1. Matthias Kovatsch Research History & Career Background

2. Dr. Matthias KOVATSCH Citizen of Germany Objective Research and development in networked embedded systems, large-scale systems, and RESTful environments and building an Internet of Things (IoT) around them Education Dipl.-Ing. in Erlangen, Germany Doctorate in Zurich, Switzerland Erlangen

3. Dipl.-Ing.Oct 2003 – Nov 2008 Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany Information and Communication Technologies (semesters 1 – 6) Signals and Systems Communications Algorithms Distributed Systems Systems of Information and Multimedia Technology (semesters 7 – 10) Elite graduate program of the Elite Network of Bavaria Wireless Communication Embedded Systems Thesis “Services for Wireless Sensor Networks” Electrical Engineering (Dipl.-Ing. degree) Computer Science Individual Curriculum

4. DoctorateMar 2009 – Mar 2015 ETH Zurich, Switzerland Institute for Pervasive Computing / Distributed Systems Group Wireless Sensor Networks (Contiki OS) Seamless Internet connectivity (6LoWPAN) A Web-like application layer for the Internet of Things (CoAP) Dissertation “Scalable Web Technology for the Internet of Things” Teaching responsibilities Supervision of 20 bachelor and master students Main teaching assistant for the Distributed Systems lecture

5. Professional Experience (selected) Freelance Web Developer, Germany Custom online shop and content management systems for small businesses Fraunhofer IIS, Germany Technical study of ZigBee and middleware for medical applications Siemens Pte Ltd, Singapore Automated testing tools for SMT placement equipment via CAN bus SICS – Swedish ICT, Sweden Protocol development for Contiki OS ETH Zurich, Switzerland Publication of scientific articles, standard documents, and open source projects

6. Publications (selected) Embedding Internet Technology for Home Automation M Kovatsch, M Weiss, D Guinard Emerging Technologies and Factory Automation (ETFA), 2010 A Low-Power CoAP for Contiki M Kovatsch, S Duquennoy, A Dunkels Mobile Adhoc and Sensor Systems (MASS), 2011 Moving Application Logic from the Firmware to the Cloud: Towards the Thin Server Architecture for the Internet of Things M Kovatsch, S Mayer, B Ostermaier Innovative Mobile and Internet Services in UbiComp (IMIS), 2012 Actinium: A RESTful Runtime Container for Scriptable Internet of Things Applications M Kovatsch, M Lanter, S Duquennoy Internet of Things (IoT), 2012 CoAP for the Web of Things: From Tiny Resource-constrained Devices to the Web Browser M Kovatsch Web of Things (WoT), 2013 Californium: Scalable Cloud Services for the Internet of Things with CoAP M Kovatsch, M Lanter, Z Shelby Internet of Things (IoT), 2014 Google Scholar Citation indicesAllSince 2010 Citations424421 h-index10 i10-index10

7. Open Source Projects Californium (Cf) CoAP framework Java project for CoAP in unconstrained environments Eclipse Foundation Erbium (Er) REST Engine C project for Class 1 devices (~100 KiB ROM, ~10 KiB RAM) Contiki OS Copper (Cu) CoAP user-agent Mozilla Firefox add-on for user interaction, testing, debugging OMA LWM2M DevKit Mozilla Firefox add-on for virtual LWM2M Clients (devices)

8. Standardization Activities Internet Engineering Task Force (IETF) IoT-related working groups Shepherd for draft-ietf-core-block European Telecommunication Standards Institute (ETSI) IoT CoAP Plugtests (involvement in TTA CoAP interop event in April 2015) World Wide Web Consortium (W3C) Web of Things Interest Group (WoT IG) Open Mobile Alliance (OMA) Collaboration with Joaquin Prado for improved openness OMA LWM2M workshop and feedback on technical specification

9. Technical Question How to get an interoperability between IoTivity and other IoT standards such as OneM2M and AllSeen?

10. Differentiation of Goals IoTivity/OIC device in other environments Other devices in OIC environment Mixed environments Small device Powerful device Local communication Cloud-based communication

11. Differentiation of IoT Standards Service layer specifications OneM2M IoT-A (EU project) Data model specifications IPSO Objects ZigBee Cluster Library* Device level specifications OIC OMA LWM2M AllSeen Bluetooth GAP/GATT Provide bindings for multiple protocols Define their own mechanisms on top Usually introduce unmanageable complexity Provide bindings for multiple protocols Reuse protocol mechanisms Often include own data model Often define own protocol Usually define the local environment * for IP-based ZigBee solutions

12. Example Solution: OneM2M Binding Solution Requirements OneM2M “gives in” Advantages No change in IoTivity Working for all device classes Disadvantages Forced into OneM2M architecture Possibly limited OneM2M service functions No direct device-to-device communication Application-level Gateway OneM2M API App OIC Devices

13. Example Solution: Additional IPSO Data Model Requirements IoTivity “gives in” Advantages IPSO devices can use IoTivity devices Re-use of code (CoAP stack) No gateway or cloud service Disadvantages IoTivity devices cannot use IPSO devices IPSO functionality limited (e.g., no discovery) Unclear security model Slightly higher resource requirements on devices

14. Example Solution: AllSeen Plug-in Requirements IoTivity “gives in” Powerful device Advantages No gateway or cloud service Disadvantages Not possible for small devices Possible IPR issues

15. Example Solution: BLE (GAP/GATT) Gateway Requirements Someone “gives in” (provides gateway) Users buy additional hardware Advantages No change in IoTivity Working for all device classes GATT to REST mapping application-agnostic (if done correctly) Disadvantages Gateway (could be part of high-end device) BLE Gateway LWM2M Server with OIC northbound interface

16. General Recommendations Reuse exiting open standards as far as possible see Internet protocol suite (IP as narrow waist) look for mechanisms instead of inventing them Keep core features lean see Web architecture (REST constraints) extend with plug-ins for special content (e.g., link to RTP for multimedia) Plan for (secure) software updates see living standard in Web browsers (apply moderately in the IoT) confirm specification with running code (good approach already in OIC)

17. Promising Future Solutions Progress in CoRE working group (Thing-to-Thing Research Group) RESTful application templates (see draft-hartke-core-apps)  Common Internet Media Types derived from individual data models e.g., SenML for accessing and understanding sensors IoTivity with text/plain sub-resources? Split compound JSON data model into atomic resources OMA LWM2M with RESTful peer-to-peer support Strict Object/Instance/Resource structure is modeled through Web Linking  Common set of link attributes describing resources serving text/plain