Research Background & Evaluation
KAIROS builds on research in dependable distributed systems, State Machine Replication, efficient state transfer, knowledge graphs, and LLM-oriented retrieval. The project aims to validate its ideas through systematic benchmarking.
Research Foundations
FlexCast
Prior work on efficient communication and ordering in partitioned, replicated distributed systems.
Read Paper →Parallel SMR Execution
Research on parallel execution strategies for replicated state machines and improved throughput under contention.
Read Paper →B+AVL Trees
Research on data structures for efficient state management, synchronization, and recovery in replicated environments.
Read Paper →Planned Evaluation Axes
Latency under Dynamic Workloads
KAIROS will evaluate whether incremental processing reduces response time when graph data changes continuously.
Scalability and Throughput
The project studies how replicated partitions, scheduling, and communication protocols affect graph-processing throughput.
State Transfer and Recovery
The research examines how optimized state structures can reduce synchronization and recovery overhead after failures.