Before Blockchains, There Was State Machine Replication
Episode
37 min
Read time
2 min
Topics
Remote Work, Relationships, Design & UX
AI-Generated Summary
Key Takeaways
- ✓State Machine Replication as blockchain foundation: Every blockchain consensus mechanism descends from state machine replication principles developed in the 1980s. The core idea — that replicas maintain identical logs of ordered operations regardless of the application layer — is why Ethereum and Solana can run arbitrary smart contracts. Builders should understand SMR as the conceptual layer beneath any consensus protocol they evaluate or implement.
- ✓View-change protocol for fault tolerance: ViewStamp Replication solved the "embarrassing pause" problem in two-phase commit by introducing a structured primary-to-backup failover mechanism. When a primary replica stops responding, backups execute a view-change protocol electing a new primary while preserving the complete prior history. Modern blockchain leader-rotation mechanisms — used in protocols like HotStuff — directly inherit this architecture.
- ✓PBFT's 3f+1 replica threshold for Byzantine tolerance: Handling malicious nodes requires 3f+1 total replicas to tolerate f Byzantine failures, versus only 2f+1 for benign crash failures. PBFT addresses lying primaries by requiring two-thirds majority certificates — collections of 2f+1 signed messages confirming the same state — before committing any operation. Blockchain engineers designing validator sets should size them against this threshold explicitly.
- ✓Cryptographic certificates as trust anchors: PBFT introduced the use of signed certificate bundles — 2f+1 matching signed messages — as verifiable proof that a protocol checkpoint was reached honestly. This mechanism, borrowed from theoretical cryptography and made practical by PBFT, is now the standard method by which proof-of-stake blockchains like Ethereum finalize blocks and detect equivocation by validators.
- ✓AI shifts coding toward specification and verification: Liskov argues that as AI generates code from specifications, the critical skill becomes verifying correctness rather than writing syntax. The MIT course she co-developed with John Guttag since the late 1970s — centered on design, modularity, specifications, and verification of large systems — describes exactly the competency profile developers need to audit and manage AI-generated code effectively.
What It Covers
Turing Award winner Barbara Liskov traces the origins of blockchain consensus protocols to her 1980s distributed systems research at MIT, explaining how ViewStamp Replication and Practical Byzantine Fault Tolerance (PBFT) became the foundational architecture underlying every major blockchain protocol operating today.
Key Questions Answered
- •State Machine Replication as blockchain foundation: Every blockchain consensus mechanism descends from state machine replication principles developed in the 1980s. The core idea — that replicas maintain identical logs of ordered operations regardless of the application layer — is why Ethereum and Solana can run arbitrary smart contracts. Builders should understand SMR as the conceptual layer beneath any consensus protocol they evaluate or implement.
- •View-change protocol for fault tolerance: ViewStamp Replication solved the "embarrassing pause" problem in two-phase commit by introducing a structured primary-to-backup failover mechanism. When a primary replica stops responding, backups execute a view-change protocol electing a new primary while preserving the complete prior history. Modern blockchain leader-rotation mechanisms — used in protocols like HotStuff — directly inherit this architecture.
- •PBFT's 3f+1 replica threshold for Byzantine tolerance: Handling malicious nodes requires 3f+1 total replicas to tolerate f Byzantine failures, versus only 2f+1 for benign crash failures. PBFT addresses lying primaries by requiring two-thirds majority certificates — collections of 2f+1 signed messages confirming the same state — before committing any operation. Blockchain engineers designing validator sets should size them against this threshold explicitly.
- •Cryptographic certificates as trust anchors: PBFT introduced the use of signed certificate bundles — 2f+1 matching signed messages — as verifiable proof that a protocol checkpoint was reached honestly. This mechanism, borrowed from theoretical cryptography and made practical by PBFT, is now the standard method by which proof-of-stake blockchains like Ethereum finalize blocks and detect equivocation by validators.
- •AI shifts coding toward specification and verification: Liskov argues that as AI generates code from specifications, the critical skill becomes verifying correctness rather than writing syntax. The MIT course she co-developed with John Guttag since the late 1970s — centered on design, modularity, specifications, and verification of large systems — describes exactly the competency profile developers need to audit and manage AI-generated code effectively.
Notable Moment
When the Google File System paper was published in the late 1990s citing Paxos, Liskov was initially frustrated — until a former student at Google recognized that the GFS replication approach was functionally identical to ViewStamp Replication, two protocols developed independently without either team realizing they had solved the same problem.
