Syndicated content in the Internet has been a huge success ever since the early days of RSS 0.9 and MyNetscape. Currently, it is the cornerstone of content push, ranging from podcasts to emerging Web 2.0 sites such as FriendFeed and Plexus. Unfortunately, the baffling simple technology that makes publication and subscription very simple and flexible, thus explaining in part its success, is also limiting its usefulness in more demanding applications.

Current protocols, like RSS and Atom, allow anyone to easily publish or subscribe to feeds by using simple tools. These protocols prove quite adequate for the large majority of feeds, featuring a few subscribers and low data throughput. However, for high throughput users or highly popular feeds, they're less adequate, as the polling mechanism is expensive for the publisher.

seeds is...

an approach to feed caching and dissemination that leverages social networks in order to perform aggregation of interests and also to promote discovery. The actual feed distribution is done using peer-to-peer protocols.

Project goals

Multiple Dissemination Protocols

Feed content type and traffic patterns vary widely depending on the source. We can have some mainly composed of text, normally for website related news, or feeds with richer content as are podcasts (feeds with audio/video attached). It makes sense not to use the same dissemination protocol for such different content, so we should be able to differentiate between them and choose the right protocol. For example, small text based entries with a very high update rate could be disseminated via NeEM and use BitTorrent for audio and video content.

No Change to the Source

One of the reasons other proposals haven't been met with great success (e.g. FeedTree ) was relying heavily on the source for being able to join the system. In our proposal, we should be able to disseminate feeds within the architecture used today, by leveraging social relationships among subscribers.

No Change to the Target

Clients shouldn't need to change their habits, removing this way, one of the biggest barriers for the adoption of our proposal. The client is also given the chance to customize the subscription process.

Enhanced Discovery and Aggregation

By leveraging social networks we enhance the subscribers' experience with recommendations for new feeds. These can be content related or stem from social connections.