Home Contact

DIP3 Project

Computer Science Department, University of Ioannina

College of Computing, Georgia Institute of Technology


FP7 People IOF Action
The 3Ps of Distributed Information Delivery: Preferences, Privacy and Performance


Research in DIP3 aimed at advancing the state-of-the-art in the following ways.

  • Whereas traditional pub/sub systems rely on a binary, match/no-match model for sending relevant data to users [EFGK03], DIP3 proposes non binary matching, where items are assigned degrees of relevance. This is a novel idea. We are aware of only two very recent approaches, namely [MVGS08] and [PPZA08], that go beyond binary matching. [MVGS08] considers the reverse problem to the one in DIP3: sending a publication to its most relevant subscribers. Our approach is user-centric: we are interested in sending the most relevant publications to each subscriber. [PPZA08] assumes that publications are already ordered, thus avoiding tackling the difficult problem that DIP3 addresses, that is, determining relevance per user. Distributed ranking of data delivery in this setting is especially challenging due to the level of distribution, the huge number of data and users and especially the potential diversity among the preferences of the users. This is expected to create a family of new research problems and solutions.
  • DIP3 has also explore different ways of ranking data including skylines, top-k and iceberg queries, as well as their variations and possible combinations. This extends the filtering (match/no match) functionality of traditional pub/sub systems.
  • Although there is extensive work on preference models (for example, [AW00, K02, C03]), preference models and algorithms for push data delivery have not been addressed yet. To this end, DIP3 will explore both quantitative and qualitative models for expressing preferences. Another novel feature of DIP3 is exploiting social networks for enhancing pub/sub systems. There is increased interest in the Web research community on exploiting the rich information created by Web 2.0 application to extend Web search and recommendation systems (for example [NXW+07, HGG08, DDGR07]). DIP3 utilizes such information to move from individual user preferences to social preferences.
  • DIP3 has addressed the interplay between system quality (as achieved through ranking) and system performance (especially in terms of response time). Enforcing performance guarantees in this setting is challenging mainly because of the large scale of the system, its diversity and dynamicity. DIP3 will exploit caching and replication at different system levels and at various granularities. Caching and replication are well studied; however, this new setting introduces new requirements. DIP3 will also develop new data structures and algorithms for clustering and indexing subscriptions and preferences that will take into account ranking and preferences.
  • Finally, although there is a lot of research on privacy, privacy-preserving push-based delivery has not been explicitly addressed by previous research. We expect that the output of DIP3 will include a set of new personalized privacy models and mechanisms.


[EFGK03] P. Th. Eugster, P. Felber, R. Guerraoui, A-M. Kermarrec: .The Many Faces of Publish/Subscribe.. ACM Computing Surveys 35(2), 2003

[MVGS08] A. Machanavajjhala, E. Vee, M. Garofalakis, J. Shanmugasundaram: .Scalable Ranked Publish/Subscribe., 34th Very Large Database Conference, (VLDB08), September 2008, To appear.

[PPZA08] K. Pripuzic, I. Podnar Zarko, K. Aberer: .Top-k/w Publish/Subscribe: Finding k Most Relevant Publications in Sliding Time Window w.. 2nd Intern. Conference on Distributed Event-Based Systems (DEBS 2008), June 2008

[AW00] R. Agrawal, E. L. Wimmers: .A Framework for Expressing and Combining Preferences.. SIGMOD 2000

[K02] W.Kie.ling: .Foundations of Preferences in Database Systems.. VLDB 2002

[C03] J.Chomicki: .Preference Formulas in Relational Queries.. ACM Transaction on Database Systems 28(4), 2003

[NXW+07] S. Bao, G-R. Xue, X. Wu, Y. Yu, B.Fei, Z. Su: .Optimizing Web Search using Social Annotations.. WWW 2007

[HGG08] P.Heymann, G. Koutrika, H.Garcia-Molina: .Can Social Bookmarking Improve Web Search? WSDM 2008

[DDGR07] A. Das, M. Datar, A. Garg, S. Rajaram: .Google News Personalization: Scalable Online Collaborative Filtering.. WWW 2007

Valid CSS! Valid XHTML 1.0 Strict