There are at least three standard models for distributed data-sharing, and as we think about the right model for us to use, these should be on the table:
- “central CBPP-style” as used by !WorldCat
([http://en.wikipedia.org/wiki/Worldcat WP page]) or the international banking system. (tightly-integrated “push” model)
- “OAI-style” as used by OAI
([http://en.wikipedia.org/wiki/Open_Archives_Initiative WP page], [http://www.openarchives.org/ homepage]) (weakly-integrated “pull” model).
- “p2p-style” as used by Chord etc. (see below)
“The great achievement of Napster was the empowerment of the peers (i.e., the fringes of the network) in association with a central index, which made it fast and efficient to locate available content.” -[http://en.wikipedia.org/wiki/Peer-to-peer Wikipedia] “An unstructured P2P network is formed when the overlay links are established arbitrarily. Such networks can be easily constructed as a new peer that wants to join the network can copy existing links of another node and then form its own links over time.” – Like Gopher or early WWW. (ibid.) “Most networks and applications described as peer-to-peer actually contain or rely on some non-peer elements, such as DNS.” (ibid.) “Peer-to-peer networks have also begun to attract attention from scientists in other disciplines, especially those that deal with large datasets such as bioinformatics.” (ibid.) “P2P network defense is in fact closely related to the “Byzantine Generals Problem”. ”
- http://en.wikipedia.org/wiki/Chord_project – Systems we have prototyped
that use Chord and DHash include CFS (Cooperative File System), !UsenetDHT, and !OverCite. CFS allows anyone to publish and update their own file system, and provides read-only access to others; !UsenetDHT allows Usenet servers to share storage instead of fully replicating articles locally; !OverCite is a distributed version of the !CiteSeer digital library.
http://freepastry.org/PAST/default.htm
– From the project homepage I learned that ” P-Grid is implemented in Java and at the moment is available upon request for research purposes only.” I don’t know if that’s what we want.
interesting (“uses locality”) but I didn’t find a source download on the MS Research development page!
- http://en.wikipedia.org/wiki/LionShare – GPL, “designed with academic
users in mind” http://lionshare.its.psu.edu/
- http://www.firstauthor.org/Downloads/P2P.pdf (a whitepaper that claims to
discuss use of P2P in science) And this looks interesting:
The Internet was designed to provide a communications channel that is as resistant to denial of service attacks as human ingenuity can make it. In this note, we propose the construction of a storage medium with similar properties. The 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. The detailed design of this service is an interesting scientific problem, and is not merely academic: the service may be vital in safeguarding individual rights against new threats posed by the spread of electronic publishing. -http://www.cl.cam.ac.uk/~rja14/eternity/eternity.html
Key based routing:
All DHT topologies share some variant of the most essential property: for any key k, the node either owns k or has a link to a node that is closer to k in terms of the keyspace distance. It is then easy to route a message to the owner of any key k using the following greedy algorithm: at each step, forward the message to the neighbor whose ID is closest to k. When there is no such neighbor, then we must have arrived at the closest node, which is the owner of k as defined above.
– http://en.wikipedia.org/wiki/Distributed_hash_table
This kind of hopping seems similar to issues with Clusions (see below).
----
How does p2p work relative to the functioning of a given ISP? Presumably every “peer” on the network has its “online image” at the ISP level, and so the servers provided by the ISPs are somehow the “effective peers” in the network. (Everything else being simple “local rendering”.)
http://www.ietf.org/rfc/rfc1.txt – details a method to avoid message flooding by limiting the use of individual “links” between peers. (Note: this suggests the possibility of using Arxana to model or “simulate” the behavior of a peer-to-peer system, not just embedding it within one.) The system described here provides 32 links for potential peers to connect to, but typically can only support a couple of connections at once.
Notice that there are two different problems when we consider “sychronous” or “asynchronous” behavior (both are just abstractions anyway). The HCI for math text doesn’t need to be GUI or in any way fancy – but some people will no doubt want full-immersion math environments. I think that we can assume that we are transmitting things like text files (or streams) and in either event they don’t sound too threatening.