1

Correct me if I am wrong, but if a malicious user keeps creating smart contracts and deploying them onto Ethereum, they would grow the size of Ethereum's ledger. Eventually, people may run out of hard disk space to store the ledger. As people run out of disk space, they drop from the network and the strength of the network diminishes, possibly leading to the system's collapse, thus it is in every user's best interest to keep ETH price point relatively high in order to make deploying a smart contract fairly costly. Is that correct?

Is there a cost associated with the size of the contract you are trying to deploy?

1 Answer 1

4

Check out this reddit comment: https://www.reddit.com/r/ethereum/comments/56f6we/explaining_eip_150/.

There was such an attack back in October 2016 because the cost of creating a particular type of contract was too small. Generally, though this type of problem is taken care of with 'gas.' 'Gas' causes every transaction, including spam, to cost money. The cost is, presumably, more expensive than it's worth.

There's also a per-block gas limit that precludes this type of spam.

4
  • I just worry that someone with a lot of liquid might convert a large chunk of their money to ETH, especially if they're upset that Ethereum has sufficiently disrupted some of their businesses, and they try to kill the system by deploying a lot of contracts at once, since presumably gas is much cheaper than ETH so you can probably do some level of damage with that already...
    – Alexandru
    Aug 4, 2017 at 15:41
  • 1
    If someone tries to spam, there are more transactions, therefore there is more gas to be 'won,' therefore, it becomes profitable to be a miner, therefor more miners enter, therefore hashing power goes up, therefore difficulty goes down, therefore block times stay the same. The system protects itself. At least, that's the theory. Aug 4, 2017 at 15:47
  • 1
    With the per-block gas limit of ~7 million gas and an average gas price of ~0.00000003 ether, it would cost around 0.2 ether to saturate a block from an attacker. At an 18 second block time, that's roughly $125 per minute, which, I suppose, isn't all that costly for an attacker fill the entire network with spam. OTOH, $125 buys a lot of storage capacity nowadays. So mining even one extra block at full saturation (and without elevated gas prices), a miner could easily justify buying a hard disk or three for his/her nodes. It would cost a little bit more for "regular" nodes to run.
    – lungj
    Aug 4, 2017 at 15:51
  • I did say "That's the theory." I always thought someone with enough motivation could make it happen. Who knows. Aug 4, 2017 at 17:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.