grc-iit/ChronoLog

Organizes log entries using physical timestamps for distributed ordering without a central sequencer, while supporting multi-tier storage that automatically balances fast and capacity tiers. Features a plugin-based serving layer enabling direct deployment of custom services—SQL query engines, streaming processors, TensorFlow modules, and LLM integration via MCP—directly on the log with zero-coordination multi-writer/reader access optimized for RDMA and TCP.