Understanding and Improving Model-Driven IoT Systems Through Accompanying Digital Twins
summary
Model-driven instrumentation generates MontiThings code that records external inputs, network delays and computation latencies via DDS topics and uses model-to-model transformations to build a deterministic digital twin that replays executions. A HVAC case study records a scenario 100 times and finds about 50% failures of naive replays; adding determinism spacing (5 ms) yields 100/100 correct orders at modest timing loss. CPU overhead scales linearly, memory stays flat, and network cost grows exponentially because of vector clocks.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:13Context
- 1:37Our Approach
- 2:43Example
- 3:37Requirements
- 4:12Transformations
- 4:44Workflow
- 5:44Digital Twins
- 7:19External Parts
- 8:53Nondeterministic Function Calls
- 9:42Evaluation
- 12:28Determinism spacing
- 13:28Overhead
- 14:08Conclusion
related talks
44mDelivering business outcomes with digital twins
41mFrom connectivity to simulations - the evolution of Digital Twin
51mAzure Digital Twins
45mDigital twin – from IoT to smart future.
49mI(o)T takes two! Meet Azure Digital Twins
56mSimplifying the digital twin journey with IoT TwinMaker
49mTechnical Integration of Digital Twins in IoT Solutions
59mInnovating power plant operations through digital twins