corona

CorONA: A High-Performance Publish-Subscribe System for Web MicroNews


CorONA is a high-performance publish-subscribe system for quick and efficient dissemination of web micronews. It is a replacement for, and is backwards compatible with, RSS. It uses Beehive, an analysis-driven framework for optimally trading off bandwidth for performance, to provide low-latency news dissemination while ensuring that the load placed on news providers does not exceed a desired limit.

The current RSS system causes serious load problems for news providers. Currently, every client periodically checks every news source, consuming significant bandwidth. Further, RSS demand is "sticky," that is, once subscribed, clients remain subscribed. Hence a site whose popularity temporarily surges ends up having to deal with a large constant load afterwards.

CorONA fixes the load problem by intelligently trading off resources for quick update performance. It operates as a ring of cooperative proxy servers, each of which periodically poll news providers. When a client signs up for a channel, a fraction of the servers are dedicated to polling the news providers, while the ring records the interests expressed by the client. The number of servers dedicated to checking the channel and disseminating the news is determined optimally, based on an analysis of web object popularity, size, and update rate.

Users interact with Corona entirely through instant messages:

-> How to use CorONA Instructions for using CorONA.
-> CorONA Deployment View of the internals of the CorONA ring.
Check out the channels, the nodes, and the ring.
-> Privacy Policy Our commitment to user privacy.
-> Papers Papers and Talks on Corona, Beehive and related projects.
-> People Project members.
-> Support Project support and funding.

Beehive

Computer Science Department
Cornell University