Workshop Program
| Start |
End |
Monday, December 1st, 2008 |
| 09:00 |
09:15 |
Welcome / Workshop Opening |
| 09:15 |
10:30 |
Keynote Speech 1 : Joe Polastre, Sentilla
The Role of Middleware in Wireless Sensor Networks
|
| 10:30 |
11:00 |
Coffee break |
| 11:00 |
12:30 |
Session 1: Context and Management (session chair: Danny Hughes) |
| 11:00 |
11:30 |
A Rule-based Acceleration Data Processing Engine for Small Sensor Node (Kenji Kodama, Naotaka Fujita, Yutaka Yanagisawa, Masahiko Tsukamoto and Tsutomu Terada) |
| 11:30 |
12:00 |
A Self-Adaptive Context Processing Framework for Wireless Sensor Networks (Amirhosein Taherkordi, Romain Rouvoy, Hai Ngoc Pham and Frank Eliassen) |
| 12:00 |
12:30 |
Towards policy-based management of sensor networks (Nelson Matthys and Wouter Joosen) |
| 12:30 |
14:00 |
Lunch |
| 14:00 |
14:45 |
Keynote Speech 2: Jacques De Kegel, IBM
Middleware: The Engine That Makes Sensor Applications Run Smoothly
|
| 14:45 |
15:30 |
Workshop Panel (Topic: Sensor Networks in Business - Research Challenges and Opportunities) |
| 15:30 |
16:00 |
Coffee break |
| 16:00 |
17:00 |
Session 2A: Communication and Security (session chair: Christophe Huygens) |
| 16:00 |
16:30 |
Comprehensible Security Synthesis for Wireless Sensor Networks (Stefan Ransom, Dennis Pfisterer and Stefan Fischer) |
| 16:30 |
17:00 |
Generic Multi-Packet Communication through Object Serialization (Leon Evers, Maria Lijding and Jan Kuper) |
| 17:00 |
17:30 |
Discussion / Workshop Closing |
Keynote Speaker Bios
Joe Polastre, Sentilla
Joe Polastre is Co-founder and Chief Technology Officer of Sentilla Corporation. He is credited with leading the wireless sensor networking and pervasive computing industries. Dr. Polastre deployed the world's first pervasive networked applications, one of which analyzed elusive seabirds on Great Duck Island. He evangelizes his vision of pervasive computing -- where everyday objects around us become smart and inter-networked, thereby making the world greener, safer, more efficient, and more fun. Providing everything necessary to interact and analyze the environment, Sentilla's software and solutions provide actionable intelligence that revolutionize business processes in the energy and transportation industries. Prior to Sentilla and Berkeley, Joe held engineering and management positions at IBM, Intel, Microsoft, and a few startups. Dr. Polastre has a Ph.D. from the University of California, Berkeley and B.S. from Cornell University.
Jacques De Kegel, IBM Benelux
Jacques De Kegel received his engineering degree at the K.U.Leuven in 1978 and is currently active at IBM Benelux as Business Development Executive.
His career at IBM started in 1983; since 2002 he is responsible for Emerging Business Opportunities (EBO), first in Belgium and Luxemburg and since 2005 in the Benelux. He leads a team that focuses on Sensor and Actuator Solutions. Jacques is certified Telecommunications Industry Solutions Sales Specialist and was involved in numerous innovative projects at Benelux and European level.
Keynote Speech Abstracts
The Role of Middleware in Wireless Sensor Networks, Joe Polastre, Sentilla
The emergence of inter-networked embedded computers has ignited a new revolution in computing with applications that make "the invisible visible" by analyzing real world conditions directly at the source. While early solutions were built as a stove-pipe that integrated all layers of the software stack, the real value of pervasive systems is the ability to change functionality, upgrade, and run new applications directly in the field. Many efforts in the research community have focused on developing new middleware, with new programming languages, APIs, communication paradigms, and operating systems. Instead, this talk shows a different approach -- clever adaptation of existing standards, familiar programming languages, and proven operating systems to meet the size and performance constraints of pervasive computers. Numerous industry applications are now deployed, in large part due to the time and cost benefits of standardized, familiar middleware technologies.
Middleware: The Engine That Makes Sensor Applications Run Smoothly, Jacques De Kegel, IBM Benelux
IBM is one of the major developers of general purpose middleware. In particular for sensor based applications, IBM has a long tradition in offering its customers the best performing, scalable and most robust platforms possible. If needed, IBM not only provides the technology but also implements and operates on-demand solutions on behalf of the customer. By consequence, a wide range of options is available to implement, configure and integrate sensor applications. Jacques De Kegel will provide an overview of IBM's expertise in this field and highlight customer solutions they offer to build sustainable and resilient solutions within an SOA environment.
