Add ReSQL: SQL-Based Relational Service Layer for ResilientDB #238
Description
Title
Add ReSQL: SQL-Based Relational Service Layer for ResilientDB
Description
Summary
ResilientDB currently exposes a key-value (KV) storage abstraction, which limits expressiveness for applications requiring relational schemas, joins, range queries, and analytical operations.
This issue proposes introducing ReSQL, a SQL-based service layer that enables relational querying capabilities on top of ResilientDB’s existing distributed ledger execution model. ReSQL integrates DuckDB as an embedded SQL backend and provides structured relational semantics without modifying core consensus or ledger protocols.
Motivation
The current KV model restricts:
- Structured relational schema modeling
- Join operations
- Range queries
- Aggregations and analytical queries
- Developer ergonomics for complex application logic
Applications needing relational features must rely on external database systems. ReSQL enables structured SQL interaction directly within ResilientDB.
Proposed Changes
Storage & Execution
- Integrate DuckDB as an embedded SQL engine
- Extend execution paths to support SQL-based read/write workloads
- Add runtime flags for enabling/disabling ReSQL mode
- Maintain backward compatibility with KV-based applications
Service Layer
- Introduce a dedicated ReSQL service responsible for:
- Managing SQL contexts
- Parsing and executing SQL statements
- Handling query lifecycle
- Returning structured tabular results
Tooling
- Add standalone startup scripts for ReSQL
- Provide
test_sqlfor validating SQL workloads - Include utilities for relational schema bootstrapping
Compatibility
- Fully optional module
- No modifications to consensus or ledger protocols
- KV applications remain unaffected
- SQL execution enabled via runtime configuration
Validation
- Successful table creation
- Insert and retrieval correctness
- Selection and projection validation
- End-to-end SQL execution via
test_sql
Expected Benefits
- Enables RDBMS-style semantics for decentralized applications
- Removes need for external relational DBMS
- Improves developer productivity via standard SQL abstractions
- Leverages DuckDB’s in-process analytical execution