> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lerian.studio/llms.txt
> Use this file to discover all available pages before exploring further.

# Operating Lerian SILOC

> SILOC's deferred-net settlement windows as regulatory context, regulated-certificate handling, gateway contingency and relay replay, and the reconciliation grains behind Lerian SILOC.

Lerian SILOC is built around one operational reality of SILOC: it is a **deferred-net, business-day** settlement system. The service maintains the gateway connection and relays the card-domain family across those windows; it computes no cycle and holds no position.

## Settlement windows

***

SILOC settles on a deferred multilateral net basis, on business days. Nuclea runs distinct daily settlement cycles for the boleto and the card products. These cycle windows are **Nuclea's** — Lerian SILOC does **not** compute them, open them, or close them. The service relays the cycle-period messages it receives verbatim; it acts on none of the cycle logic itself.

## Regulated certificates

***

The gateway's certificates are registered as a **public certificate plus an external custody reference** — the service holds no private key. A certificate is parsed for its subject, serial, and validity window when registered, and revoked over the API when it is retired. A **credential-disabled** state halts the gateway, so a disabled or revoked certificate fail-closes the connection rather than leaving it running on invalid credentials.

## Contingency and recovery

***

The gateway and the relay are built to hold under failure without losing or duplicating a frame:

* The gateway **circuit breaker** trips on repeated failure, and **automatic reconnect** with a capped backoff brings the connection back.
* An **unhandled message** — any code outside the card family — becomes a **fail-closed operator exception**: never dropped, never mis-routed, held for review.
* A **relay failure** on a sustained downstream outage parks the frame durably and raises a **relay-down** alert; an operator lists the parked failures and **replays** them once Lerian SLC recovers.
* **At-least-once redelivery** combined with **NUOp deduplication** guarantees no loss and no duplicate across the whole path.

## Reconciliation

***

Reconciliation runs at several grains so the connection's state never drifts:

* **Per-operation dedup ledger.** Every inbound operation is keyed by its NUOp, so a redelivered frame is recognized and not relayed twice.
* **Relayed-message audit feed.** The list of relayed card-domain messages is the audit record of what reached Lerian SLC.
* **Per-participant status and status history.** Each participant carries its operational status, and every status change is kept as a status-event history entry.