Cassandra: Consensus with Partial Progress via Robust Partitionable View Synchronization

2026-07-02

AI Summary

This work addresses the loss of liveness in traditional Byzantine fault-tolerant protocols during network partitions, which arises from their reliance on strong quorums. The authors propose a two-layer authentication framework that decouples consensus availability from commitment, enabling each partition to independently extend its local chain while ensuring safe reconciliation upon network recovery. To support progress during partitions without requiring strong quorums, they introduce a view synchronization pacemaker that integrates proof-of-authority (PoA)-based speculative execution with timeout calibration along non-critical paths. Experimental results demonstrate that the system achieves up to 900,000 TPS with 16 nodes and 480,000 TPS with 104 nodes under stable network conditions (latency: 0.31–0.75 seconds), while sustaining non-zero speculative throughput during severe partitions—all speculative work remaining safely reconcilable afterward.

Abstract

Replicated databases and permissioned blockchain systems rely on Byzantine Fault-Tolerant (BFT) consensus to maintain a globally consistent order of transactions across distributed replicas. These protocols preserve safety even under asynchrony, as they commit a transaction only after agreement among a strong quorum of replicas. During network partitions, however, when no strong quorum is reachable, they lose liveness and cannot make useful progress. In this paper, we present Cassandra, a consensus protocol that enables partial progress without sacrificing safety. Cassandra achieves this through a two-tier certification framework that decouples availability from commitment, allowing each partition to extend its own chain and reconcile these chains once the network is restored. To support this, Cassandra introduces a pacemaker that advances views without requiring a strong quorum and calibrates each replica's timeout off the critical path. Our evaluation results show that Cassandra remains competitive with state-of-the-art BFT protocols under stable conditions, sustaining 900K TPS at 16 replicas and 480K TPS at 104 replicas, with latency ranging from 0.31s at 16 replicas to 0.75s at 104 replicas. Under severe partitions, Cassandra maintains non-zero speculative throughput through PoA-backed progress, preserving work that can be reconciled once connectivity is restored.

Problem

• Byzantine Fault-Tolerant consensus
• network partitions
• liveness
• partial progress
• safety

Innovation

• Byzantine Fault Tolerance
• Partial Progress
• View Synchronization
• Partitionable Consensus
• Two-tier Certification