LoginSign Up
← Back to all sparks
I

immudb

DEVOPS
Velocity0.0

Immutable database with cryptographic verification and tamper evidence

immudb.io ↗

immudb made itself speak Postgres — while its announced v2 rearchitecture sits unfinished.

tamper-evidentpostgres-compatibilityaudit-loggingorm-supportstalled-v2
◆Current state
immudb is a tamper-evident database that recently taught itself the PostgreSQL wire protocol in earnest: pg_catalog and information_schema resolvers, PostgreSQL compatibility functions, SHOW emulation, and subquery support, enough for Django, SQLAlchemy, GORM, and ActiveRecord to introspect it. Its own verification primitives — state, row and transaction proofs, history — are now callable as SQL functions over that same wire. The follow-on 1.11.1 release adds opt-in value-level audit verification and index-only COUNT(*).
◆Where it's heading
The direction is unmistakable: stop asking developers to adopt a bespoke API and instead present as Postgres with integrity guarantees attached. Every recent addition either removes an adoption barrier (ORM introspection, compatibility functions) or makes the integrity story operationally usable (structured audit logging, SHA-256 value re-verification during audit cycles). Sitting against that is a v2.0.0 release candidate announcing a page-based indexing redesign and multi-ledger storage that has not reached GA while the 1.11 line kept shipping — the rearchitecture is announced, not delivered.
◆Prediction
Expect continued PostgreSQL surface-filling on the 1.11 line, since each compatibility gap is the concrete thing blocking an ORM. What the entries do not support is a confident call on v2: the RC has been outpaced by 1.11 releases, and nothing in the changelogs indicates whether it is being finished or quietly shelved.

◆Recent moves

  1. 1mo ago

    Opt-in value verification during audits, index-only COUNT(*)

    The audit path gets teeth: an opt-in flag re-reads and SHA-256-verifies each committed value after the chain proof, so on-disk tampering fails an audit cycle rather than passing a chain-only check. Sysadmin bootstrapping from the CLI and index-only COUNT(*) round out an operations-focused follow-up to the PostgreSQL compatibility release.

    View source ↗
  2. 3mo ago

    PostgreSQL wire compatibility opens immudb to mainstream ORMs

    ⚡ SPARK

    This is the release that redefines who can adopt immudb: by resolving pg_catalog and information_schema and emulating the compatibility functions ORMs probe for, it becomes a database existing application stacks can point at rather than integrate with. Every subsequent release builds on this footing.

    View source ↗
  3. 3mo ago

    Release candidate for the PostgreSQL compatibility work

    The second release candidate for 1.11.0, restating the same PostgreSQL compatibility and audit-logging notes verbatim. No independent content — the GA release four days later is the one to cite.

    View source ↗
  4. 3mo ago

    First 1.11.0 release candidate, docs commit only

    A release candidate whose entire visible content is a truncated docs commit subject about ScanRequest.sinceTx semantics. Contentless as a changelog entry; the feature set only becomes legible at RC2.

    View source ↗
  5. 1y ago

    v2 RC announces page-based indexing and multi-ledger storage

    ⚡ SPARK

    A release candidate announcing a rebuilt indexing layer on fixed 4KB pages and a multi-ledger storage architecture where databases share indexing resources, write buffers, and caches. Read against the 1.11 releases that shipped after it, this is a direction announced but not yet delivered — the 1.x line remains where actual work lands.

    View source ↗