|
Title: Possibility of a DDoS attack Post by: coder0x15 on July 16, 2019, 04:41:23 PM Theoretically. This is not a mechanism of any kind of attack, but It's about the "lost" coins. May be. It is just for an educational experiment.
What if someone has the ability to send transactions to the each node's memory pool. He have also the ability to send TCP packets from some list of the internet protocol addresses (65533 IPs against 9686 nodes, for example). He send transactions, but R and S in it's signature are not formed by the such way as they can be formed by the real owner of the private key. Each node needs to get unsigned TX hash and then check it's Sig for to confirm the output's amount spending. It will take some time. Is it a theoretically possible DDoS attack to [any]coin network? Each node need to verify signature and it takes some time. Provided that attacker has the full UTXO set locally on his [XXXX] machine IPs. Or IPv6 folds this "problem"? Or what is really means "bitcoin full-node"? Title: Re: Possibility of a DDoS attack Post by: achow101 on July 16, 2019, 04:53:48 PM Node's have rate limiting and will disconnect and refuse connections to and from peers that misbehave (i.e. they ban that node). So if a node sends a bunch of transactions too quickly, they will be disconnected and banned. If they send too many invalid transactions, they will be disconnected and banned. At worst, such an attack would take up a few seconds of CPU time before every node that this attacker has connected to disconnects and bans him.
Also, signature verification isn't all that slow. Nodes already fully validate every transaction they receive which includes signatures. In fact, it's faster to check an invalid signature because it will exit the signature check algorithm early. |