting" maybe? Not quite what I'm after because I also wanted to make the point that it is used as a competitive point of difference between ISPs so it is a little more than "tacit acceptance", even if it is a little less than "encouraging". Anyway, you know what I'm driving at.
CoreyPlover wrote:It is not "impossible", as two forum respondants have already indicated. I'm sure you meant "unlikely"
Yeah, yeah. Nothing is impossible. Physics can even tell you the probability of all the air in my room right now suddenly ending up in one corner leaving me asphyxiating. I'm not saying the probability of using 60GB off peak legally is quite as small as that, but, as you say, you get the drift.
CoreyPlover wrote:I think there are three main drivers of copyright infringement in society today:

* monetary gains (not having to pay for movies, DVDs or CDs)
* perceived anonymity (users mistakenly think that they cannot be traced, when in fact it is quite trivial to track down infringing users).
* perceived immunity (the argument that "they never target me because I'm too little in the grand scheme of piracy")

You can be assured that over time, these last two drivers will disappear as the processes become more and more automated (not just in Exetel but worldwide)
Completely agree with everything there. "Perceived anonymity" and "perceived immunity" essentially combine to form my "numbers" argument - they'll never get me because even though they can trace me and/or prosecute me, why would they pick me rather than the gazillion others. As you say, this is changing as it becomes easier and easier for automated processes to "make the bust" and maybe law changes to allow the bust to be made.

But, as we have seen from the evolution of Napster->Kazaa->BT->eMule the pirates will probably stay one step ahead in the game. Or, if they fall behind, it won't be for long. Look at all the failed attempts that have fallen by the roadside:

Apple II games using data on half tracks - cracked quickly by Locksmith.
DVD encryption - a month or two to crack.
DVD Region Encoding - I still laugh at the thought of how low the managerial IQ levels must have been to think of that gem.
SecuRom, SafeDisk et al - all useless within a month or two. Safedisk 4 took a little longer.
Almost every gaming console - mod chipped within a year. (Congrats to PS3 though, 2 years in and still not broken)
BluRay encryption - cracked.
DRM - Don't make me laugh. Even though I think its still uncracked the response from users defeated it without the need for the hackers.

So its only a matter of time before a better P2P protocol is developed with no need for any tracker maintaining the swarm's IP addresses. I see a few problems in establishing that the guy at the other end is a genuine thief and not a copyright cop but a bit of imagination will solve that problem.
peterh_oz wrote:Everything I quoted is a "download now, watch later" product, other than internet radio.

Most "internet tv" is downloadable. EVERY show on revision3.com is downloadable.

Ever heard of video podcasts?

I download LEGALLY Good News Week every week from channel 10.

I also download some shows (video and audio) from the ABC - LEGALLY.
Well I can't make those numbers add up to very much, 1 hour episode of TV is less than a GB but alright maybe you do use close to 60GB of legal content. I'm impressed that you have the time to absorb all of that. I'd never heard of revision3.com so I checked it out. Don't take this as insulting, I've got some weird tastes myself, but I don't think anyone would classify this as mainstream content that everybody is dying to see. You are obviously unusual in your tastes and in your morals. You have never been even the slightest bit tempted? All you have to do is press that button and get that really cool game/movie that you've had your eye on? What about that $100 bill on the pavement, did you stop to pick that up? If so, you probably said "Does this belong to anybody?" And then some bastard like me lied through their teeth and said "Yeah that's mine, ta". Anyway, good for you - you have dented my faith in the lowness of human nature.
peterh_oz wrote: So, using your analogy, EVERY CAR on the road should be speed limited to 100kph. No-one *should* go over the limit, so anyone who builds a car which CAN, is aiding people in breaking the law.
That's not at all what I said, implied or is even remotely related to the thrust of my argument. The point was that in traffic I am protected by numbers, there are too many people to bust, the same as with pirating.

By the way, have you seen the ads from the car manufacturers? Car going really fast on some beautiful country road somewhere? A Skyline GTR has 300+kW of power, goes 0-100 in a femtofart. Funnily enough I've never seen one towing a van.

