Record labels and movie studios are willing to pay serious cash to protect their content from being shared on BitTorrent and other file-sharing networks. They have paid millions of dollars to anti-piracy outfits such as MediaDefender who in return promise to do all they can to distribute fake and polluted downloads.
According to a recently published paper by Prithula Dhungel, Di Wu and Keith Ross, these effort are a waste of time and money. In the paper titled “Measurement and mitigation of BitTorrent leecher attacks,” the researchers show that BitTorrent swarms are hardly influenced by attacks from anti-piracy outfits.
The research looked into the effectiveness of two popular attack methods used by companies such as MediaDefender. The first is a ‘piece attack’ where the hostile leecher attempts to slow down downloads by creating as many hash fails as possible. The second method is the ‘connection attack’ where the hostile leechers try to tie up as many TCP connections as possible in order to make it impossible for downloaders to connect to real peers.
The different methods were tested in a real-life BitTorrent swarm of a popular music album that was targeted by these attacks. “We present measurement results for a torrent for a new album, which was verified to be under attack,” the researchers report, adding “This popular album was released a few weeks before our experiments. At the time of the experiment, it held the number 1 position on the UK album chart and iTunes ranking list.”
The researchers then downloaded the ‘attacked’ torrent several times with both Azureus (Vuze) and uTorrent. For each download they recorded the time it took to complete, both with and without using blocklist software that bans (some) of the attackers’ IP-addresses.
The results were quite remarkable. The researchers found that, on average, downloads with a blocklist were 30 to 35% faster. In other words, the efforts of the anti-piracy outfits do slow down the targeted swarms, but only for a few minutes at most, and not long enough to deter anyone from downloading.
A more detailed look at the peer distribution of the two BitTorrent clients further reveals that without the IP-filters, uTorrent encounters only 2% of malicious peers, who all use the ‘piece attack’ method. Azureus on the other hand encountered no ‘piece attack’ peers at all, but 18% ‘connection attack’ peers.
Not surprisingly, the researchers conclude from their research that the methods used to attack BitTorrent swarms are highly ineffective. “The anti-P2P companies are not currently successful at stopping the distribution of targeted assets over BitTorrent. We have also found that blacklist-based IP filtering is insufficient to filter out all the attackers,” the researchers write.
What the researchers have overlooked is that both Azureus and uTorrent have implemented various technological measures against these automated attacks. The results may differ for other BitTorrent clients. Azureus (now Vuze) has put a lot of work in preventing ‘piece attacks’ and uTorrent has implemented similar anti-pollution measures.
The overall conclusion put forward in the article is most likely the right one, and to most people not even that surprising. The millions of dollars spent by the entertainment industry to protect their works from being shared on BitTorrent is at best only a mild annoyance to the ‘pirates’.