Uptime for a national currency is not an SLA to be bought — it is a property to be engineered. Sovex is built so that no single failure, and no single operator, can take the currency down.
Availability begins with removing every component whose loss could halt settlement.
Ledger, key services, and settlement paths are replicated, with no single component whose failure stops the system.
Multiple nodes serve live traffic at once, so failover is the absence of one participant rather than a cold restart.
Sensitive actions require distributed approval, removing the single operator whose error or coercion could stop settlement.
Geographic resilience is delivered without data or control ever leaving the sovereign's jurisdiction.
Replicas are placed across sites inside the sovereign's territory, providing geographic redundancy while data stays resident.
Nodes span independent power, network, and facility domains, so a correlated failure cannot take a majority at once.
The state runs the infrastructure and holds the keys, so availability never depends on a foreign vendor's platform staying up.
Durability is a property of the hash-chain itself, verifiable at every replica rather than trusted from a coordinator.
The tamper-evident chain is the replication unit; every replica can independently verify it holds the true, unbroken history.
A transaction is acknowledged only once it is safely persisted across enough replicas to survive the loss of nodes.
Replicas prove agreement cryptographically rather than trusting a coordinator, so divergence is detected, not discovered later.
The architecture makes recovery a rehearsed, predictable procedure instead of an incident-time scramble.
Recovery-point and recovery-time targets are explicit design constraints the architecture is built to meet, not aspirations set after an incident.
Failed nodes are detected and evicted, and replacements rejoin and resynchronize from the hash-chain automatically.
Failure paths are exercised deliberately, so recovery behavior is known before a real outage rather than improvised during one.
Most outages come from maintenance, so the system is built to evolve without stopping settlement.
Software and configuration roll out across nodes without halting settlement or breaking the ledger's continuity.
The post-quantum stack can rotate algorithms and keys in service, so security maintenance never forces an outage.
Health, integrity, and consensus are monitored continuously, surfacing degradation before it becomes downtime.