In fact, I'd say it was a perfect analogy, the ISPs are advertising in the same way the car manufacturers do - reading between the lines of those car ads:

"this car can really hump. But we're not saying that you take it to some nice quiet out of the way place to see what she can do. No. We're just showing you what it could do if you were of that mind"

And between the lines of the ISP ads:
"this plan goes really fast and you can download heaps. Now we're not suggesting that you use it for illegal purposes. But if you did, gee you'd get heaps of stuff"
peterh_oz wrote: And I'm not sure how speeding in Sydney's blocked traffic jams helps, but if you think it does, enjoy your licence cos you will get caught one day,

And I hope you're not the type who write to protest "I've got a perfect driving record, please let me off". By your own admittance, you have a perfect driving record because you've been lucky, not lawful.
Now where did that come from? I don't have a perfect driving record. I get 'done', on average, once every two years for a 3 pointer, just like most other people I talk to. I've been particularly lucky for the last 5 years. I've probably got a 6 pointer waiting for me around the next corner, I'm getting paranoid now - I'm "due".

Obviously I'm not talking about Sydney's permanent parking lots. I'm talking about roads at peak hour that still sort of move - Mona Vale Rd through St Ives, Penshurst St and Willoughby Rd through Chatswood West if you want particular examples that I drive or have driven every morning, there's plenty of others, they're not all as perenially bad as the M5.

My point was that everywhere I drive in Sydney I HAVE to drive faster than the speed limit - I'm a nuisance otherwise, because I'm out of step with the rest. I'll never get caught in the daily grind because I've never even seen a cop on the road at that time, they know to stay away because their very presence causes traffic chaos with everyone slamming on the anchors at the first sign of a flashing light (when I do get 'done' its late at night or some other odd time when I'm not concentrating as hard as I should be).

Again, the analogy to pirating, protection by numbers. Now, as Corey rightly said that is changing with automation. Ready to stretch the analogy even further :D . How about this then? Automation in traffic control with speed cameras; slowed things down for a bit, now we all know where they are, we hit the brakes just before the camera, and take off again like startled rabbits a few metres later. Same deal with the pirates, it'll slow down for a bit but the "next big thing" won't be far away.

Dazzled
Volunteer Site Admin
Posts: 6003
Joined: Mon Nov 13, 2006 1:16 pm
Location: Sydney

Re: Maths Problems

Post by Dazzled » Mon May 04, 2009 11:19 pm

A great deal of newer DRM in fact has been broken. I have even watched some dreary rubbish just because some clown had tried to stop me, and I have little interest in copyrighted media. Thorne Smith in the 30s once built a novel around the notion that the invention of the chastity belt was responsible for the problem of burglary - still an amusing read.

datax1969
Posts: 21
Joined: Sat May 02, 2009 8:53 am
Location: Sydney

Re: Maths Problems

Post by datax1969 » Mon May 04, 2009 11:23 pm

JasonM wrote: Approximately 350 notices may have been processed in April - about 10 a day for April.
This compares to October 2008 - about 20 a day.
Thanks for that. I really didn't think I'd see any numbers at all with my question, it was really more tongue in cheek.

With those numbers you've restored my faith in human nature that was so severely dented by peterh_oz. I know that as a percentage of users it looks like a very small number but I think that's about what I'd expect based on my own notice:downloads ratio. So it probably means you have a reasonable amount of pirating going on because:

a) Any savvy user is using ipfilter.dat with BLM or equivalent. Certainly cuts out a lot of chances to get sprung.
b) Its hard to tell but I don't think that many torrents are actually policed. I've been caught on the latest and greatest stuff, but I really don't think they're taking the trouble to catch people on old TV episodes.
c) Not everybody is downloading all the time. You get that season of House and then spend the next month watching it.
JasonM wrote: Based on bandwidth growth, I say the users are either changing to paid or free content, or getting smarter.
:) No prizes for guessing which of those options I'm betting on.

CoreyPlover
Volunteer Site Admin
Posts: 5922
Joined: Sat Nov 04, 2006 2:24 pm
Location: Melbourne, VIC

Re: Maths Problems

Post by CoreyPlover » Tue May 05, 2009 12:42 am

Extracting some points from your post that make for some good continued discussion:
datax1969 wrote:But, as we have seen from the evolution of Napster->Kazaa->BT->eMule the pirates will probably stay one step ahead in the game.
...
Look at all the failed attempts that have fallen by the roadside:<snip>
...
So its only a matter of time before a better P2P protocol is developed with no need for any tracker maintaining the swarm's IP addresses. I see a few problems in establishing that the guy at the other end is a genuine thief and not a copyright cop but a bit of imagination will solve that problem.
There has always been a continuing struggle between pirates and commercial authorities...and I agree that DRM will always get cracked; it is a losing battle trying to lock down distributed content.

