Nomadic Applications Traveling in the Fog

Presentation at the CN4IoT Conference 2017 in Brindisi, Italy

Transcript

1. NOMADIC APPLICATIONS TRAVELING IN THE FOG Christoph Hochreiner, Michael Vögler, Johannes M. Schleicher, Christian Inzinger, Stefan Schulte, and Schahram Dustdar
2. Motivation 2
3. Transnational Manufacturing Company 3
4. Application Characteristic Task ‣ Process Sensor Data to Recalibrate Manufacturing Machines 4 Required Characteristics ‣ Data Privacy ‣ Application Upgrades ‣ Data Aggregation of Anonymized Data ‣ Cost Efficiency
5. De-Centralized Approach 5
6. Centralized Approach 6
7. Centralized vs. De-Centralized Approach Centralized - No Data Privacy - High Data Transfer + Resource Pooling + Easy Application Updates 7 De-Centralized + Data Privacy + Low Data Transfer - Expensive Hardware - Complex Application Updates Goal: Combine both Approaches
8. System Design 8
9. Nomadic Applications Approach 9
10. Enabling Concepts ‣ Mobile Agents ‣ Fog Computing ‣ Container Technologies 10
11. Requirements 11 Application Management ‣Census and Recovery ‣Evolution ‣Housing ‣Transport
12. Requirements 12 Data Management ‣Deposit Data ‣Shared Data ‣On-site Data ‣Application Data
13. System Design 13 Fog Fog Fog Computational Ressources Nomadic Application Application Data Nomadic Application On-site Data Internet Cloud Application Housing Nomadic Application Application Data Nomadic Application Application Evolution Application Census & Recovery Deposit Data Shared Data Application Transport
14. Open Research Challenges ‣ Scalable Privacy Protection for On-site Streaming Data ‣ Target Scheduling for Stateful Applications ‣ Support Different Speeds in the Transport Layer ‣ Recovery Mechanism 14
15. Conclusion & Outlook 15
16. Conclusion Nomadic Applications are suited to solve privacy and data transfer related issues for the IoT while using computational resources which are already in place. 16
17. Outlook ‣ Finalizing the Infrastructure for Nomadic Applications ‣ Evaluate the Suitability of Docker Container and Unikernels ‣ Investigate towards Recovery Mechanism ‣ Investigate towards Privacy Protection Mechanism for Streaming Data 17
18. Q & A Christoph Hochreiner c.hochreiner@infosys.tuwien.ac.at