G-Lab (Real-World G-Lab)

Dauer:   01.09.2009 - 31.08.2012
Projektleiter:  Prof. Dr. Horst Hellbrück
Mitarbeiter: Torsten Teubler

Hintergrund

Das heutige Internet besitzt eine große wirtschaftliche Bedeutung, basiert jedoch größtenteils noch auf Mechanismen und Algorithmen, die in den 70er und 80er Jahren entwickelt wurden. Zukünftige Anwendungen im geschäftlichen und privaten Bereich stellen Anforderungen, für die das Internet ursprünglich nicht konzipiert war. Diese Anwendungen sehen sich Problemen gegenüber, die auf Defizite in der Architektur des Internet hinweisen.

Eine zentrale Anforderung des zukünftigen Internets wird die Einbindung diverser Netzwerktechnologien zur Unterstützung unterschiedlichster Anwendungen sein. Eine wesentliche Rolle wird dabei die Erfassung von Daten der realen Welt spielen, um beispielsweise Informationen über bestimmte Umweltphänomene zu sammeln oder um einem Benutzer aktuelle Kontextinformationen zu vermitteln. Ein wesentliches Instrument dieses zukünftigen Internets werden daher drahtlose Sensor- und Mesh-Netze sein, mit denen die Daten erfasst und verfügbar gemacht werden.

Ziel

In diesem Projekt wurde untersucht, wie sich drahtlose Sensor- und Mesh-Netze in zukünftige Internet-Architekturen einpassen lassen. Die Fachhochschule Lübeck hat sich hier besonders die Entwicklung neuer Anwendungen und die Unterstützung von Mobilität zum Ziel gesetzt. Es wurden unterschiedliche Aspekte von grundlegender technischer Einbindung über Kommunikations- und Service-Infrastrukturen bis hin zu Anwendungsproblemen betrachtet.

Viele Forschergruppen verfolgen den Ansatz, leichtgewichtige Protokolle und Anwendungen direkt auf den Sensorknoten zu implementieren und die Sensornetze so direkt ins Internet zu integrieren. Die physikalische Anbindung erfolgt über eine Art „Gateway“, das IP Pakete aus dem Internet in 6LoWPAN (IPv6 über Funkstandard IEEE 802.15.4) Pakete wandelt und über die Funkschnittstelle in das Sensornetz leitet und umgekehrt.

Wir hingegen verfolgten mit EZgate den Ansatz der Anbindung und erweitern die Aufgaben des Gateways. Es dient nun als Anwendungs-Proxy. Das heißt, dass Anwendungen auf dem Gateway ausgeführt werden. Wir sehen in diesem Vorgehen eine Vereinfachung der Anwendungsentwicklung und Wartbarkeit im laufenden Betrieb der Sensornetze.

Weiterhin wurde im Rahmen des Projekts wurde eine Experimentierplattform (Testbed) aufgebaut, die aus verschiedenen Sensor- und Mesh-Netzen besteht. So können andere Forschungseinrichtungen und auch Firmen eigene Algorithmen für solche Netze entwickeln und unter realistischen Bedingungen testen. Dieses Sensornetz Testbed wurde in die bereits existierende G-Lab (German-Lab, nationales Testbed für Internet-Algorithmen) Testbed-Infrastruktur integriert. Real-World G-Lab stellt somit eine Erweiterung von G-Lab dar.

Schematische Schichten-Darstellung des EZgate Stapels zwischen Internet und Sensornetz und zwei fiktive Anwendungen oberhalb des Gateways

Schematische Schichten-Darstellung des EZgate Stapels zwischen Internet und Sensornetz und zwei fiktive Anwendungen oberhalb des Gateways

Stationäres und Mobiles Testbed

Das mobile Testbed an der Fachhochschule ist ein Indoor-Netz, die Infrastrukturknoten des mobilen Testbeds an der Fachhochschule Lübeck sind permanent im Testbed Verbund verfügbar. Die mobilen Knoten können nach Absprache verfügbar gemacht werden. Als mobile Plattform für das Testbed an der Fachhochschule Lübeck dienen speziell modifizierte Roomba Roboter. Durch die vorgenommenen Modifikationen können die Roboter autonom gesteuert und über WLAN administriert werden. Neben den mobilen Sensorknoten (die Geräte, die das Sensornetz bilden) gibt es noch fest installierte Sensorknoten im Testbed.

