Task Scheduler Layer
At the core of the system is a decentralized scheduler that categorizes and matches tasks to nodes. Incoming tasks are classified into:
AI inference
Data processing
Web3 validation
Edge compute
Developer-defined workloads
Scheduling decisions are made based on:
Device performance profiles (CPU/GPU benchmarks)
Live availability and uptime
Network latency/bandwidth
Task completion history and reputation
Geographic locality (for edge tasks)
While scheduling logic is off-chain for performance reasons, task assignment logs are anchored periodically on-chain to ensure traceability.