Systems & Infrastructure / Independent exploration
Kamish
A collaborative event-streaming system exploring signed residency policies, broker enforcement, and auditable local operation.
Problem
A streaming platform needs to decide where data may move as well as how quickly it can move. Kamish encodes placement and transfer policy as explicit broker decisions.
Architecture
The Rust workspace separates policy, enforcement, append-only storage, ownership, outboxes, routing, and audit records. A gateway, CLI, TypeScript SDK, and frontend expose the system. Earlier research used the name Hydra Stream; this existing route is retained for continuity.
Implementation
The inspected enforcement function evaluates ingress, replication, export, fetch, and promotion operations against class and destination rules. Denials and accepted operations are recorded through the audit layer.
Engineering decisions
Policy checks precede transfer decisions rather than being inferred from a dashboard. The reviewed PR also distinguishes the local in-memory authority from an external metadata service and verifies trusted policy issuers before installation.
Contribution
Saksham's merged PR and commit history document gateway, CLI packaging, policy-trust, static-path containment, and release-evidence work. Kamish remains a collaborative project; the portfolio does not attribute the complete platform to one author.
Results
VERIFIED: enforcement decisions exist in source, and Zylus08's PR #1 was merged. DOCUMENTED: the PR handoff records Rust, SDK, frontend, and clean-room checks. These counts describe that repository's run, not a rerun by this portfolio audit.
Limitations
The inspected handoff explicitly does not verify external etcd, KMS, independent witnesses, a production browser deployment, or persistent multi-region rollout. Cryptographic policy checks do not by themselves establish legal compliance, zero data loss, or production readiness.
Reproducibility
Reviewed at commit 5682594cb504 on 2026-11-09. Source inspection establishes implementation structure; repository-authored results are not independent measurements. The upstream application and benchmark suite were not rerun during this review.
Next investigation
Resolve the limitations above, preserve the reviewed commit and exact experiment configuration, and reproduce correctness checks before promoting performance or deployment claims.