This work was inspired by Anderson's seminal paper on The Eternity Service [2]. As Anderson wrote, ``[t]he basic idea is to use redundancy and scattering techniques to replicate data across a large set of machines (such as the Internet), and add anonymity mechanisms to drive up the cost of selective service denial attacks.''
A publisher uploads a document and some digital cash, along with a requested file duration (cost would be based on document size and desired duration). In the simple design, a publisher would upload the document to 100 servers, and remember ten of these servers for the purposes of auditing their performance. Because he does not record most of the servers to whom he submitted the file, there is no way to identify which of the participating eternity servers are storing his file. Document queries are done via broadcast, and document delivery is achieved through one-way anonymous remailers.
There are issues which are not addressed in his brief paper: for instance, if documents are submitted anonymously but publishers are expected to remember a random sample of servers so they can audit them, what do they do when they find that some server is cheating? Anderson passes this responsibility on to the digital cash itself, so servers do not receive payment if they stop providing the associated service. He does not elaborate on the possible implications of this increased accountability to the anonymity of the publishers.
Eternity has several problems that hinder real-world deployment. Most importantly, Eternity relies on a stable digital cash scheme, which is not available today. There is no consideration to maintaining a dynamic list of available servers and allowing servers to smoothly join and leave. Anderson further proposes that a directory of files in the system should itself be a file in the system. However, without a mechanism for updating or revising files, this would appear very difficult to achieve.