5.4.4. Analysis of Possible attacks

*1. Double spending attacks*: In a double spending scenario, an adversary attempts to revert a transaction that has been finalized on the distributed ledger. In Econledger, a voting-based chain finality mechanism ensures the total order and persistence of data recorded on the distributed ledger. Thus, once a transaction is finalized in the checkpoint block, all other honest validators will work on the finalized main chain and disregard any double spending transactions from attackers.

*2. Free-riding attacks*: There is a possibility of free-riding attacks that some lazy nodes only gain benefits by using the security service without fulfilling their responsibilities in the EconLedger network, such as forwarding messages or submitting ENF proofs. The punishment strategies can prevent against free-riding attacks by reducing credit stake of dishonest nodes or even isolating them from the entire network.

*3. Selfish-mining attacks*: In a selfish-mining attack, the adversary tries to withhold blocks and release them strategically to reduce chain growth and increase the relative ratio of his proposed blocks. In PoENF consensus, only valid blocks generated in the current round can be accepted by honest validators, while those outdated blocks are discarded. Moreover, withholding blocks is a type of misbehavior in PoENF, and it decreases both profits and credit of a dishonest node. Therefore, selfish-mining is unprofitable for rational validators according to reward and punishment strategies.

*4. ENF-proof replay attacks*: The adversary can launch replay attacks by sending duplicated ENF proofs. As ENF fluctuations of power grid vary as time changes, the duplicate ENF proofs generally output large ENF scores. As a result, these Byzantine validators have marginal chances to propose valid blocks. Furthermore, ENF-proof replay attacks can be detected by analyzing ENF proofs on EconLedger. Thus, identifying misbehavior and isolating suspicious nodes can improve system robustness, while we leave ENF-based detection topics to future work.