Übersicht über die fest installierte Testbed-Sensorknoten im Gebäude 18 der FH Lübeck
Fest installierte Testbed-Sensorknoten im Gebäude 18 der FH Lübeck

Verwertung der Ergebnisse

Das im Rahmen des Projektes aufgebaute Testbed wird weiterhin aktiv genutzt. So wird es im Rahmen des Projektes DataCast, wo datenzentrische Ansätze in drahtlosen Sensornetzen untersucht werden, eingesetzt, um Protokollimplementierungen zu testen.

Das Projekt SoCoR (Strategies for cooperative spectrum sensing in Cognitive Radio networks) welches ebenfalls am CoSA Kompetenzzentrum beheimatet ist, nutzt das Testbed regelmäßig für Experimente und Messungen von kooperativen Systemen. Dies zeigt die Vielseitigkeit unserer Installation und Ergebnisse.

Weiterhin werden Implementierungen von Protokollen, die im Rahmen von Abschlussarbeiten entstehen, im Testbed evaluiert. Ein Beispiel für eine Arbeit, die zur Produktreife weiterentwickelt wurde und sich seit Dezember 2012 im produktiven Einsatz in der Industrie befindet ist „Entwurf, Implementierung und Bewertung eines sensornetzbasierten Messsystems für Flächenneigungen“. Das System wurde unter Zuhilfenahme der TriSOS Sensorknoten und der generischen Algorithmenbibliothek Wiselib entwickelt. Die Sensorknoten werden mit EZgate im IP basierten LAN verfügbar gemacht.

Durch die konsequente Verwendung der generischen Algorithmenbibliothek Wiselib können Anwendungen unabhängig von der Sensorknotenhardware entwickelt werden. Da die Wiselib auch auf die Sensornetzplattformen der Projektpartner (iSense) portiert worden ist, lassen sich Algorithmen, die in diesem Projekt implementiert worden sind, auf allen Plattformen evaluieren und weiterverwenden. Durch die Verwendung der Testbed Föderationssoftware Wisebed lassen sich die Testbed-Installationen auch auf Experimentebene in einem Verbund betreiben, um so die Ausdehnung des Testbeds zu erweitern.

Veröffentlichungen

Folgende Veröffentlichungen der Ergebnisse sind im Rahmen des Projektes erfolgt:


