DECENTRALIZED CONTROL AND Felix Leif Keppmann ADAPTATION IN DISTRIBUTED DECENTRALIZED APPLICATIONS CONTROL AND via Web and Semantic Web Technologies ADAPTATION IN DISTRIBUTED APPLICATIONS via Web and Semantic Increased use of mobile devices, wearables, Web Technologies and sensors characterizes current develop- ments in multiple domains. In this context, the visions of the Internet of Things, Web of Things, and Semantic Web of Things as well as related visions such as Industry 4.0 promise interconnection and collaboration between billions of “things”. Still, what we are currently witnessing is the proliferation of isolated is- lands of custom solutions that cannot be easily integrated or extended. The work presented in this book provides an approach and an implementation for enabling decentralized control in distributed applica- tions composed of heterogeneous components APPLICATIONS IN DISTRIBUTED AND ADAPTATION CONTROL DECENTRALIZED by benefiting from the interoperability provid- ed by the Web stack and relying on semantic technologies for enabling data integration. In ISBN 978-3-7315-0966-0 particular, the concept of Smart Components enables adaptability at runtime through an adaptation layer and is complemented by a reference architecture as well as a prototypi- Gedruckt auf FSC-zertifiziertem Papier auf FSC-zertifiziertem Gedruckt 9 783731 509660 cal implementation. Keppmann Felix Leif Keppmann Decentralized Control and Adaptation in Distributed Applications via Web and Semantic Web Technologies Decentralized Control and Adaptation in Distributed Applications via Web and Semantic Web Technologies by Felix Leif Keppmann Decentralized Control and Adaptation in Distributed Applications via Web and Semantic Web Technologies Zur Erlangung des akademischen Grades eines Doktor der Ingenieur- wissenschaften (Dr.-Ing.) von der KIT-Fakultät für Wirtschaftswissen- schaften des Karlsruher Instituts für Technologie (KIT) genehmigte Dissertation von M.Sc. Felix Leif Keppmann Tag der mündlichen Prüfung: 03. September 2018 Hauptreferent: Prof. Dr. Rudi Studer Korreferent: Prof. Dr. Oscar Corcho Impressum Karlsruher Institut für Technologie (KIT) KIT Scientific Publishing Straße am Forum 2 D-76131 Karlsruhe KIT Scientific Publishing is a registered trademark of Karlsruhe Institute of Technology. Reprint using the book cover is not allowed. www.ksp.kit.edu This document – excluding the cover, pictures and graphs – is licensed under a Creative Commons Attribution-Share Alike 4.0 International License (CC BY-SA 4.0): https://creativecommons.org/licenses/by-sa/4.0/deed.en The cover page is licensed under a Creative Commons Attribution-No Derivatives 4.0 International License (CC BY-ND 4.0): https://creativecommons.org/licenses/by-nd/4.0/deed.en Print on Demand 2020 – Gedruckt auf FSC-zertifiziertem Papier ISBN 978-3-7315-0966-0 DOI 10.5445/KSP/1000097534 Acknowledgments My work on this doctoral thesis has been supported by several people. They provided me with research opportunities, motivated me, kept me focused, and supported me both in my professional and private life. I thank my doctorate supervisor, Prof. Dr. Rudi Studer, the second doctorate supervisor, Prof. Dr. Andreas Harth, and the reviewer of my thesis, Prof. Dr. Oscar Corcho. All three are respected researcher and well-known in the re- search community. Rudi Studer provided me with the opportunity to work as a research associate at the Institute of Applied Informatics and Formal Descrip- tion Methods (AIFB) of the Karlsruher Institute of Technology (KIT). His guidance, experience, and notable leadership style provided me with both the freedom of researching my own topics and the opportunity to take over respon- sibility in projects and management. Andreas Harth guided and challenged my research in close collaboration, enabling me to benefit from his founded knowl- edge in this research area. Oscar Corcho provided me with valuable feedback on my thesis and was a integral member of the examining board. I thank my wife, Dr. Maria Maleshkova, for her strong support in both my professional and private life. In our professional life, we collaborated on dif- ferent research topics and on hosting events for the research community. But especially in my private life, Maria has been always supporting and motivating me to stay focused and to complete my research. Without her support, my work on this doctoral thesis would have not been finished with the same success. I thank my parents, Vera and Dr. Hans Adolf Burbach, for their strong support during all chapters of my life. They always encouraged me to keep on my journey through bachelor, master, and doctoral studies. Finally, I would like to thank all people that have been supporting me but have not been mentioned here. i Abstract