mcaspi Posted October 20, 2010 Report Posted October 20, 2010 Show the user somewhere (Statistics? Status bar?) how many clients are banning him/her. This can tell the user if he/she has any problem causing him/her to be banned.
DreadWingKnight Posted October 20, 2010 Report Posted October 20, 2010 Get a reliable method to detect being banned first. No there isn't one.
mcaspi Posted October 20, 2010 Author Report Posted October 20, 2010 Why not?. The banning client sends a +1 message to the banned client. The banned client displays the total bans.Let me tell you why I suggested this. I am testing my uTP connections which some of them don't transfer well for some reason. At one point I noticed that some peers are missing. My guess was that uTP errors caused me to be banned by those peers. So I changed my ip and and boom, saw the disappearing peers again. Information like this can also give BitTorrent information on how their implementation is working.Buggy implementation should produce many bans.
DreadWingKnight Posted October 20, 2010 Report Posted October 20, 2010 Except that no such message exists in the protocol, and if it did, things would be EXTREMELY strongly abused by anti-p2p groups once it gets to the wild.
mcaspi Posted October 20, 2010 Author Report Posted October 20, 2010 Regarding the protocol, it needs some thinking. It might be implemented like PEX. Peer A sends information to peer B. Not sure. Again, it needs some thinking about if and how. One thing is almost sure. No one, including BitTorrent Inc., knows today, from how many clients, each client is banned from and why. I am also almost sure, that I was banned from those peers because of sending "bad data". "Sending" data at a max speed of 0.2 kB/s can't produce good results. Such an indication will give both the users and the developers, a better understanding of what is going on in the swarm. Just think of how much technical information you can get, if users come to the forum and complain that they are banned for some reason and don't know why. In my opinion it's a great diagnostic and informative tool.
Switeck Posted October 21, 2010 Report Posted October 21, 2010 ISPs will love this...and will be spoofing that within days of it being in a stable build of uTorrent.
mcaspi Posted October 21, 2010 Author Report Posted October 21, 2010 Switeck and Knight, it is not implemented yet and if it will be implemented, you don't know how. So at this point, I can't understand how you guys can know that ISP's and anti-p2p groups can abuse it. I gave an idea and BitTorrent can adopt it, some of it or not adopt it at all. I don't care. Just gave an idea. This is my last post regarding this.
DreadWingKnight Posted October 21, 2010 Report Posted October 21, 2010 So at this point, I can't understand how you guys can know that ISP's and anti-p2p groups can abuse it.It's called modus operandi.These organizations have a HISTORY of attempting to use functions like this found in clients to ruin the experience for users.You can't understand how people who have been dealing with torrenting for 5+ years can know how interest groups who have also been working for 5+ years against torrenting will behave?
Ultima Posted October 22, 2010 Report Posted October 22, 2010 it is not implemented yet and if it will be implemented, you don't know how.The implementation is irrelevant. The solution space for your problem is such that the only way to solve it is to get someone else to tell you what you want to know, so it's open for easy spoofing. Knowing this number isn't terribly interesting, and your given use cases aren't all that compelling. Bans only happen when bad data is sent, don't happen all the time, and don't have to have anything to do with uTP to still happen. Given the first point, do you seriously think uTP doesn't even have the most basic functionality (data transfer) down?If abused by 3rd parties, this becomes absolutely worthless. At least if PEX messages were spoofed, the worst that would happen is a failed connection -- nothing really lost except a tiny bit of time and bandwidth. This one would result in a misleading number displayed in the UI, which you have no means of validating. Yes, tons of other numbers shown in the UI can be inaccurate, but none of them are features such that the number is the feature, so being spoofable isn't feature-breaking for them.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.