🚀 Getting Started
Getting Started Guide
Prerequisites (Rust 1.75+), building from source, running the server, first queries, and opening the admin console.
Start HereConfiguration
JSON-based config (skeindb-config.json): server bind address, ports, storage paths, cluster settings, and feature toggles.
Project README
High-level project overview, architecture summary, feature matrix, test results, and contribution guidelines.
Overview⚡ Querying & API
SkeinQL Reference
Complete reference for the native typed RPC protocol: 76 methods across 15 families, request/response schemas, JSON-RPC envelope, and batch requests.
Core APIMySQL Compatibility
Supported SQL statements, wire protocol details, prepared-statement metadata coverage for supported scalar subqueries, compatibility matrix, known deviations from MySQL 8+, and dialect bridging.
CompatibilityCompatibility Telemetry
How SkeinDB detects and reports unsupported features, compatibility scoring, and the migration advisory system.
CompatibilityPostgreSQL Compatibility
Current partial PostgreSQL v3 baseline: startup/auth handshake, SSL rejection, simple query protocol, transaction stubs, extended-protocol scaffolding, and common bootstrap probes such as SHOW server_version and current_setting(...), with parser/catalog/auth hardening still open.
OpenAPI Specification
Machine-readable OpenAPI 3.0 schema for all SkeinQL endpoints. Use with Swagger UI or any API client generator.
Schema🧬 Storage & Internals
On-Disk Format
Binary record formats, version tags, ValueID encoding, cell interning strategy, LSM level layout, and WAL record structure.
InternalsDelta-Chained Values
Delta encoding for MVCC version chains: base values, forward/reverse deltas, chain topology optimization, and dedup interactions.
InternalsCompaction Scheduler
Workload-guided compaction strategy: trigger conditions, level sizing, I/O throttling, and concurrent compaction management.
InternalsConvergent Encryption
Content-derived keys for encrypted deduplication: threat model, key derivation, and integration with cell interning.
Internals🖥️ Operations & Administration
SkeinAdmin (Web Console)
26-panel web admin: schema browser, data editor, settings explorer, telemetry/security centers, dialect-aware SQL console, cluster dashboard, and dedicated panels for all research features, including an Easy Viewer with inline DB creation + live create-table preview and a live Index Advisor page.
AdminClustering
Multi-node deployment: join tokens, topology lifecycle, shard placement, replica promotion, and RPC fanout with recursion suppression.
DistributedObservability
Prometheus metrics, structured logging, tracing spans, health endpoints, and integration with monitoring stacks.
OperationsPerformance
Benchmark methodology, throughput numbers, latency percentiles, optimization guidelines, and profiling tips.
Operations🔬 Advanced Features
Audit WAL
Hash-chained write-ahead log for tamper-evident audit trails. Forensic query language for WAL analysis.
SecurityWasm UDFs
WebAssembly sandbox for user-defined functions: safe cancellation, memory limits, and WASI integration.
ExtensionsWasm Query Operators
WebAssembly-compiled query operators for portable execution across server, browser, and edge runtimes.
ExtensionsETag Validators
Cache-coherent reads via HTTP ETag chains: row/query validators, cacheable prepared-query GETs, and 304 Not Modified responses.
CDC Changefeed
Baseline CDC today: table subscriptions with long-poll delivery, acked cursors, and close semantics; query-scoped streaming remains planned.
StreamingTraffic Reduction
Encoding strategies for network efficiency: delta patches, column projection pushdown, and result compression.
Web-NativeQUIC Transport
HTTP/3 and QUIC-native database protocol: stream multiplexing, 0-RTT, and transport-level parity with HTTP.
TransportColumn Snapshots
Hybrid row+column storage: columnar materialization strategy, scan optimization, and storage format.
Storage📐 Research & Design
Research Agenda
Overview of all 20 research tracks: objectives, status, and interconnections between tracks.
ResearchSkeinIR
Intermediate representation for SkeinQL: AST, type system, optimization passes, and lowering to execution plans.
CompilerSkeinQL Research Extensions
Experimental SkeinQL methods for research features: privacy queries, learned index hints, and Wasm operator registration.
ExperimentalAutoparameterization
LLM-assisted query autoparameterization: literal extraction, parameter reuse, and plan cache hit optimization.
AI📋 Additional Topics
Incremental Views
Dependency-graph–based incremental view maintenance: change propagation, invalidation, and materialized refresh.
Index Advisor
Automatic index recommendations from workload analysis, including the live SkeinAdmin advisor page with ranked suggestions, action history, and observed-before/expected-after reports.
Learned Indexes
Machine-learned index models for ValueID lookup: training, serving, fallback strategies.
Merge Functions
CRDT-inspired merge semantics for conflict resolution: register types, set operations, and custom merge UDFs.
Time-Travel & Replay
Point-in-time queries, MVCC as_of reads, and the maintenance.history.* retention + GC surface (T180–T182), plus deterministic replay bundles.
Oblivious Execution
Access-pattern–hiding query execution for multi-tenant privacy: ORAM-based scans and dummy operations.
Query Coalescing
Merging concurrent identical queries into a single execution with shared result broadcast, including cacheable prepared-query GETs and query.patch.
Query Patch
Sending query result patches instead of full results: diff generation, client reconciliation.
Offline Write Queue
Client-side write buffering for offline operation: queue management, conflict resolution on reconnect.
CAS Replication
Content-addressed storage replication: dedup-aware sync, Merkle-tree verification, and bandwidth efficiency.
Telemetry & Migration
Compatibility telemetry collection, migration readiness scoring, and advisory system for MySQL-to-SkeinQL transition.
Project Backlog
Complete engineering backlog: prioritized tasks, milestones, and implementation roadmap.
True Status Matrix
Authoritative status for all 20 research tracks: 14 hardened (R02–R11, R13–R16) with evidence and test links.
Status🔬 20 Research Tracks — Detailed Pages
Each track has a dedicated page with novelty analysis, technical phases, testable hypotheses, and key references.