Freenet: A Distributed Anonymous Information Storage and Retrieval System
Reviewer: Robert Dugas
Although numerous distributed file-sharing application have been
developed, none provide anonymity for both information providers and
The freenet system represents an attempt to implement an anonymous
distributed file sharing system without prohibitive overhead for
storage and retieval operations.
Stores data according to lexical hashes of document keys
Replicates information along path of retrieval to help ensure
Peer-to-peer architecture protects identity of requestors and privders of
Freenet represents one of the first implementations of a viably effective
distributed file-sharing system. Although no particularly innovative
algorithms are presented the combined aspects of Freenet's architecture
represent an effective and robust means of achieving the system's goals.
Freenet seems to have succeeded in spite of the methodology behind its
creation, rather than because of it. Testing and analysis of performance
characteristics seems to be just now taking place after implementation rather than during design and planning.
The primary limitation, admitted but not fully addressed in the paper,
is the flat namespace of the file system.Information providers must
produce globally unique identifiers on their own through trial and
error before successfully storing new information. In addition, the system has few provable bounds on performance in general. Finally,
while much emphasis is placed on the anonymity provided by Freenet,
the authors admit to the system's vulnerability to concerted attack.
The primary lesson of freenet is that distributed, relatively anonymous, peer-to-peer file sharing is a feasible endeavor.