Artikel and Buchkapitel
[2012] Transparent Integration of Non-IP WSN into IP Based Networks (Torsten Teubler, Mohamed A. Hail, Horst Hellbrück), In International Conference on Distributed Computing in Sensor Systems and Workshops IEEE Computer Society, 2012. [bib] [ppt] [abstract]
Embedded devices connected to the Internet will start an increasing growth of the Internet in near future. Wireless Sensor Networks (WSN) will play a major role in that growth. In the past several solutions were proposed to make sensor networks IP capable. Today there are IPv6-Stacks available including web servers running on sensor nodes. However, a gateway is always needed to convert the routing protocols and MAC-Layer Protocols including compression of IP packets to run on these devices. The overhead using IPv6 on the nodes is very high in respect of code size and message overhead. Therefore, in our approach we design and implement a system based on simple protocols target for sensor network nodes and a flexible gateway working in a hybrid fashion for our sensor network testbed. We successfully integrated this non-IP WSN in the Internet and our testbed is productive available from any computer connected to the Internet for reference. In this paper we present the architecture of our solution and present the implementation details of a standard WSN application that can be used for evaluation.
[2010] Cooperative Virtual Memory for Sensor Nodes (Torsten Teubler, Jan Pinkowski, Horst Hellbrück), Chapter in Real-World Wireless Sensor Networks Springer Berlin / Heidelberg (Pedro Marron, Thiemo Voigt, Peter Corke, Luca Mottola, eds.), volume 6511, 2010. [bib] [pdf] [abstract]
Wireless sensor networks (WSN) have unique challenges and constraints. Sensor nodes e.g. have tough memory limitations. However, the latest advances in WSN research direct for an implementation of lightweight versions of Internet protocols like IPv6, TCP, and HTTP on sensor nodes. These protocols have challenging requirements. Especially, memory consumption of these protocols is often higher than the physical RAM that microcontrollers have integrated. Therefore, we suggest an approach for virtual memory providing more memory than the available RAM. As microcontrollers do not include a memory management unit the usage of memory is implemented in cooperative fashion based on the C standard library function malloc and free. We suggest an underlying file system and a hardware abstraction layer to support various external or internal memory devices like Flash or EEPROM. In this work in progress we present an API, some implementation details and preliminary results including future work.
Konferenz Beiträge
[2013] Wiseman - A Management and Deployment Approach for WSN Testbed Software (Torsten Teubler, Horst Hellbrück), In 2013 IEEE INFOCOM Student Poster Session (INFOCOM'2013 Student Posters), 2013. [bib] [abstract]
Wireless Sensor Networks (WSNs) are an emerging technology. Today research in this field focuses on WSN testbeds to evaluate algorithms under realistic conditions. Numerous WSN testbed platforms allow remote deployment of WSN code and control of WSN experiments. However, one major aspect for testbeds was not addressed until now, namely the deployment and management of the testbed software itself. By deployment we mean installation and configuration of software. Once deployed on the testbed machines executables or services need maintenance and management. During testbed lifetime, periodic redeployment of testbed software is necessary due to new software versions, configuration changes, or an extension of the testbed. In this work, we present Wiseman a management and deployment approach and an implementation of Wiseman for the Wisebed WSN testbed software.
[2013] CCN-WSN - a lightweight, flexible Content-Centric Networking Protocol for Wireless Sensor Networks (Zhong Ren, Mohamed A. Hail, Horst Hellbrück), In 2013 IEEE Eighth International Conference on Intelligent Sensors, Sensor Networks and Information Processing (IEEE ISSNIP 2013), 2013. [bib] [abstract]
In future Internet research, content centric networking (CCN) is a new promising approach. CCNx has been introduced recently as an open source protocol suite for CCN and implementation base for practical research. In wireless sensor networks (WSNs) research, data or content centric approaches like in-network processing and data aggregation are important. While the principle of CCN is a suitable approach in WSNs, the CCNx protocol suite designed for PCs is not applicable to resource-constrained WSNs. Additionally, gateways necessary between CCNx and WSN are difficult to implement. Therefore, we design, implement and evaluate a lightweight variant of a CCN protocol specifically for WSNs called CCN-WSN. Key concepts of CCNx protocol are integrated but a variety of aspects are revised to meet the memory and computational constraints of sensor nodes and communication patterns in WSNs. E.g. the message format is simplified and some fields are omitted completely. Instead, we propose a flexible naming strategy which extends the functionality of content names to add small amount of data in interest messages. For performance evaluation a challenging time-synchronization application was implemented with CCN-WSN to demonstrate the flexibility of the approach and a comparison with a reference protocol for data dissemination called AutoCast is presented.
[2011] API for Data Dissemination Protocols - Evaluation with AutoCast (Sebastian Ebers, Mohamed A. Hail, Stefan Fischer, Horst Hellbrück), In The Third World Congress on Nature and Biologically Inspired Computing (NaBIC 2011) IEEE, 2011. [bib] [pdf] [abstract]
In the past various protocols inspired by nature and biology have been proposed to disseminate or transfer data in mobile or static ad-hoc networks. Many of them are designed for usage in wireless sensor networks or vehicular ad-hoc networks. Recently, we have developed and designed a general purpose data dissemination protocol called AutoCast in this field that we evaluated in detail by simulations. When we started to use AutoCast in real applications, we found out that the description of AutoCast is incomplete, as we provided the algorithms of AutoCast in details but did neither provide nor describe a suitable Application Programming Interface (API) and AutoCast was closely coupled to the application. The focus of this article is twofold. First, we propose an appropriate API to encapsulate data dissemination protocols like AutoCast and we specify the service interface of AutoCast in detail. This API can serve as a reference model for other nature and biologically inspired data dissemination approaches and applications. Second, we evaluate two applications based on our API with AutoCast in the field of wireless sensor networks and vehicular ad-hoc networks to illustrate the usage of the API and demonstrate the flexibility of this approach.