Episode Transcript
DARPA had recognized that this was a serious problem, the problem of malicious attacks, and was looking for research in that area. I had a student, Miguel Castro. He came to me and he said, why don't we see whether we can figure out a way to do replication that handles these malicious attacks? And that seemed like a great idea. The problem, if the primary failed, the primary that was running the protocol, the whole thing came to a crashing halt. We came up with a protocol that if the primary seemed to not be doing its job, the backups then carried out another protocol in which a different replica became the primary. We thought that at some point, people would start to use this, and then along came blockchains. And so that was very funny. Someone who loves computer science, loves research, wants to do important work, wants to have an impact, What would you suggest as kind of a mindset, as an approach? Computer science is in a very strange place right now with the advent of AI. I think that as far as research is concerned As part of our summer feed drop series, we're sharing another episode from the a 16 z crypto show. Every blockchain today relies on replication techniques first developed in the nineteen eighties, years before anyone imagined cryptocurrencies. In this episode, Tim Roughgarden and Atai Abraham speak with Turing Award winner, Barbara Liskov, whose pioneering work on distributed systems, state machine replication, and practical Byzantine fault tolerance laid the foundation for many of today's blockchain protocols. Whether you're new to crypto or already building in the space, this conversation offers a deeper understanding of the computer science behind modern blockchains. If you enjoy this episode, be sure to follow the a sixteen z crypto show for more conversations like this. Hi, everyone, and welcome to First Principles, a series by the team at a sixteen z Crypto. I'm Tim Roughgarden. I'm the head of research at a sixteen z Crypto. And today, we're exploring a core challenge in distributed computing. How can a system remain reliable when some of its parts fail or behave maliciously? In this episode, we're going to speak with the Turing award winning computer scientists whose work helped answer this question, Doctor. Barbara Liskoff. She's famous for many things, but today we're gonna focus on her work that laid the foundations for modern Blockchain systems. For example, from view standard application from the nineteen eighties, to Practical Byzantine Fault Tolerance or PBFT a decade later. A16z crypto research partner, Ittai Abraham, joins me as we talk with Doctor. Liskov about her path from programming languages to distributed systems, the evolution of practical replication protocols, the state machine replication paradigm underlying blockchains, and how emerging technologies like AI are reshaping systems research. Here's our conversation. Professor Liskopf, thanks so much for joining us for the conversation today. I really appreciate you taking the time. Sure. I thought maybe …
Get the full transcript (6,709 words) + summary by email — free
One-time email with the complete transcript and AI summary of this episode. No account needed.
One email, no spam. We’ll also show you what SignalCast does.
You just read a 3-minute summary of a 34-minute episode.
Get a16z Podcast summarized like this every Monday — plus up to 2 more podcasts, free.
Pick Your Podcasts — FreeKeep Reading
More from a16z Podcast
The State of AI: Macro, Apps, and Consumer
Aug 26 · 37 min
Odd Lots
Brendan Greeley on the Real 500-Year History of the Dollar
May 28
More from a16z Podcast
The New Economics of AI | Martin Casado & Steven Sinofsky
Aug 25 · 63 min
Deep Questions with Cal Newport
AI Reality Check: Are LLMs a Dead End?
Mar 26
Books, tools, and gear mentioned in this episode
SignalCast may earn commission on purchases via these links. As an Amazon Associate, SignalCast earns from qualifying purchases.
Books
“When the Google File System paper was published in the late 1990s citing Paxos, Liskov was initially frustrated — until a former student at Google recognized that the GFS replication approach was functionally identical to ViewStamp Replication”
course
- MIT Course on Design, Modularity, Specifications, and Verification of Large SystemsRecommendedBy guest
by MIT
“The MIT course she co-developed with John Guttag since the late 1970s — centered on design, modularity, specifications, and verification of large systems — describes exactly the competency profile developers need to audit and manage AI-generated code effectively.”
More from a16z Podcast
We summarize every new episode. Want them in your inbox?
The State of AI: Macro, Apps, and Consumer
The New Economics of AI | Martin Casado & Steven Sinofsky
Why Medical AI Needs a Referee | Protege's Engy Ziedan
Martin Casado on Where the Value Is Going in AI
How Microsoft Is Securing the Agentic Enterprise | Aaron Zollman
Similar Episodes
Related episodes from other podcasts
Odd Lots
May 28
Brendan Greeley on the Real 500-Year History of the Dollar
Deep Questions with Cal Newport
Mar 26
AI Reality Check: Are LLMs a Dead End?
Everything Everywhere Daily
Mar 12
The Purple Heart
Dwarkesh Podcast
Sep 26
Richard Sutton – Father of RL thinks LLMs are a dead end
Stuff You Should Know
Aug 26
SYSK Tribute: Why does everyone love Dolly Parton?
Explore Related Topics
This podcast is featured in Best Business Podcasts (2026) — ranked and reviewed with AI summaries.
You're clearly into a16z Podcast.
Every Monday, we deliver AI summaries of the latest episodes from a16z Podcast and 192+ other podcasts. Free for one show.
Start My Monday DigestNo credit card · Unsubscribe anytime