Perfromance Benchmarking and Improvement
| MPS | 0020 |
|---|---|
| Status | Proposed |
| Category | Core |
| Authors | Karmel E - karmoola |
| Created | 06-08-2026 |
| Requires | none |
| Replaces | none |
Benchmarking – Improvements
Abstract
This problem statement concerns two coupled performance dimensions of the Midnight protocol: (1) the speed at which a node synchronizes the chain (“sync throughput”) and (2) the growth rate of the nullifier set.
Goals and Success Criteria
Satisfied when the following are established (the how belongs in a MIP):
- A reproducible sync-throughput baseline on defined reference hardware at a defined chain height, expressed in × realtime.
- A reproducible nullifier-set baseline: size and growth rate (per block / epoch / transaction), plus measured contribution to synced-state size, sync wall-clock time, and per-transaction validation cost.
- A quantified nullifier-set growth-rate or cost-ceiling target, set once the baseline is known.
- A defined improvement limit for each metric, justified by operator benefit and consensus-correctness constraints.
Motivation
- Sync time gates participation: Every new or recovering node operator must replay and validate historical chain state before participating. Slow sync raises the barrier to running a node as history accumulates.
- No benchmarking capability: There is no standard benchmark for monitoring either sync throughput or nullifier-set cost and impact of any current/future network updates. Without a benchmark, no change can be evaluated as an improvement or a regression.