Small logo of ETH main building ETH Zurich : Computer Science : Pervasive Computing : Distributed Systems : Research : Research Activities

Distributed Systems Group

Ubiquitous Computing Infrastructures

Prof. Friedemann Mattern

[Also available as a PDF document. For a more detailed list of activities, visit our Research Topics overview.]

The Distributed Systems Group conducts research in distributed algorithms, Internet applications, middleware systems, and concepts to enhance reliability, security, and privacy in distributed systems.

Of particular interest is the field of ubiquitous computing, which aims at making computers available throughout the environment, while rendering them effectively invisible. One of the main goals is to incorporate computing power into everyday objects in order to create "smart things" real-world objects that provide novel ways of accessing information, react to their environment, or provide new emergent functionality when interacting with other smart things.

This vision of ubiquitous computing is grounded in the belief that microprocessors and advanced sensors will soon become so small and inexpensive that they can be embedded in almost everything. It is expected that billions of such objects will be interwoven and connected together by wireless networks, forming a world-wide distributed system several orders of magnitude larger than today's Internet.

To enable communication and cooperation among smart objects, new information infrastructures are required. They have to cope with a highly dynamic environment and should, among other things, provide location information to mobile objects, represent context information, and enable reliable and scalable service creation. The Distributed Systems Group addresses the challenges of designing and implementing such infrastructures.

Middleware

Figure 1: MICO is a widely used open source CORBA implementation.

Application development for distributed systems relies on software frameworks and middleware systems that provide high-er-level abstractions. Based on our experience with the open source CORBA implementation MICO, we are exploring novel middleware approaches for dynamic networks of resource-constrained mobile computing devices. In particular, we are working on middleware concepts for sensor networks and smart cards, and on frameworks for remote identification of real-world objects.

Smart Devices

Figure 2: The Smart-Its project was one of 16 projects conducted under the European Union's Disappearing Computer initiative, whose aim was to explore how everyday life can be supported and enhanced through the use of collections of interacting artifacts.

We participated in the European Smart-Its project whose goal was to develop unobtrusive, deeply interconnected smart devices that can be attached to everyday items in order to support new functionality, novel interaction patterns, and intelligent collaborative behavior. Eventually, Smart-Its should be as cheap and as small as state-of-the-art radio tags (RFIDs), but in addition they will also be able to communicate with peers, and they will be customizable in their behavior. In order to facilitate a meaningful integration in their environment, Smart-Its are equipped with various sensors providing context information.


Building on our experiences in the Smart-Its project, we are participating in the MICS (Mobile Information and Communication Systems) National Center of Competence in Research (NCCR). The Center studies fundamental and applied problems that are raised by new types of mobile communication and information services and devices. Together with other research groups at ETH Zurich and EPFL we are working on software architectures and infrastructures for mobile networks of smart, location-aware devices. The BTnode platform, consisting of a microprocessor, a Bluetooth radio, and various sensors is used as the basis of several prototypes built within the scope of this project.

Figure 3: The second generation of Smart-Its prototypes, called BTnodes, are based on an Atmel microcontroller with 128 kB of in-system programmable flash memory and 64 kB of SRAM. Bluetooth modules allow communication between different devices.

Sensor Networks

A sensor network consist of a large number of tiny autonomous computing devices, each equipped with sensors, a wireless radio, and a processor. Sensor networks are envisioned to be deployed unobtrusively in the physical environment in order to monitor a wide range of environmental phenomena with unprecedented quality and scale. The close integration of sensor networks with the physical world imposes a number of challenging research problems.

Our group is working on basic issues such as energy-efficient and scalable time synchronization and localization of sensor nodes, as well as on higher-level mechanisms for integrating these basic services into a common framework for the development of applications for sensor networks. For practical experiments we use collections of cooperating BTnodes.

Implications of Ubiquitous Computing

Figure 4: A typical RFID tag has the size and flexibility of an address label and allows remote reading and writing from up to several meters without requiring line of sight. The rising use of RFID tags in both industry and retail has led to a number of controversies over the impact of ubiquitous computing on society.

Ubiquitous computing technologies will have a strong impact on future society. In a project funded by the Gottlieb Daimler and Karl Benz Foundation, an interdisciplinary group of researchers explores its social, economic, and legal implications. One of the main tasks is to communicate these findings to relevant groups in society and politics. Within the project we are also developing scenarios and building prototypes, demonstrating new types of human-computer interaction.

Security and privacy will be of prime concern in a world of highly interconnected, invisible devices that will eventually permeate our everyday lives. We are working on a privacy-aware infrastructure (based on P3P, a W3C Web standard for exchanging privacy policies) that is able to keep track of any ongoing data exchange, while providing the user with powerful means to selectively manage such collected personal information. Also, with the growing number of interacting devices and smart appliances, the probability of failure for any single device increases. We are therefore investigating the role of redundancy and fault-tolerance in ubiquitous computing, with the goal to provide concepts for reliable infrastructures and services.

Business Applications

The M-Lab, a joint venture with the University St. Gallen, was created to bring the ideas of ubiquitous computing into companies that could profit from the availability of smart devices, radio frequency identification, and related technologies. We are focusing on areas such as logistics, retail, life sciences, and automotive. The M-Lab brings together researchers from academia and practitioners from industry, contributing to the research area and creating competitive advantage.

Teaching

Besides teaching fundamentals of distributed systems, distributed algorithms, and Internet technology, our graduate level curriculum focuses on state-of-the-art research in ubiquitous computing. Lectures are accompanied by hands-on programming experience. In-depth seminars allow students to individually research and present fundamental works in the field, as well as investigate recent developments. In our Ubicomp Lab, students can devise and build their own smart environments using handheld devices, ID systems (e.g., smart cards or RFID tags), and wireless communication equipment such as Bluetooth or WLAN.

More information on the Distributed Systems Group is available at http://www.vs.inf.ethz.ch/

Friedemann Mattern studied computer science at the University of Bonn and became a Faculty Research Assistant at the University of Kaiserslautern in 1983. He obtained his Ph.D. in 1989 with a thesis on distributed algorithms. From 1991-1994 he was Professor of practical computer science at Saarland University, Saarbrucken, and from 1994-1999 Professor of practical computer science and distributed systems at Darmstadt University. He was appointed full Professor at the ETH Zurich in July 1999.

ETH ZurichDistributed Systems Group
Last updated August 14 2009 05:40:50 PM MET webvs