LARK -- Linearizability Algorithms for Replicated Keys in Aerospike

TL;DR

LARK is a new replication protocol for distributed key-value stores that achieves linearizability with lower latency and higher availability by eliminating ordered logs and reasoning over the entire cluster.

Contribution

LARK introduces Partition Availability Conditions (PAC) and a log-free design, enabling higher availability and immediate partition readiness compared to traditional consensus protocols.

Findings

LARK improves partition availability by roughly 3x with one failure and 10x with two failures.

LARK enables immediate partition readiness after leader changes, reducing downtime.

LARK maintains commit capabilities during data-node failures and allows zero-downtime rolling restarts.

Abstract

We present LARK (Linearizability Algorithms for Replicated Keys), a synchronous replication protocol that achieves linearizability while minimizing latency and infrastructure cost, at significantly higher availability than traditional quorum-log consensus. LARK introduces Partition Availability Conditions (PAC) that reason over the entire database cluster rather than fixed replica sets, improving partition availability under independent failures by roughly 3x when tolerating one failure and 10x when tolerating two. Unlike Raft, Paxos, and Viewstamped Replication, LARK eliminates ordered logs, enabling immediate partition readiness after leader changes -- with at most a per-key duplicate-resolution round trip when the new leader lacks the latest copy. Under equal storage budgets -- where both systems maintain only f+1 data copies to tolerate f failures -- LARK continues committing through data-node failures while log-based protocols must pause commits for replica rebuilding. These properties also enable zero-downtime rolling restarts even when maintaining only two copies. We provide formal safety arguments and a TLA+ specification, and we demonstrate through analysis and experiments that LARK achieves significant availability gains.