pull down to refresh
OK, so have you published your findings and your recommendations? Also, what was your point in running Knots and signalling for BIP-110 with your sybil nodes when you claim it was just for research purposes? Clearly you made that change at some point in the middle of your "research".
Also, is this you: https://www.bitprojects.io/
If so, why has it been offline since at least when you shutdown your 3000+ nodes?
Article coming soon.
To summarize though: I've been running these nodes long before the spam war started. I switched to knots last year because I believe that's the right choice, 100k op_return dramatically increases the attack surface of bitcoin and must be stopped. I then upgraded to knots 29.3+bip110 about 2 weeks ago now, and that's when everyone noticed all the nodes.
These nodes may be viewed as "fake" but they were still handling 80k connections from 35k unique sources. I actually did control 3k nodes, and I chose to signal bip110.
I think even the puritans are irritated by your behavior in that regard. Hurts their numbers and accomplished nothing but reinforcing the narrative that what little support there is for the RDTS fork was artificial.
Looking forward to your published findings based on the experiment.
The vulnerability is that I was able to do this for a much lower cost than assumed would be required. IP addresses are easy to rent/lease for $0.30-0.40, ASNs are easy to obtain and advertise from, and the bitcoin blockchain is easily deduplicated at both the block and file levels.
Only a small fraction of nodes are IPv4 reachable (meaning, accept inbound connections via IPv4), and nodes by default only make 8-12 outbound connections. With a little bit more scale, a single person or entity could become a large majority of the IPv4 reachable nodes that other nodes connect to, and that would allow them to control/restrict/decide/manipulate all traffic flow between nodes.
One action that would help change this is if more people made their bitcoin nodes IPv4 reachable -- able to accept inbound connections on TCP/8333 from the public internet
Another action would be to increase the outbound connection count to 24-48.
There are other combined actions/changes that help solve this but these two are ones that individual node runners can do.