A brief introduction to Yenta

Every user runs their own copy of Yenta, and the various Yentas communicate with each other, arranging introductions and passing messages. Personal information is tightly protected, without requiring large amounts of trust either in the network or in any Yenta not under your own control. The eventual goal is the ubiquitous distribution of Yenta across the Internet.

There are three major motivations for this work:

• Technical. I'd like to advance the state of the art in building distributed, decentralized, multi-agent systems. There are a number of problems to be addressed in making these sorts of applications, especially if one cares about protecting people's privacy at the same time.
• Sociological. The net is a big place, and it's sometimes difficult to find the communities you'd like to find. As the net gets bigger and the signal-to-noise ratio decreases, this seems to be getting even harder. Investigating how Yenta can help, and also how people use its built-in reputation system, could be very interesting.
• Political. Too many systems that handle personal information think about privacy and civil liberties issues last, when it's too late to fix often horrendous gaffes in these areas. This project is an attempt to show that it's possible to build such a such in which privacy issues are taken seriously. Yet building robust, secure, private, and distributed systems requires strong cryptography, so part of the motivation here is also to show that strong cryptography is not just for the "bad guys", whatever your definition of "bad guys" might be.

Getting your own copy of Yenta

If you think you might be interested in helping to port Yenta to a new platform, please let me know.

Finding out more about Yenta

• Information for programmers:
  • Internal documentation.
  • Help critique the code. This uses a brand-new approach called Yvette, written as part of the Yenta project, which allows many people to collaborate on reviewing a large body of code.
• A PhD dissertation.
• A variety of short papers:
  • Yenta: A Multi-Agent, Referral Based Matchmaking System, presented at The First International Conference on Autonomous Agents (Agents '97), Marina del Rey, California, February '97.
  • A Security Architecture for Multi-Agent Matchmaking (Postscript or PDF), presented at The Second International Conference on Multi-Agent Systems at Keihanna Plaza, Kansai Science City, Japan, December '96.
  • A Multi-Agent Referral System for Matchmaking (Postscript or PDF), presented at The First International Conference on the Practical Applications of Intelligent Agents and Multi-Agent Technology, London, UK, April '96.
  • Clustering and Information Sharing in an Ecology of Cooperating Agents (Postscript or PDF), presented at The AAAI Spring Symposium '95 on Information Gathering in Distributed, Heterogeneous Environments, Palo Alto, CA.
  • Clustering and Information Sharing in an Ecology of Cooperating Agents, or How to Gossip without Spilling the Beans (Postscript or PDF), presented at The 1995 Conference on Computers, Freedom, and Privacy. This paper was the Student Scholarship Paper Winner at CFP '95.
• A complete description of the basic idea, written in late 1994 and early 1995, as a large, bushy bunch of hypertext. Note that this material should be considered historical and hence I make no claims to keep it regularly updated. As such, it will contain citations to old papers and old results.