Independent tech intelligence, checked against primary sources.

TechPulseMind Useful technology.
No manufactured hype.

Cloudflare’s Meerkat Picks QuePaxa: A Leaderless Consensus Algorithm for Global Scale

Cloudflare explains why its experimental Meerkat consensus service uses QuePaxa instead of Raft or Paxos, citing leader dependence and global network volatility.

Cloudflare’s Meerkat Picks QuePaxa: A Leaderless Consensus Algorithm for Global Scale

Before you start: versions, package names and storage identifiers change over time. Check the commands against the official documentation for your setup before running them, especially anything that creates, deletes or overwrites data.

Cloudflare has explained why its experimental distributed consensus service, Meerkat, runs on QuePaxa rather than Raft, (Multi-)Paxos or Viewstamped Replication. In a post published November 11, engineers João Pedro Leite, Bob Halley and James Larisch said the choice comes down to how those older algorithms behave when the network, not the servers, is the problem.

Meerkat is a distributed system in which replicas agree on an ordered log of events, with each event decided by a consensus algorithm. Cloudflare says the service is designed to manage its internal global control plane and to expose applications such as a key-value store, where each log entry is a get or a put. Comparable consensus-powered systems include Consul, etcd and Apache ZooKeeper.

What sets Meerkat apart, according to the company, is global scale. A cluster may span continents over links with high and variable latency — an unusual setup for consensus systems, which are usually deployed in smaller, more local networks with strict latency requirements.

Why the familiar algorithms struggle

Cloudflare says common consensus algorithms often suffer in wide-area networks with volatile connectivity and latency, such as the Internet. They rely on a single leader process being available and having predictable latency. If that leader goes down, slows down or is cut off, the whole system stalls or goes unavailable until a replacement takes over.

The company points to its own history as evidence. In 2020, the partial failure of a core Cloudflare network switch caused a consensus-powered system to degrade and the dashboard to run 80x slower than usual for six hours. No data was lost, but writes were unavailable until the issue was resolved — something Cloudflare calls unacceptable.

According to the post, QuePaxa promises better stability under volatility, largely because it does not depend on a leader to make progress. The authors also describe two traits that can degrade conventional algorithms under hostile network conditions — destructive interference and fault detection — and say QuePaxa sidesteps both.

What it means if you run distributed systems

Meerkat is still described by Cloudflare as experimental, and the post is an explanation of the design rationale rather than a product launch. The full write-up, including the authors’ walkthrough of how QuePaxa works, is available on the Cloudflare blog. Cloudflare has separately been expanding its agent and analytics tooling, as covered in our recent report on its personal assistant analytics and AI bot access policies.

Related reading

Sources

Some links on this page may be affiliate links. If you buy through them we may earn a commission at no extra cost to you. See our affiliate disclosure.