MD-5: Fast Finality Settlement - Long Range Attacks
- Description: Consider and model long-range attacks in the Fast Finality Settlement protocol.
- Authors: Liam Monninger, Andreas Penzkofer
Overview
Long-range attacks are an inherent vulnerability to Proof-of-Stake (PoS) systems. Without an anchoring mechanisms blockchains are exposed to the risk of these.
The Fast Finality Settlement (MIP-34) is a PoS based settlement mechanism and separates into
- a Fastconfirmation protocol (MIP-64) on L2. Fastconfirmations confirm L2Blocks.
- a Postconfirmation protocol (MIP-37) on L1. Postconfirmations confirm superBlocks, which are a sequence of L2Blocks. Postconfirmations act as an anchoring mechanism.
Postconfirmation The stake and finality of Postconfirmations is protected by the underlying L1 finality. However, since the L1 is itself is a PoS system it may be susceptible to reorgs. Consequently, in combination with a reorg on L1, Postconfirmations may be reverted, and the anchoring invalidated. The invalidation of anchors can permit long-range attacks on superBlocks.
Albeit unlikely due to the high security of L1, long-range attacks in the Postconfirmation protocol should be considered and modeled.
Fastconfirmation Fastconfirmation on L2 are anchored through Postconfirmations on L1. An anchoring mechanism can protect against long-range attacks on L2Blocks. However, Fastconfirmations should consider the Postconfirmations.
L2Blocks that are not postconfirmed are susceptible to long-range attacks.
Introduction
This reference categorizes long-range attacks into
- Simple Attack: The attacker forges an alternative blockchain branch from some earlier block, bypassing timestamp checks to produce blocks ahead of time and overtake the main chain.
- Posterior Corruption: A retired validator is bribed or hacked to sign blocks on an alternate chain, allowing the attacker to increase their block production and potentially surpass the main chain.
- Stake Bleeding: The attacker stalls the main chain by skipping block production while accumulating stake and rewards on a hidden branch, eventually overtaking the main chain.
Desiderata
D1: Evaluate the opportunity for “simple” long-range attacks and detection thereof
User Journey: A researcher or protocol implementer can understand how a simple long-range attack could be executed and detected.
Justification: Providing a model which demonstrates how a simple long-range attack could be executed and detected could enable us to derive preventative measures.
D2: Evaluate the opportunity for “posterior corruption” long-range attacks and detection thereof
User Journey: A researcher or protocol implementer can understand how a posterior corruption long-range attack could be executed and detected.
Justification: Posterior corruption long-range attacks are a known vulnerability to PoS networks. Providing a model which demonstrates how a posterior corruption long-range attack could be executed and detected could enable us to derive preventative measures.
D3: Evaluate the opportunity for “stake-bleeding” long-range attacks and detection thereof
User Journey: A researcher or protocol implementer can understand how a stake-bleeding long-range attack could be executed and detected.
Justification: Stake-bleeding long-range attacks are a known vulnerability to PoS networks. Providing a model which demonstrates how a stake-bleeding long-range attack could be executed and detected could enable us to derive preventative measures.