But, copyright infringement is not about DRM. It is about whether the piraters can continue to create an anonymous means of distributing content without detection. And this represents more than just a "few problems". Trackerless torrents, IPFILTER.dat, etc are trivially circumvented by authorities.

I can only think of three ways in which copyright infringers can continue unabated:
  • If the efforts of tracking an infringer (cost) outweigh the benefits of catching them (fines)
  • Mass obfuscation like onion routing
  • Minimising exposure time to detection (as you say, download fast, then remove yourself from the network)
The first is a matter of technology and I think we are already at the point where the costs of detecting infringers is low enough to warrant mass detection (as you have already discovered). The second is a completely inelegant mechanism as it increases n-fold the amount of bandwidth and traffic needed to transmit information and you can rest assured that packet inspection at the ISP level will circumvent it quite easily (there is no way they will sit idly by and watch bandwidth spike n-fold and not implement countermeasures). The third is my favourite. I hope that EVERYBODY who wants to leech stuff does this, because it will lead to the demise of P2P networks (P2P can only survive if there are high enough proportion of uploading peers).

datax1969
Posts: 21
Joined: Sat May 02, 2009 8:53 am
Location: Sydney

Re: Maths Problems

Post by datax1969 » Tue May 05, 2009 10:52 pm

CoreyPlover wrote: I can only think of three ways in which copyright infringers can continue unabated:
  • If the efforts of tracking an infringer (cost) outweigh the benefits of catching them (fines)
  • Mass obfuscation like onion routing
  • Minimising exposure time to detection (as you say, download fast, then remove yourself from the network)
Exactly. I'll leave it to those smarter than me to come up with a solution to these problems. If I had to bet on what the solution it would either be:

1) Some sort of authentication mechanism similar to the way the eBay rating system works. The tracker is still necessary but rather than keep track of who has what files, it keeps track of everyone's trust rating. This mechanism wouldn't have to be secure against authorities, publishing your IP on some sort of list is not enough evidence, they actually have to connect to you and exchange data, if they're not trusted you don't let them do that.

I see a host of problems with this so there's no need to make any arguments against it, I've thought of them already. But in some form or other it may be possible. The point being to make it hard for automated bots to get a good rating - method 1 in your list. Maybe actual user interaction is needed to establish a network of trusted thieves - I trust A, and he trusts B, therefore I trust B. It wouldn't stop the authorities from trying to get a good rating, but it might make it hard enough for them not to bother.

2) I'm not too shabby when it comes to the IP protocol but my practical knowledge is limited to smallish LANs of 10s of computers. I know, in black box theory, how IP is used in a WAN but exactly what happens in ISP land is a bit of a mystery so this idea could be dead before I even start. Still here's the rough idea:

I'm thinking that Method 2 on your list can be accomplished without too much extra traffic by using UDP multicasting (IGMP). There's no TCP connection and nobody needs to talk directly to you, so you can spoof your IP address in any IGMP packet that you send. The various swarms each take over an unused multicast address and everyone just joins that IGMP group and listens and transmits to that. Rather than transmit a requested block of the file, as you currently do with BT, you just cycle through all the blocks you have on an regular basis. Obviously, there can be no monitoring of uploads:downloads so you have to rely on "honour amongst thieves" to prevent too many leeches. I don't think this would increase total traffic too much. Yes, one multicast transmit packet eventually turns into n packets if there's n members of the group but when I'm connected to a 100 peers/leeches via TCP that's also a lot of replicated data packets - no ACKs or any other TCP overhead needed either.

Big problem #1 - Do ISPs even allow little users like me to send an IGMP packet, or rather does it get routed like its supposed to or does it die at the first switch? Would the ISP care if the Sender IP address is spoofed?

Big Problem #2 - There's not enough unreserved multicast addresses. These could be multiplexed in some way - not using port numbers though, because that WOULD drastically increase the volume of traffic. So some sort of time division multiplex system managed by a tracker could be done. A tracker in this situation doesn't hurt, no one knows anyone's IP address.

Big Problem #3 - I'm really just blowing it out of my posterior. I'm a programmer with a enough of a clue about networks to keep me going. So I could well be spouting complete BS.

I know nothing about IPv6, maybe that has a whole bunch of multicast features that could be (ab)used by the pirates and that could be the solution, when (if) it happens.

Hmm, just Wiki'd IPv6. From what I understand bring it on! Pirate heaven in multicast land there. Assuming the ISPs play along, the very tired argument - "you can't ban such-and-such a protocol because it has legitimate users" - might come in handy, there's got to be a good P2P protocol in the making there.


