How Does Lightning Network Work?

Bitcoin 2-of-2 multisig UTXO with revocable commitment transactions and penalty (justice) transactions for breach remedy

Well-established since 2016 Poon-Dryja paper; penalty mechanism requires watchtower monitoring to enforce; channel capacity limited by on-chain UTXO amount

Script-enforced hash preimage reveal with CLTV/CSV time-lock expiry for multi-hop atomic routing

Core routing primitive; stuck HTLCs during path failure can lock liquidity for hours; downstream failure can cascade upstream

BOLT #7 channel announcements and updates broadcast to maintain distributed routing table across all LN nodes

Routing efficiency degrades as network scales; source-based routing means sender must have accurate global channel capacity knowledge, which is probabilistic

Third-party monitoring services detect old commitment broadcasts and submit penalty transactions on behalf of offline channel parties

Users who go offline without a watchtower are vulnerable to counterparty broadcasting stale states; watchtower trust and liveness are critical dependencies

Sphinx packet construction for privacy-preserving multi-hop payment routing — intermediate nodes see only adjacent hops

Provides payment privacy but probing attacks can infer channel balances; timing correlation is a known deanonymization vector

Interactive transaction construction (BOLT #2 updates) allowing in-place channel capacity adjustments without closing and reopening

Splicing is a newer feature with limited implementation maturity; reduces on-chain footprint but introduces new edge cases in concurrent splice state machines

Channel jamming attacks: an attacker routes HTLCs through victim channels with no intent to settle, holding slots and liquidity hostage. Gossip network provides no spam prevention, so attackers can identify and target high-value routing paths at low cost.

Watchtower liveness dependency: if a user's watchtower is offline when their counterparty broadcasts a revoked commitment transaction, the justice transaction window may expire and the attacker can steal channel funds. Single point of failure for offline users.

Force-close cascade during Bitcoin fee spikes: HTLC expiry deadlines force channel closures onto the base layer simultaneously. During high-fee periods, small HTLC outputs become economically unrecoverable as fees exceed value, causing real fund losses.

Balance probing attacks: adversaries send probe payments to infer private channel balances from failure messages despite onion routing privacy. Combines with gossip-derived topology to map liquidity positions across the network.

Splice state machine race conditions: concurrent in-flight HTLCs during an active splice can create ambiguous commitment state between peers, potentially leading to disagreement on valid channel state and unilateral closes.