[2011] GAAP - Generic Android Application Programming (Horst Hellbrück, Philipp Krummenauer, Torsten Teubler), In Proceedings of WWW/Internet 2011 (Bebo White, Pedro Isaías, Flávia Maria Santoro, eds.), 2011. [bib] [pdf] [ppt] [abstract]
Today, smartphones are one of the fastest growing markets in the world where Android is receiving more and more attention. In the beginning when users of smartphones use Internet applications with generic WWW-Browsers recently there is a trend that they are been replaced by special applications. For each service in the Internet e.g. Android users need to download, install and maintain individual applications. However, there are too many service providers that have their own application that need to be updated regularly because new functionality is added. Many users lose control of this administrative process which bears additionally security risks. In this paper, we suggest a generic android application programming (GAAP) comparable to a middleware where the application logic and presentation layer is placed on the server. By this approach users do not need to update their installed application. Additionally, we suggest using android-like syntax and message format for layout and GUI following the KISS-principle (Keep it simple and stupid). For evaluation purpose we developed an application providing typical GUI elements for users to demonstrate the effectiveness of GAAP approach.
[2010] Poster Abstract: Real-World G-Lab: Integrating Wireless Sensor Networks with the Future Internet (Daniel Bimschas, Sándor Fekete, Stefan Fischer, Horst Hellbrück, Alexander Kröller, Richard Mietz, Max Pagel, Dennis Pfisterer, Kay Römer, Torsten Teubler), In TridentCom 2010: The 6th International ICST Conference on Testbeds and Research Infrastructures for the Development of Networks & Communities, 2010. [bib] [pdf] [abstract]
Today's Internet is approaching architectural limits that are set up by its legacy architecture. It is based on technologies and algorithms that were developed about 30 years ago. Thus intensive research is done in the field of new protocols and algorithms fulfilling the needs of the Future Internet. We believe that wireless sensor networks (WSNs), sensor-equipped devices such as cellphones and other embedded systems will have a vital part in this context, providing a way to obtain information about the real world. All of these new devices will produce a huge amount of data by sensing and interacting with their environment. At present, complex low-level sensor node programming and algorithm knowledge is necessary to access sensor data, and only a few mature techniques exist to integrate heterogeneous WSNs with the Internet. The goal of Real-World G-Lab, which is part of the German-Lab (G-Lab) project, is to overcome these obstacles by working on the different levels of protocols, services and applications. We will enable developers to write applications that rely on sensor data input, without knowledge of the underlying hardware platform and the network communication algorithms. This implies that sensors are able to participate in the Future Internet as peer hosts. This enables new fields of applications but likewise opens a set of new challenges in the context of efficient request processing by WSNs. Efficiency here means the optimization of query-latency, energy efficiency and general service-guarantees regarding the access to sensor nodes. Beyond this we need the ability to scale to a large amount of requests from the Internet while maintaining low latencies. We will verify our concepts and applications inside the controllable environment of the G-LAB research network, by adding several outdoor WSN deployments to the experimental facility of the G-LAB project. In summary, Real-World G-LAB will contribute to the integration of resource-constrained (wireless) sensor devices into the Future Internet by investigating several key challenges, ranging from low-level energy efficiency to improved high-level application development.
Workshop Beiträge
[2011] Using and Operating Wireless Sensor Network Testbeds with WISEBED (Horst Hellbrueck, Max Pagel, Alexander Kröller, Daniel Bimschas, Dennis Pfisterer, Stefan Fischer), In Proceedings of the 10th IEEE IFIP Annual Mediterranean Ad Hoc Networking Workshop, 2011. [bib] [abstract]
Current surveys and forecast predict that the number of wireless devices is going to increase tremendously. These wireless devices can be computers of all kinds, notebooks, netbooks, Smartphones and sensor nodes that evolve into realworld scenarios forming a "Real-World-Internet" in the future. In our work we focus on the Future Internet with small battery driven devices forming the "Internet of Things". In recent networking research, testbeds gain more and more attention, especially in the context of Future Internet and wireless sensor networks (WSNs). This development stems from the fact that simulations and even emulations are not considered sufficient for the deployment of new technologies as they often lack realism. Experimental research on testbeds is a promising alternative that can help to close the gap. The deployment of testbeds is challenging and user and operator requirements need to be considered carefully. Therefore, the goal is to design an architecture that allows operators of WSN testbeds to offer numerous users access to their testbeds in a standardized flexible way that matches these requirements. In this paper we first identify some of the requirements, then introduce the architecture and general concepts of our WISEBED approach and show how this architecture meets the requirements of both groups. We give an overview of existing WISEBED compatible WSN testbeds that can be used for experimentation today. Main focus in this paper compared to previous work is to address the perspective of both users and operators on how to experiment or respectively operate a WSN testbed based on WISEBED technology.
