projects
Detail
Publication date: 1 de June, 2021P2P Dissemination of Web Syndication Content
LiveFeeds focuses on Web Syndication – a concept that enhances the World Wide Web experience by helping users to keep track of the updates made to their favorite websites, blogs and web content in general.
The project will devise ways to incorporate P2P principles to form the basis of a new Web Syndication event dissemination infrastructure, which will be modeled primarily around a (true) push-based model. Additionally, the project will seek ways to enhance the Web Syndication model with support for fine-grained filtering and prioritization of syndication content in ways compatible to its large-scale dissemination goals and, in particular, without cumbering the servers with that additional effort.
LiveFeeds focuses on Web Syndication – a concept that enhances the World Wide Web experience by helping users to keep track of the updates made to their favorite websites, blogs and web content in general. Web Syndication is attractive and becoming increasingly popular, as testified by the number of top sites that already offer RSS [1] feeds. In effect, the ability to automatically aggregate and update many web feeds, summarizing the content of many different sources, makes Web Syndication a convenient and natural addition to the World Wide Web, especially since the underlying model of the latter has relied implicitly on the users to poll the web sources themselves in order to stay up-to-date.
In time, Web Syndication could very well evolve into a full fledged notification service for the World Wide Web. Although that is still somewhat speculative, an informal survey of the syndication content provided by major web content providers already shows that it is being used to disseminate popular, time-critical events, such as news alerts and sport results. Such usage is surely appealing but raises important scalability concerns, which are worsened if more aggressive requirements in liveliness (event freshness) are factored in. The root of the problem lies in the fact that Web Syndication is currently being pursued based on models [1, 2] that essentially only detail the web feed structure and, for all practical purposes, still rely on client-side polling as the basis for collecting updates. This straightforward model had the advantage of allowing a quick, incremental deployment of the service on top of the already existing server infrastructure but, in the future, it could lead to considerable strains on the communication budgets of content providers and users alike. Improving liveliness is expensive using polling techniques and multiplied by large-scale it translates to the costly problem of dealing with large numbers of users polling frequently and systematically many web feeds. This could ultimately lead to a backlash or complete abandonment of Web Syndication, reenacting a similar episode to PointCast ?s debacle [3].
There are other issues in the current approaches to Web Syndication. An important one pertains to its poor suitability when considered in conjunction with Mobility. Again, client-side polling is, generally, a poor choice for mobile environments. It is wasteful in terms of energy and bandwidth, both of which are often scarce and expensive in those situations. In mobility-driven scenarios, where disconnection and poor connectivity are the norm, the user is likely less interested in a full report of each of the web feeds he/she subscribes but, instead, would prefer to be alerted for particularly important events. In fact, it is arguable that the ability to filter web feeds and prioritize events is desirable even for the generality of users. Addressing these issues introduces interesting challenges to Web Syndication, mainly in the form of large-scale notification with filtering requirements, which is a current research topic, e.g. [4, 5].
We intend to address the problems discussed above by pursuing large-scale Web Syndication in a cooperative manner. Under such principle, the server-side model will be preserved and it will be up to the clients with similar interests to act in concert towards the desired goal. Together, they will be involved in collecting, filtering and distributing web feeds content among themselves. In order to achieve these goals, we will devise ways to incorporate P2P principles to form the basis of an alternative Web Syndication event dissemination infrastructure, which will be modeled primarily around a (true) push-based model [3]. Additionally, the project will seek ways to enhance the Web Syndication model with support for fine-grained filtering and prioritization of syndication content in ways compatible to its large-scale dissemination goals and, in particular, without cumbering the
servers with that additional effort.
We also want to explore social networking concepts to allow users with similar interests (subscriptions) to share information automatically and benefit from their experience as a collective. This opportunity is missed in the prevalent Web Syndication model, in which users act alone and interact with servers completely isolated from each other, yet it becomes feasible and natural with the move to a P2P-based infrastructure, as we advocate. However, this can be a potential source for abuse, so this aspect must be addressed in the proposed solution in a suggestive, non-intrusive spirit and include proper safeguards to avoid spam-like problems. In addition to that, it will be important to ensure that the move
from an individualistic subscription model to a distributed, cooperative one does notjeopardize other basic security requirements, such as privacy, authenticity and fair use.
[1] RSS 2.0 Specification. http://blogs.law.harvard.edu/tech/rss.
[2] M. Nottingham and R. Sayre, The Atom Syndication Format. RFC 4287, 2005.
[3] M. Franklin and S. Zdonik, Data in your face : push technology in perspective, in SIGMOD 98, 1998.
[4] I. Aekaterinidis and P. Triantafillou, Pastrystrings: A comprehensive content-based publish/subscribe dht network, in ICDCS 06, 2006.
[5] E. Anceaume, et al., A semantic overlay for self- peer-to-peer publish/subscribe, in ICDCS 06, 2006.
Funding Total | 50 |
---|---|
State | Concluded |
Startdate | 01/01/2008 |
Enddate | 30/06/2011 |