|
|
Nocturnal
Nocturnal provides automated information
sharing between Messenger users.
The vision behind the Nocturnal project is to harness the power of social
links for sharing information such as recommendations and reviews. The key idea
is to leverage the existing messaging network in order to bootstrap the system
with natural, automatic social communities. The communities are formed by a
user's buddies list, his buddies' buddies and their buddies, and so on. Each
social circle shares much in common, without requiring users to subscribe to
any new service, eliciting information about users' hobbies and social
behavior, etc.
Nocturnal automates information exchange within these social circles using the
existing messaging channels. A user Alice exchanges data with her immediate
contacts, and stores the information she pulls from them on the local disk.
Alice's contacts exchange information with their contacts, and in this way,
information from Alice propagates to her contacts' contacts, and on into the
network.
In order to address trust concerns, recommendations shared in Nocturnal include
an origin and the number of hops they transfer through. A Nocturnal user
controls how far in the network to pull information from, usually bounding it
to several hops away.
The research component of this project is complete.
Nocturnal uses the Messenger folder-sharing feature to facilitate exchanges
between a user and her Messenger contacts. Information can propagate to
the contacts' contacts, to three-hop contacts, and so on, until any hop-limit
set by the user. The information arriving at any user is stored in a local
database, along with tags indicating the distance it came from and the path. In
this way, Nocturnal builds a social information-sharing infrastructure in
a peer-to-peer manner.
Nocturnal is used for sharing URL recommendations among Messenger peers and
form a collaborative web-search tool. Collecting URL recommendations is done in
either automated mode or manual (or both): For the former, the user may allow
automated copying of her `favorites' files and/or browser histories. For
the latter, she may voluntarily click a `recommend' button on her browser.
A browser enhanced with our collaborative-search tool bar uses the local
recommendations database to filter MSN-search results and bring URLs
recommended by peers to higher importance.
- P2P Rating
By Danny Bickson, Dahlia Malkhi and
Lidong Zhou.
7th IEEE International Conference on Peer-to-Peer Computing. PDF
- Click here for PowerPoint slides that briefly describes
Nocturnal.
|
