Durability & trust

A database earns trust by not losing data.

SynergyDB is built on one principle: a committed write stays committed. Transactions are ACID, writes go through a crash-safe log before they're acknowledged, and the same guarantees apply whether the data is a table, a document, a cached key, a graph edge, or a vector.

The foundation

One set of guarantees, across every model

Consolidating six systems into one only helps if that one is dependable. These are the properties that make it safe to put your data — all of it — in a single engine.

ACID transactions

Multi-version concurrency control gives readers a consistent view while writers work, and a transaction either fully commits or leaves nothing behind.

Crash-safe write-ahead log

Every write is recorded in a durable log before it's acknowledged, so a committed change is replayed on restart if the server stops unexpectedly.

Replication & failover

A consensus-replicated cluster keeps copies in step and promotes a replica if a node goes down, so the data survives losing a machine.

Encryption at rest

Data at rest can be encrypted with AES-256, with a distinct key derived for each tenant and purpose so one key can't unlock another's data. Keys are held separately from the data.

Role-based access control

Ownership and grants are enforced by the engine across every SQL wire, with no default superuser password — a new install generates its own.

Audit logging

Actions are recorded so you can see who did what and when — one audit trail covering every model in the engine, not one per system.

In plain terms: putting every kind of data in one engine only pays off if that engine is dependable. SynergyDB is built so a committed write is durable, access is controlled, and actions are logged — the same way, for tables, documents, cache, graph, and vectors alike. Encryption and access controls are configured for your deployment; a walkthrough covers exactly how. See what connects on the Compatibility page.

Questions about durability?

We'll walk your team through the transaction, replication, encryption, and access-control model, and how it's configured for your environment.

Book a technical walkthrough