CBSE Consensus Algorithm Security 3 — Questions and Answers
Question 1: What is a 'grinding attack' in the context of Proof-of-Stake consensus?
- An adversary iterates over many block header values to bias the next validator selection (Correct answer)
- An attacker floods the network with low-fee transactions to stall consensus
- A miner repeatedly re-mines the same block to earn double rewards
- An adversary degrades peer connections to force network partitions
Correct answer: An adversary iterates over many block header values to bias the next validator selection
In a grinding attack, a validator tries many candidate block values to manipulate the randomness source used for future leader selection.
Question 2: Delegated Proof-of-Stake (DPoS) introduces which unique centralization-related security risk compared to standard PoS?
- Collusion among a small set of elected delegates to censor transactions or double-spend (Correct answer)
- Validators losing stake due to network latency rather than malicious behavior
- An inability to fork the chain in response to protocol bugs
- Block rewards being distributed unevenly among token holders
Correct answer: Collusion among a small set of elected delegates to censor transactions or double-spend
DPoS concentrates block-production power in a small elected group, making delegate collusion or cartel formation a realistic attack vector.
Question 3: In Practical Byzantine Fault Tolerance (PBFT), the protocol phases are prepare, pre-prepare, and commit. Which phase ensures that all honest nodes agree on a single value before committing?
- Pre-prepare phase
- Prepare phase (Correct answer)
- Commit phase
- View-change phase
Correct answer: Prepare phase
The prepare phase collects 2f+1 matching prepare messages (where f is the fault threshold), ensuring quorum agreement on the proposed value before the commit phase finalizes it.
Question 4: What security property distinguishes 'safety' from 'liveness' in distributed consensus?
- Safety guarantees no two honest nodes commit conflicting values; liveness guarantees the system eventually makes progress (Correct answer)
- Safety guarantees fast block finality; liveness guarantees resistance to Sybil attacks
- Safety prevents double-spending; liveness prevents eclipse attacks
- Safety requires BFT node counts; liveness requires PoW hash rate thresholds
Correct answer: Safety guarantees no two honest nodes commit conflicting values; liveness guarantees the system eventually makes progress
Safety means the system never produces contradictory finalized outputs; liveness means honest nodes will eventually reach a decision even under adversarial conditions.
Question 5: Ethereum's Gasper consensus combines Casper FFG with LMD-GHOST. What does LMD-GHOST stand for?
- Latest Message Driven Greediest Heaviest Observed SubTree (Correct answer)
- Longest Message Directed Graph Heuristic Ordering System Tree
- Least-Median Distance Greedy Hash Ordered Subtree
- Latency-Minimized Directed Gossip Heuristic Over Stake Trees
Correct answer: Latest Message Driven Greediest Heaviest Observed SubTree
LMD-GHOST uses each validator's latest message to weight the fork-choice rule, selecting the subtree with the most accumulated validator support.
Question 6: An attacker who controls 34% of staked ETH in Ethereum's PoS can threaten which consensus property?
- They can immediately double-spend any transaction
- They can prevent finality by withholding enough votes to block the 2/3 supermajority (Correct answer)
- They can rewrite all blocks in the last 24 hours
- They can halt the P2P gossip layer entirely
Correct answer: They can prevent finality by withholding enough votes to block the 2/3 supermajority
Ethereum's Casper FFG requires a 2/3 supermajority to finalize checkpoints; 34% stake allows an attacker to indefinitely block finality without achieving majority control.
Question 7: What is the primary security purpose of a 'view-change' protocol in PBFT-based consensus systems?
- To rotate validators in and out of the active set based on stake changes
- To replace a faulty or unresponsive primary (leader) node while preserving safety (Correct answer)
- To trigger an emergency fork when more than 1/3 of nodes go offline
- To reset all pending transactions when a network partition is detected
Correct answer: To replace a faulty or unresponsive primary (leader) node while preserving safety
The view-change protocol allows replicas to elect a new primary when the current leader is suspected of failure or malice, maintaining liveness without sacrificing safety.
CBSE Consensus Algorithm Security 3
What is a 'grinding attack' in the context of Proof-of-Stake consensus?