Paper Abstracts
A Rule-based Acceleration Data Processing Engine for Small Sensor Node.
Kenji Kodama, Naotaka Fujita, Yutaka Yanagisawa, Masahiko Tsukamoto and Tsutomu Terada
In recent years, various small sensor nodes have been developed to recognize situations and events occurring in the real world for the development of context-aware application systems. Many systems apply sensor data to provide pervasive computing services. We consider acceleration sensors one crucial element on a node because their data include rich and simple information to accurately recognize various types of situations such as motions, directions, and the positions of objects. An application system using acceleration data requires the following: 1) a node that can rapidly process data without a large amount of memory since the amount of acceleration data is much greater than the amount of other sensor data; 2) a node that can reduce the amount of data sent to a server; and 3) systems that can be easily configured by users with low cost. Existing sensor nodes, however, do not have enough functions to satisfy these requirements. In this paper, we propose a rule-based data processing system specified to rapidly process acceleration data. Our proposed system can rewrite the rules on each node with a few bytes of data. To evaluate our proposed mechanism, we experimented with our rulebased system implemented on our small sensor node called MoCoMi-Chip.
A Self-Adaptive Context Processing Framework for Wireless Sensor Networks.
Amirhosein Taherkordi, Romain Rouvoy, Hai Ngoc Pham and Frank Eliassen
Wireless sensor networks are increasingly been exploited in the ubiquitous computing environment as an important platform for gathering the context data. To collect continuously the environment context information for a long time, the sensor node itself should be considered as a context-aware device having its own particular context parameters, such as residual energy or sample rate. Existing work in the field of context-aware computing considers sensor node as a context data collector agent, regardless to the concern of node’s context elements. In this paper, first we propose an approach for modeling sensor network context information, and then a middleware framework is introduced for mapping the context model to software components, processing the context data, and executing the context model. For this purpose, we propose the notion of context node, which is the building block for context execution. The proposed solution is exemplified in the shape of a home monitoring application. Using the proposed framework, sensor application can adapt itself to the current situation in the environment through executing a high-level context model describing the context information and the respective adaptation actions.
Towards policy-based management of sensor networks.
Nelson Matthys and Wouter Joosen
Contemporary distributed software systems for realistic business applications have become extremely heterogeneous, dynamic and large scale. They offer services to many different types of users and include various hardware such as back-end servers, regular PCs and various mobile and embedded devices, as well as diverse network infrastructures, such as sensor networks. However, the management of these systems is typically a complicated task. This paper proposes a policy-based approach that offers a high-level of abstraction for managing these distributed applications. We first give an overview of existing research on policies in traditional middleware and describe some of the shortcomings when applying it for managing sensor networks. Secondly, we propose an architecture for a middleware that addresses these limitations and provides a solution for them.
Comprehensible Security Synthesis for Wireless Sensor Networks.
Stefan Ransom, Dennis Pfisterer and Stefan Fischer
Providing usable security mechanisms for Wireless Sensor Networks (WSNs) is one of the important tasks to foster WSN development, as they will increasingly be deployed in real-world settings in the future. Generally, due to its complexity, expert knowledge in the field of security is needed to plan, implement, and deploy a sound security setup. Frequently, no or only basic security mechanisms leave the WSN open to attacks of which the application designer might not be aware, especially if no security evaluation has been done. Furthermore, available security mechanisms have a great impact on the design of the application itself, i.e. to properly secure a protocol. We therefore propose a framework that first devises feasible security setups based on the characteristics of the envisioned application and second provides a security evaluation of theses possibilities from which the application designer is able to choose the fitting setup for his application. We have prototypically implemented the proposed scheme into the WSN middleware synthesis tool FABRIC and thus support the application designer in including a sound security solution into his application.
Generic Multi-Packet Communication through Object Serialization.
Leon Evers, Maria Lijding and Jan Kuper
To this date, sensor networks communication protocols and abstractions deal only with payloads that fit in a single packet. Although this may be a convenient way to build simple communication protocols, it does not always reflect the application requirements and may not be the most energy--efficient way to communicate. Hence the need for a communication mechanism that enables splitting payload across multiple packets. This paper presents a communication abstraction that enables multi-packet communication, while minimizing memory requirements by using an object serialization mechanism to integrate memory management and communication functionalities. Evaluation shows reduction of communication of up to 3.5 times using our method compared to state of the art, and improvement of application performance through the use of reliable communication.
|
