The Renaissance of Decentralized Systems 

Peter Druschel, Max Planck Institute fo Software Systems

Decentralized systems are symmetric distributed systems, where each participating node can in principle assume any of the required tasks. Such systems have enjoyed tremendous interest in recent years, sparked by the fact that decentralization is an enabler for new paradigms like cooperative computing (e.g., peer-to-peer systems), ad hoc networking applications, and highly resilient network services.

These application domains present formidable new challenges to decentralized system design. Practical systems must be self-organizing and extend to planetary scale. They must tolerate heterogeneity, dynamic membership and mobility. And, they must provide provable guarantees and consistency despite complex, correlated failures, mutual suspicion, lack of trusted authorities, and selfish or malicious participants.

In this talk, I will outline the state-of-the-art in modern decentralized system design, sketch existing and future applications and identify key challenges that remain.

About the speaker:

Peter Druschel is scientific director at the Max Planck Institute for Software Systems in Kaiserslautern and Saarbruecken, Germany.

Previously, he was a Professor of Computer Science and Electrical and Computer Engineering at Rice University in Houston, Texas.  He received the Dipl-Ing. (FH) in Data Systems Engineering from Fachhochschule Munich, Germany in 1986 and the M.S. and Ph.D. degrees in Computer Science from the University of Arizona in 1990 and 1994, respectively. 

His research interests include distributed systems and operating systems. Peter's current work centers around resilient, large-scale, decentralized services and applications in cooperative and ad hoc environments.  He is the recipient of an NSF CAREER award and an Alfred P. Sloan Fellowship.

Data Management for Mobile Services---Location Tracking and Geo-Content Modelling

Christian S. Jensen, Department of Computer Science, Aalborg University (

We are reaching the unique point in history where a global computing and information infrastructure will emerge that supports the delivery of location-enabled services to mobile users always and everywhere, including services that rely on the tracking of the continuously changing positions of entire populations of mobile-service users. Qualitatively new services will become available that either make little sense or are of limited interest in the traditional context of fixed-location, desktop-based computing.

Following a coverage of underlying drivers and example applications, this talk describes an approach to tracking that offer accuracy guarantees for the positions of the moving objects as stored on the server side. In this approach, the server and each mobile client share a representation of the client's changing position, and each client issues an update to the server when its position as given by this shared representation deviates from its GPS position by more than a certain threshold.

A key issue is how to represent the changing position of a moving object so that tracking can be done with as few updates as possible. A range of techniques are described that utilize progressively more information about an object's movement, ranging from assuming that an object remains in its most recently reported position to assuming that past movement trajectories of an object through a road network are known. The talk reports on experimental evaluations of the tracking techniques based on real GPS data and a corresponding real road network.

Users typically have small screens and no standard keyboard, and they are engaged in some primary activity such as driving safely. For these and other reasons, it is important to deliver the ``right'' service at the right time, with minimal user interaction. This may be achieved by making services context aware. A route-aware service may suggest points of interest near the user's route, rather than merely to the user's current location. This talk describes key techniques underlying
a software component that builds routes for individual users, with associated usage data, based on traces of GPS coordinates, so that the likely route(s) for a user can be identified and used in mobile services without user interaction.
About the speaker:

Christian S. Jensen, Ph.D., Dr.Techn., is a Professor of Computer Science at Aalborg University, Denmark, an Honorary Professor at Cardiff University, UK, and an adjunct Professor at Agder University College Norway.

His research concerns data management and spans issues of semantics, modeling, and performance. With his colleagues, he receives substantial national and international funding for his research, and he has authored or coauthored more than 150 scientific papers. He is a member of the Danish Academy of Technical Sciences, the EDBT Endowment, and the VLDB Endowment's Board of Trustees. He received IbHenriksens Research Award 2001 for his research in mainly temporal data management and Telenor's Nordic Research Award 2002 for his research in mobile services.

His service record includes the editorial boards of ACM TODS, IEEE TKDE and the IEEE Data Engineering Bulletin. He was the general chair of the 1995 International Workshop on Temporal Databases and a vice program committee chair for ICDE 1998. He was program committee chair or co-chair fore Workshop on Spatio-Temporal Database Management, held with VLDB 1999, for SSTD 2001, EDBT 2002, and VLDB 2005. He serves on the boards of directors and advisors for a small number of companies, and he serves regularly as a consultant.


Applying components / frameworks to the ObjectWeb Persistence Support

Pascal Déchamboux, Senior architect, France Télécom R&D, Grenoble  

(abstract not available).

About the speaker: 

Pascal Déchamboux , Ph.D., is a senior architect at the research centre of France Telecom.

His research work concerns middleware support for telecommunication service delivery environment. He has been the manager for several years of research team that focuses on two main middleware topics:
- work on component support that lead to the proposition of the Fractal component model and its Julia implementation, complemented with management capabilities.
- work on persistence support that lead to the proposition of several frameworks such as mapping subsystem managing typed I/Os (e.g., O/R mapping), transactional data manager, cache manager or concurrency manager, for building persistence supports complying with different standards like EJB or JDO.
He has been involved in many national and international collaborative projects, especially with Bull and INRIA. He is one of the co-founders of the ObjectWeb open source community within which results of the previously mentioned works can be found and used with no restriction!