On: Malicious attack on Wikipedia – what we know and w...

[mpfundstein]

I love ipfs. Can this actually be ddos’d as well?

[progval → mpfundstein]

Yes. It's based on a bittorrent-like protocol with a DHT, so you can get a list of nodes hosting a particular file; then you can DoS them.

[smitop → mpfundstein]

The ipfs.io is just a web-based way to access IPFS, called a gateway. There are a bunch of different gateways. In addition, you can run an IPFS node locally, and then as long as just one node holds the content you're looking for you are looking for you're good. There are also browser extensions to re-write gateway URIs to localhost URIs.

[alexeldeib → smitop]

Does this actually answer the question? If there's a node online that means I can reach the content, but would it help with DDOS? Not so sure.

[whitexn--g28h → alexeldeib]

A popular IPFS file might be available on thousands of nodes similar to how popular torrents have thousands of seeders. DDoS attack against thousands of servers across multiple countries and networks would be nearly impossible to perform.

[viraptor → whitexn--g28h]

It depends on the kind of node storing the data and how many there are. It's likely easier to DDoS 100 people on DSL than a single Wikimedia endpoint.

[nine_k → alexeldeib]

Ipfs rarely stores 100% of content on one node.

[vasili111 → smitop]

So there is no central place like bittorent tracker which if down the network does not works?

Or is it like DHT which does not need central tracker?

[shakna → vasili111]

IPFS uses DHT on a fairly fundamental level.