Current developments in multiple domains are characterized by the increased use of mobile devices, wearables, and sensors. In this context, the visions of the IoT, the WoT, and the SWoT as well as related visions such as the I4.0 promise the interconnection and collaboration between billions of “things”. Still, what we are currently witnessing is the proliferation of isolated islands of custom solutions, which support a restricted set of protocols and devices, and cannot be easily integrated or extended. In general, these visions mark a shift towards more modularized and distributed application designs, in which appli- cations are composed of several smaller, virtually or even physically separated, components with distinct domain-specific capabilities that communicate via a network to provide value-added functions. The design of distributed applications, built on top of a diverse landscape of components with a multitude of involved stakeholders, poses a number of chal- lenges. These include overcoming the data and communication heterogeneity of the involved components, dealing with requirements of multi-stakeholder scenarios, where a priori we hardly know the needs and constraints of all possible integration scenario, and enabling decentralized control in distributed applications that are composed of several distinct developed components. To this end, this thesis aims to benefit from the basic interoperability provided by the Web stack and relies on semantic technologies for enabling data integra- tion to provide an approach and an implementation for enabling decentralized control in distributed applications composed of heterogeneous components. In particular, we introduce the novel concept of Smart Components, which enable adaptability at run-time through an adaptation layer, and give a reference archi- tecture with a specific prototypical implementation. The presented solutions are thoroughly evaluated in terms of function, performance, and scalability. For the scalability tests, we design and implement a benchmark environment that enables us to easily evaluate use cases with only a few or multiple components. iii Contents Acknowledgments .......................... i Abstract ................................ iii Figures ................................. ix Tables ................................. xi Listings ................................ xiii Algorithms .............................. xv Equations ............................... xvii 1 Introduction ............................ 1 1.1 Challenges . 2 1.2 Hypothesis . 6 1.3 Research Questions . 6 1.4 Methodology . 8 1.5 Contributions . 10 1.6 Outline . 12 1.7 Publications . 12 2 Foundations ............................ 15 2.1 The Web and Related Visions . 15 2.1.1 World Wide Web (WWW) . 16 2.1.2 Semantic Web (SW) . 17 2.1.3 Web of Data (WoD) . 18 2.1.4 Internet of Things (IoT) . 19 2.1.5 Web of Things (WoT) . 22 v Contents 2.1.6 Semantic Web of Things (SWoT) . 24 2.1.7 Positioning of the Thesis . 27 2.2 Paradigms, Architectures, and Technologies . 28 2.2.1 Representational State Transfer (REST) . 28 2.2.2 Semantic Web Technologies (SWT) . 47 2.2.3 Linked Data (LD) . 58 2.3 Concepts and Terminology . 64 2.3.1 Component . 64 2.3.2 Application . 66 2.3.3 Interaction . 68 2.3.4 Meta-interaction . 69 2.3.5 Processing . 70 2.3.6 Lifecycle . 71 3 Component Adaptation and Decentralized Application Control ....................... 73 3.1 Introduction . 73 3.1.1 Scenario . 75 3.1.2 Challenges . 77 3.1.3 Related Work . 86 3.1.4 Contributions . 87 3.2 Approach for Smart Component-based Integration . 88 3.2.1 Requirements . 90 3.2.2 Smart Component-based Integration Architecture . 97 3.2.3 Smart Component . 100 3.3 Implementation of the Smart Component Adaptation Framework . 112 3.3.1 Smart Component Adaptation Layer . 112 3.3.2 Smart Component Adaptation Ontology . 118 3.3.3 NIREST Smart Component . 121 3.4 Evaluation . 123 3.4.1 Evaluation of Function . 124 3.4.2 Evaluation of Performance . 135 3.5 Summary . 136 4 Interaction Optimization and Mapping ............ 139 4.1 Introduction . 140 4.1.1 Scenario . 142 vi Contents 4.1.2 Challenges . 145 4.1.3 Contributions . 149 4.2 Approach for Frequency-based Interaction Optimization . 149 4.2.1 Optimization Scenario . 150 4.2.2 Requirements . 151 4.2.3 Frequency-based Network Model and Optimization Algorithm . 152 4.3 Approach for Domain-specific Architecture Mapping . 160 4.3.1 Mapping Scenario . 162 4.3.2 Requirements . 164 4.3.3 ROS Architecture Mapping . 166 4.4 Implementation of the ROS-REST Proxy . 173 4.5 Evaluation of Frequency-based Interaction Optimization . 175 4.6 Summary . 179 5 Distributed Benchmark Generation and Provisioning .. 181 5.1 Introduction . 181 5.1.1 Scenario . 183 5.1.2 Challenges . 185 5.1.3 Related Work