Steward: Scaling Byzantine Replication to Wide-Area Networks

Overview

Steward is a hierarchical Byzantine fault-tolerant replication system developed as part of DARPA's Self-Regenerative Systems project. The system is suited to deployments consisting of multiple wide-area sites, each with a cluster of replication servers. Steward continues to work correctly even if some of the replication servers in each site are compromised and act in an arbitrary manner.

Steward deploys a hierarchical architecture. Each site runs local Byzantine fault-tolerant protocols to confine the effects of malicious replias to their local site. Thus, each site acts as a trusted logical unit. A benign fault-tolerant replication protocol runs among the logical units over the wide-area network.

Contributors

• Johns Hopkins University
  • Yair Amir
  • Claudiu Danilov
  • Jonathan Kirsch
  • John Lane
• Purdue University
  • Cristina Nita-Rotaru
  • Josh Olsen
  • David Zage
• The Hebrew University
  • Danny Dolev

Funding

Our work on Steward was partially funded by the Defense Advanced Research Project Agency (contract FA8750-04-2-0232), and by the National Science Foundation (grants 0430271 and 0430276).

Software

Steward was fully implemented and underwent a Red Team evaluation in December 2005. The code was written in C and runs on Linux.

License

Download

Related Publications

• STEWARD: Scaling Byzantine Fault-Tolerant Replication to Wide Area Networks
      Yair Amir, Claudiu Danilov, Danny Dolev, Jonathan Kirsch, John Lane, Cristina Nita-Rotaru, Josh Olsen, David Zage
      IEEE Transactions on Dependable and Secure Computing, Volume 7 Number 1, pp. 80-93, 2010.
   
• Scaling Byzantine Fault-Tolerant Replication to Wide Area Networks
      Yair Amir, Claudiu Danilov, Danny Dolev, Jonathan Kirsch, John Lane, Cristina Nita-Rotaru, Josh Olsen, David Zage
      IEEE International Conference on Dependable Systems and Networks (DSN 2006), Philadelphia, PA, June 2006, pp. 105-114.
      Award paper.