VMworld vBrownBag vmtn5105e - Towards optimal IoT workload placement in a fog computing environment
- 1. Towards optimal IoT workload placement in a fog computing environment Kenneth Moore 7th November 2019 vmworld VMTN5105E
- 2. © Copyright 2019 Dell Inc. Background Kenneth Moore Data Scientist & PhD Student @ Dell EMC Open HCI Solutions Researching workload modelling and simulation of Converged Infrastructure SDDC workloads for large-scale IoT deployments
- 3. © Copyright 2019 Dell Inc. What considerations are required for Placing IoT workloads on Fog / Edge nodes?
- 4. © Copyright 2019 Dell Inc. Fog Computing Architecture Source: OpenFog Consortium Higher Lower Higher Lower
- 5. © Copyright 2019 Dell Inc.Source: OpenFog Consortium
- 6. © Copyright 2019 Dell Inc. Fog Computing Challenges Mobility Task Placement Security & Privacy Interoperability Resource Management Energy Management QoS & QoE Fog Computing Challenges Mobility Task Placement Security & Privacy Interoperability Resource Management Energy Management QoS & QoE Fog/Edge Research Topics
- 8. © Copyright 2019 Dell Inc. Workload Placement Locations Broker/Auction Approach Fog Edge Node offering best solution to a desired optimisation costing will process the workload ? ? ??
- 9. © Copyright 2019 Dell Inc. Workload Placement Locations Distributed Approach Grouped Fog Edge Nodes best suited to a desired optimisation costing technique share workload processing ? ? ??
- 10. © Copyright 2019 Dell Inc. Workload Placement Locations Hierarchical Approach Workload offloaded to next level Fog/Cloud node infrastructure ? ? ??
- 12. © Copyright 2019 Dell Inc. Who makes the placement decision?
- 14. © Copyright 2019 Dell Inc. Optimisation Costing Methods Resource Awareness • Cost: Maximise resource usage • CPU (# cores or % utilisation) • Memory (capacity) • Disk (capacity) • Benefit: Maximum throughput ensuring more workloads can be processed closer to the IoT device
- 15. © Copyright 2019 Dell Inc. Optimisation Costing Methods QoS/QoE Awareness • Cost: Minimise QoS/QoE metrics • Latency • Jitter • Delay • Benefit: IoT device service requirements achieved
- 16. © Copyright 2019 Dell Inc. Optimisation Costing Methods Energy Awareness • Cost: Minimise energy usage • Benefit: reduced energy consumption • Reduced energy costs • Reduced carbon footprint • Battery capacity reserves
- 17. © Copyright 2019 Dell Inc. Optimisation Costing Methods Infrastructure Centric IoT Device Centric Environment Centric
- 18. Further Study
- 19. © Copyright 2019 Dell Inc.
- 20. © Copyright 2019 Dell Inc. Questions