Help and Support
Ask a question, report a problem, request a feature...
<<  Back To Forum

IP Filtering Observations/Suggestions

by Guest on 2023/01/16 11:09:59 AM    
I usually seed about 100 or so transfers at once. Many of them have 100+ peers. I use an IP filter from Github and keep it updated. It's not unusual to have half of the peers filtered out for any given transfer, but many other problematic peers remain. I spend hours manually "ignoring" peers that the IP filter has missed.
It would be great if Tixati added the ability to easily filter peers based on IP address, range, country, or some combination thereof. For example, I have to manually filter all IPs beginning with 103.x.x.x that originate from Africa and several southeast Asian countries. There are literally thousands of them, and playing whack-a-mole for hours on end gets old really fast. Other examples are all IPs from Switzerland that begin with 213.55.x.x and most U.S. addresses beginning in 18.x.x.x, 35.x.x.x, 52.x.x.x, and 54.x.x.x.
Ironically, most of the addresses used by my VPN somehow ended up in the ipfilter.dat file. I have to log on and off several times before I can get an address that isn't filtered by other peers.
Does anyone else have this issue?
by notaLamer on 2023/01/17 03:36:23 PM    
1.  you can filter the subnets
2. When you add a new IP filter it tells you that you can filter countries like @US @PL
3. The blocklists found online are filled with trash to 99%. How did you not come to this conclusion when you started blocking your own VPN addresses? These are lists of abusive hosts, Bittorrent has no such concept as abusive hosts. Instead it has the problem of surveillance hosts that are unique to bittorrent and nobody does an automatic detection of them to my knowledge.
Many people must use VPNs and your list blocks VPNs. Counter productive!

If you want to efficiently block bittorrent botnets then start by blocking all AWS IPs. Then you can look up previous posts on these forums to see what blatant surveillance bots look like and block their IPs or /24 subnets.
by shag00 on 2023/11/11 03:02:50 AM    
I emphasise with your dilemma. There are a number of sub issues that you need to consider:
1/ private or public tracker: for this discussion only public trackers are discussed
2/ general country internet infrastructure: every country has a few fast peers but most countries have an average or general speed, eg Africa, South Africa has a few reasonable peers but in general all of Africa is bad news, Middle East, I have yet to find a good peer in the entire area, America, a minority of good peers but mostly bad, Western Europe almost all bad.
3/ country pricing structure: countries that charge on a per MB usage are always bad and heavily influences country infrastructure bias/rating
4/ spy bots: mainly US and Western Europe, reasonably easy to filter out.
5/ consumers & torrents: points 2 and 3 end up giving a huge consumer base of shit peers. They are generally the most populous group, being poor, having poor hardware and cannot afford 15 streaming services. These people go to places like EZTV to download TV episodes that are 500MB in size and upload at less than 5 Kb/s.

What could Tixati do? Be far more aggressive in dropping poor uploading peers/seeders.

What can you do? Be far more selective on where you go to download/upload your files. Only torrent premium quality files, as a rough guide everything should be multi Gb in size because all(most of) the poor people won't download them.

It is practically impossible to IP ban all the bad peers on a site like EZTV as the banned IPs would total in the millions. RARBG would be(was a) a good example of a curated high quality site that discouraged the great unwashed and gave superior throughput.

This web site is powered by Super Simple Server