Isaac Sheff

Heterogeneous Consensus

The Heterogeneous Consensus project has invented and implemented a Consensus algorithm in which not all participants agree on who may fail, and how. We in some ways resemble the Stellar project, but our algorithm tolerates mixed Byzantine and Crash failures, one message-send lower latency, and a different model of Observers and Participants. At the moment, the best way to learn about Heterogeneous Consensus is to read the presentation from 2017 IC3 Bootcamp:

• Keynote
• PDF (with notes)
• PDF (slides only)

Implementation

We are working on an (unoptimized) reference implementation, featuring interoperable language-agnostic APIs and generic across cryptographic primitives. The (incomplete, not even alpha-level) code is on github.

Members

We are computer Scientists at Cornell:

• Isaac Sheff,
• Andrew Myers, and
• Robbert van Renesse.