Anyway, if P2P ends up getting clobbered we'll just go back to the bad old days. Setting up a network of friends, we all buy one game/movie, then copy them and share them around. I'll miss the easy access to TV episodes, I might actually have to yeeech, actually pay to rent the season DVDs at Video Sleezy and then copy them.

It could be worse than that if P2P loses out. The people I consider the real pirates, ie. those that onsell and make money from their ill-gotten booty (even, I'm not low enough to think about doing that, plus the chances of getting caught skyrocket) will see a windfall. All the Usenet operators in non-signatory countries, all the Singaporean street traders etc etc, might end up making even more squillions. That would be a real shame.

I know one thing for certain - you could stop P2P tomorrow and any increase in profit by anyone would hardly be noticeable. The pirates will either find another way - a friend network, paying the real pirates or just living without. Because one thing I reckon we all have in common is that we're tighter than a nun's you know what.

CoreyPlover
Volunteer Site Admin
Posts: 5922
Joined: Sat Nov 04, 2006 2:24 pm
Location: Melbourne, VIC

Re: Maths Problems

Post by CoreyPlover » Tue May 05, 2009 11:37 pm

datax1969 wrote:I'm thinking that Method 2 on your list can be accomplished without too much extra traffic by using UDP multicasting
Interesting concept. From very brief research, I can see development is being done into multicast P2P networks (mainly for streaming radio and video servers) because they are more efficient compared to unicast networks and I think this development is great in light of increasing bandwidth demands.

The unacknowledged nature of multicast makes it more prone to packet loss, but I'm sure this can be overcome in a myriad of ways. One-way IP multicast is very useful for efficient content delivery, and has been implemented in internal networks, but this is centralised and would effectively be a warez server broadcasting pirated content (quickly shut down by authorities).

One pitfall with multicasting is that it doesn't scale very well. Essentially, you need to build the multiple IP destinations into a datagram. That, or you need to still have a tracker that manages the routing tree. Either of these options is not anonymous and again, easy to monitor participating seeds. You *could* do it anonymously in a number of ways (broadcast to all IP addresses within a subnet, or via Onion routing) but these become subject to Denial of Service attacks (http://en.wikipedia.org/wiki/Anonymous_ ... s_networks) so you can distribute material anonymously but not reliably.

The trusted network proposal you gave is known as a Friend-to-Friend network (F2F as opposed to P2P) or Darknets. Yes, more difficult for authorities to infiltrate, but it naturally hinders the spread of material, so some might consider it to be a partial win against piracy.

datax1969
Posts: 21
Joined: Sat May 02, 2009 8:53 am
Location: Sydney

Re: Maths Problems

Post by datax1969 » Thu May 07, 2009 7:55 am

Looks like we've bored everyone else to tears, so I'll keep this brief.
CoreyPlover wrote: The unacknowledged nature of multicast makes it more prone to packet loss,
Trivially solved using modern checksumming and hashing. Any retransmissions required don't actually add anything to the total bandwidth, a seeder can't be told what to send, so receivers just get what they get. It just means the download takes longer. Plus you hardly ever see packet loss in modern networks.
CoreyPlover wrote: One pitfall with multicasting is that it doesn't scale very well. Essentially, you need to build the multiple IP destinations into a datagram. That, or you need to still have a tracker that manages the routing tree.
Not if you're using IGMP. If all the routers and switches in the world are playing the game then multicast should "just work". A receiving host joins a multicast group (an IP address with some particular bits set - 224.168.x.y is an example) by sending a join request to the network using the IGMP manager address (224.168.0.0 from memory), an IGMP manager (always the switch/router/gateway) picks this up and now knows that host X, plugged into physical hole Y, is a member of group Z. All multicast packets hitting the switch with a destination == group Z get forwarded down hole Y. The protocol expands in a well-defined way to let multiple switches on different networks let each other know the set of groups of interest on their local networks.

The problem I referred to earlier is that I don't think many switches and routers are playing the correct IGMP game. And even if they are, I can't be bothered doing the maths, but I think the number of free IGMP multicast addresses is not large enough to manage all the individual swarms.

* If you're still using an old style bus Ethernet architecture rather than each host directly connected to a switch then there's a defined way of mapping IGMP IP addresses to Ethernet multicast MAC addresses, in this case, the gateway host will Ethernet multicast on its local subnet rather than pump it out a particular hole as a switch would.


That sees this topic done to death I reckon.

Post Reply