Trevor Perrin

The Quiet Architect of Private Conversation

Most people who changed the internet did it loudly. Trevor Perrin did it in a spec document with a salamander joke buried in the footnotes. The Double Ratchet Algorithm - the cryptographic engine inside Signal, WhatsApp, and a dozen other secure messaging apps - was originally called the Axolotl Ratchet. Named for a Mexican amphibian that can regenerate not just limbs but heart tissue and parts of its brain. The algorithm heals itself forward the same way. Compromise a key today and tomorrow's messages are still private.

That's not metaphor. That's forward secrecy working the way it's supposed to. And that's the lens through which Perrin builds things: elegant, rigorous, designed to survive contact with reality.

Perrin spent years at Cryptography Research - the firm Paul Kocher built after breaking smartcard security with timing attacks in 1996 - before linking up with Moxie Marlinspike at Open Whisper Systems. When they sat down to design the protocol that would become Signal, they weren't trying to build something popular. They were trying to build something correct.

The timing turned out to be everything. 2013. The Snowden documents had just hit. The world discovered that "secure" and "private" were not the same thing. Perrin and Marlinspike had already written the math.

Three Protocols. One Billion Users.

Most cryptographers produce papers. Perrin produces infrastructure. The protocols he has designed or co-designed now run inside the apps that billions of people trust with their most sensitive conversations.

The Protocol That Changed Everything

Before Signal Protocol existed, secure messaging meant OTR - Off-the-Record Messaging. OTR was good. It gave you forward secrecy for a session. But it assumed continuous connection and required both parties online simultaneously. Smartphones had already killed that assumption.

Perrin and Marlinspike had to solve asynchronous delivery while preserving all the security properties that made OTR worth using. The result was TextSecure v2, introduced in February 2014. It used Triple Diffie-Hellman for the initial key agreement - combining identity keys, signed prekeys, and ephemeral one-time keys into a handshake elegant enough that academic cryptographers spent years writing papers proving why it works.

WhatsApp adopted it in 2016, encrypting end-to-end for 1 billion users overnight. That was the largest single cryptographic deployment in history at the time. Perrin's name appeared in the acknowledgments. The quiet ones usually are in acknowledgments.

Build Your Own Secure Channel

After Signal Protocol, Perrin stepped back and asked a bigger question. Everyone building secure communications was reinventing the same wheel. TLS was too complex and carried 30 years of backward compatibility baggage. What if there was a framework - not a protocol, but a language for building protocols - that let you specify exactly the handshake you needed?

That's the Noise Protocol Framework. Released in 2015, refined through 2016. Public domain. No license. No restrictions. A set of handshake patterns that compose like building blocks, each with formally provable security properties.

Jason Donenfeld, building WireGuard, chose Noise_IKpsk2 as the foundation for what became the most praised VPN protocol in a decade - eventually merged into the Linux kernel. Lightning Network, the Bitcoin payment layer, uses Noise to authenticate node connections. Slack uses it for client transport. The public domain release was deliberate: maximum adoption, zero friction.

The citation understated it. By 2017, the Signal Protocol had become the de facto standard for secure messaging - not because regulators mandated it, not because it was easy to adopt, but because it was simply the best option and the authors had made it open enough that anyone could use it. That's a different kind of prize.

Before Signal, There Was the Specification

Perrin graduated from UC Berkeley and arrived at Cryptography Research - Paul Kocher's firm in San Francisco - at a time when applied cryptography was beginning to diverge sharply from academic cryptography. The field had great theory. It had terrible implementation. Kocher's group was obsessed with the gap between the two.

In 2007, Perrin co-authored RFC 5054: the Secure Remote Password (SRP) protocol for TLS Authentication. Not glamorous work. Exactly the kind of cryptographic plumbing that the internet runs on invisibly. That same year, he presented on public key distribution through "cryptoIDs" at NSPW 2003 - work that anticipated identity-binding problems that still haven't been fully solved.

The TACK TLS extension, which Perrin designed, proposed a form of key continuity for TLS certificates - a way for servers to build up trust over time rather than relying entirely on certificate authorities. It was ahead of its time. Certificate Transparency, which Google later deployed widely, addressed some of the same problems through different means.

On Stage

The Quantum Problem

The encryption that Perrin helped build is very good. Good enough that the primary way adversaries bypass it is to store encrypted traffic and wait for a quantum computer powerful enough to break the key exchange. "Harvest now, decrypt later." The threat isn't today. But the decisions you make today about what you encrypt and how determine whether 2030's quantum computers find anything useful in your archived traffic.

This is where Perrin is now: independent consultant, working on quantum readiness and cryptographic inventory. Helping organizations understand what they actually have deployed, where the exposure is, and what needs to be upgraded before the window closes. The Signal team added post-quantum cryptography to the Signal Protocol in 2023 using the PQXDH specification - Perrin's foundational work made that upgrade tractable.

The pattern is consistent. He builds things that can be upgraded. The axolotl can regenerate. The ratchet keeps turning.

Five Things

Links