Most current network designs and protocols presume that the agents within will watch out for each other's interests. An international network with software from untold and unknown hoards hardly suits this concept. The key missing link is feedback: unless good behavior is rewarded and bad behavior is penalized, a network of diversely interested parties cannot collaborate effectively.

One solution is to introduce a multi-tiered economy, similar to that of our physical world, into cyber-space. Ad revenue will not support these lower layers. Contracts help but cost time and money to establish. They do not serve the individual entrepreneur very well, especially in early stages. A finer-grain, more flexible mechanism is needed at lower levels of the network.

Digital Silk Road is a micropayment-based network design that attends to many of these issues at the lowest layers. When we are advised to ``follow the money,'' big money is usually meant. In Digital Silk Road, little money flows together with data to enable a set of possibilities that networks do not currently aspire to. Following the small money leads to fascinating possibilities. This talk will describe the core ideas of Digital Silk Road, and discuss some of the applications that it enables.

About the Speaker:

Norman Hardy has had a distinguished career spanning processor architecture, operating systems, distributed systems, and things at NSA that he still can't talk about. He was a member of the Advanced Computing Systems (ACS) group at IBM -- the group which created both the IBM 360 and the 801 processors (the 801 was the first RISC machine). At Tymshare, Inc., Norm played a key role in the creation of Tymnet, one of the first wide-area terminal switching networks. Later at Tymshare, he was the principal architect of the KeyKOS operating system, the first workable capability-based operating system and one of the systems studied by the National Security Agency as a basis for the Trusted Computer Security Evaluation Criteria (TCSEC, better known as the ``Orange Book''). KeyKOS has influenced most later microkernel research, including our own work on EROS here at JHU. Most recently Norm has worked for Agorics, Inc, a Silicon Valley-based software company that builds distributed, market-based computational systems.

Mr. Hardy's peculiar range of experience gives him a fairly unusual outlook on software systems design. His designs simultaneously consider not only functional requirements, but also resource management, precise architectural specification, and the demands of high-performance implementation. This leads routinely to system designs that are initially non-obvious and later discovered to be quite beautiful as their end-to-end cohesiveness emerges in hindsight.