[fsflover]

I prefer torrents inside I2P [0], because otherwise my IP will be shown to everyone. I wonder why there are so few people there. It's of course slow, but it's worth it. There are already quite a few torrent trackers.

[0] https://geti2p.net

[knaik94]

What kind of speeds do you get? Because last time I tried it getting above a few megabits per second was a struggle. When we're talking about even an Ubuntu ISO, it's a couple gigs. It's too slow to want me to keep using it.

[fsflover]

Indeed, few Mbps is typical for the current number of users. It can take days to download relatively large files. Still, is speed more important than anonymity? Definitely not always, perhaps never.

I hope it gets better with more users, as it happened with Tor.

[Eli_P]

> I wonder why there are so few people

Maybe because people understand that it's security model is flawed: the peers in the mesh treat all the other peers as not trusted, that's okay; but the network design doesn't seem to take into account potential hostility of the physical medium itself.

Say, your ISP can tarpit packets or shape traffic, or shut down the power or cell tower in your block temporarily, and then measure how the mesh congestion changed. At the same time the mesh relies on building hop chains with TTL of 10 minutes iirc, which I think makes a peer a sitting duck. It is (was) all documented/leaked.

[fsflover]

AFAIK garlic routing is still harder to crack than Tor: https://geti2p.net/en/comparison/tor. The main advantage of Tor is the number of users.

[saurik]

The network security people I know stopped bothering publishing attacks on I2P's directory service--which was(/is) in fact subject to eclipse attacks--because I2P didn't seem to care to fix the issues. Here's a great paper.

https://sites.cs.ucsb.edu/~chris/research/doc/raid13_i2p.pdf

[fsflover]

>The network security people I know stopped bothering publishing attacks on I2P's directory service

Maybe those you know stopped, but others didn't: https://geti2p.net/cs/papers/.

[saurik]

Great... are the issues getting fixed? The point I was making about I2P working so poorly that some people have given up trying to help--in addition to specifically noting that eclipse attacks do, in fact, affect the design of I2P--doesn't seem affected by "there exists people who still spend time demonstrating I2P doesn't work well" ;P.

[Eli_P]

I doubt the number of nodes is important when you can split a large group of them into a few subgroups and analyze them or alter how they interact with each other, then repeat subdivision; because again, them overlay networks are like smaller sandboxes inside the big one.

[fsflover]

Don’t forget that you have to do all this within 10 minutes, because after that time the network configuration will be totally different.

[Eli_P]

Yes, overlay network's configuration will be somewhat different, but physically it will be the same. It's like you're rearranging apple bits in a pie to save them; the pie will be sliced and eaten anyways.

[fsflover]

> the pie will be sliced and eaten anyways.

The difference is how long it takes and how much resources. There is no perfect security. Security is not a "yes/no" thing.