Consensus Algorithm Security Flashcards

In a Proof-of-Authority (PoA) blockchain, what is the primary attack surface that does not exist in permissionless PoW networks?

Answer: Compromise or coercion of a small set of approved validator identities

PoA relies on a small, known set of authorized validators, so compromising or coercing even a few identities can give an attacker majority control.

Tendermint BFT achieves instant finality by requiring validators to sign what type of message before a block is committed?

Answer: A pre-commit vote witnessed by 2/3+ of validator voting power

Tendermint requires a pre-commit round where ≥2/3 of voting power signs the block; once this threshold is reached, the block is finalized immediately.

Which attack exploits predictable block timestamps or weak randomness beacons in PoS to manipulate validator selection?

Answer: Grinding attack

A grinding attack lets a block proposer iterate through candidate block headers to find a value that biases the randomness function toward selecting themselves or allies as future proposers.

What does 'weak subjectivity' mean in the context of Proof-of-Stake security?

Answer: New nodes joining the network must trust a recent checkpoint rather than syncing from genesis to avoid long-range attacks

Weak subjectivity acknowledges that PoS nodes syncing from scratch cannot cryptographically verify which chain is canonical without a trusted recent checkpoint.

An adversary who successfully executes a block withholding attack in a mining pool gains what strategic advantage?

Answer: They receive all block rewards while contributing no valid blocks to the pool's output

In a block withholding attack, a malicious pool member submits partial proofs (shares) but discards full solutions, collecting pay-per-share payments without contributing revenue to the pool.

Avalanche consensus achieves high throughput by using repeated sub-sampled voting. Which property makes it resilient to Sybil attacks?

Answer: It samples a random subset of nodes, so adding fake identities does not proportionally increase an attacker's influence without controlling real stake or compute

Avalanche's security is parameterized by the fraction of adversarial stake or nodes; simply creating many fake identities without proportional resources does not bias the sampling outcome.

In the context of consensus security, what is a 'DAG-based' consensus approach's primary security advantage over linear chain consensus?

Answer: DAG structures allow multiple blocks to be confirmed concurrently, reducing the window during which a double-spend can be introduced

DAG-based protocols (like IOTA, Phantom, or Conflux) confirm many transactions in parallel, increasing throughput and reducing the time an attacker has to build a competing chain.