[2011] RoombaNet - Testbed for Mobile Networks (Mohamed A. Hail, Jan Pinkowski, Torsten Teubler, Maick Danckwardt, Dennis Pfisterer, Horst Hellbrück), In Proceedings of the Workshops der wissenschaftlichen Konferenz Kommunikation in verteilten Systemen 2011 (WowKiVS 2011) Electronic Communications of the EASST (Tiziana Margaria, Julia Padberg, Gabriele Taentzer, eds.), volume 37, 2011. [bib] [abstract]
The design and deployment of wireless networks needs careful planning including various tools for analysis, simulation and evaluation. Therefore, development of software to support deployment of wireless networks has been subject of intensive research for several years. In particular the evaluation of the influence of mobility remains a challenging task. For deployment of mobile communication networks operators perform simulations and measurements during the planning process with large efforts. In the past the research community based their decisions on development of new protocols on simulations exclusively. While network simulators provide fast investigation of huge and also mobile networks they rely on theoretical models which are often considered as inaccurate and too optimistic. Therefore, more and more real wireless network environments called testbeds are established worldwide most of them with static nodes. Testbeds dedicated towards mobile networks remain a challenge as the effort to build such a network increases with mobility. The work here presents an approach for a fully automated real-world mobile network testbed where nodes are piggybacked on mobile robots. The platform with up to 30 mobile nodes and additional 30 static nodes can emulate various scenarios especially suited for pedestrian scenarios or for slow car movements. In this paper we introduce this testbed which is integrated into the larger Real-World GLab Internet testbed facility. We provide first details of the hardware and software components and provide first evaluations as well as present application examples.
[2011] EZgate - A Flexible Gateway for the Internet of Things (Torsten Teubler, Ulrich Walther, Horst Hellbrück), In Proceedings of the Workshops der wissenschaftlichen Konferenz Kommunikation in verteilten Systemen 2011 (WowKiVS 2011) Electronic Communications of the EASST (Tiziana Margaria, Julia Padberg, Gabriele Taentzer, eds.), volume 37, 2011. [bib] [abstract]
Two years ago a survey of the wireless world research forum predicted that in the year 2017 there will be seven trillion wireless devices for seven billion humans which is equivalent to 1000 devices per human being on the average. The future will show if this incredible number will be reached but for sure we will see an increasing number of wireless devices forming the Internet of the future. The new evolving ``Internet of Things'' is one of the challenging research topics today. With many wireless resource constraints devices, smart gateways integrating these small battery-powered devices into the future Internet will play a major role for the success of the Internet of Things. These gateways will work as a communication endpoint or proxy enabling transparent services including mechanisms for semantic service discovery, Quality of Service (QoS), and performance enhancing proxies (PEPs). In this work we will introduce a fully operable TCP/IP-Stack EZgate written in Java that allows designing and implementing such gateways for wireless networks in a flexible and fast approach and compare it with related work. We will demonstrate how the protocols in this stack can be assembled in a flexible manner, creating various types of gateways and can be easily extended to implement cross layer techniques. Finally, we evaluate the performance of the implementation for delay and throughput performance to show that EZgate is suitable for use in a productive environment.
[2010] Middleware for Smart Gateways Connecting Sensornets to the Internet (Daniel Bimschas, Horst Hellbrück, Richard Mietz, Dennis Pfisterer, Kay Römer, Torsten Teubler), In MidSens'10: The fifth international workshop on Middleware Tools, Services and Run-Time Support for Sensor Networks, 2010. [bib] [pdf] [abstract]
There is an increasing trend to integrate sensor networks into the Internet, eventually resulting in an Internet of Things. Recent efforts of porting IPv6 to sensor networks turn sensor nodes into equitable Internet peers and RESTful Web Services on sensor nodes allow a distribution of the application logic among sensor nodes and more powerful Internet nodes. The touching point between a sensor network and the Internet is the gateway which translates between the link-layer protocols used in the Internet (Ethernet, Wi-Fi) and sensor networks (IEEE 802.15.4). So far, the functionality of those gateways was fixed and simple. We propose to turn these gateways into smart gateways by enabling them to execute application code. As only the gateway has full knowledge of and control over both the sensor network and the Internet, smart gateways can act as performance-enhancing proxies and intelligent caches to preserve the limited resources of the sensor network. Also, the smart gateway can perform application-specific protocol conversion between highly optimized but non-standard protocols in the sensor network and standardized, but less efficient protocols in the Internet. In this paper we present the design of a middleware for smart gateways that allows the execution of application code on the gateway by offering simplified interfaces to the sensor network and the Internet. We also report preliminary performance results for key functions of the middleware.
Powered by bibtexbrowser

Projektpartner

  • Institut für Telematik, Universität zu Lübeck
  • Institut für Technische Informatik, Universität zu Lübeck
  • Institut für Betriebssysteme und Rechnerverbund, TU Braunschweig
  • coalesenses GmbH, Lübeck
  • Institut für Informatik, Freie Universität Berlin

Dieses Projekt wird gefördert vom Bundesministerium für Bildung und Forschung. Förderkennzeichen 01BK0906, GLab

BmBF-Logo