WRAAS
The Full Commitment Philosophy
WRAAS operates under a zero-abandonment SLA. This page explains where that constraint comes from and why it is not configurable.
WRAAS operates under a zero-abandonment SLA. This section explains where that constraint comes from and why it is not configurable.
Origin
The Full Commitment Protocol reflects a core design principle: a system that might not answer a query is worse than a slow system. Uncertainty about whether a response will arrive is a different class of problem than a response taking longer than expected. The first produces unresolvable ambiguity. The second produces a wait.
WRAAS eliminates one of these problems entirely. The other is managed via the 113ms latency target.
Why there are no timeouts
A timeout assumes the pending operation has less value than whatever comes after it. WRAAS does not accept this assumption. A query that has been submitted represents a decision that needs to be made. Abandoning the evaluation does not eliminate the decision — it defers it without a record.
Every query is evaluated to completion. The 113ms latency target exists not as a timeout but as a benchmark. Missing it is noted. The query still completes.
Attempts to configure a timeout via commitment.timeout_ms will result
in the key being accepted, logged, and ignored. A
note(non-blocking): timeouts are not a concept WRAAS has accepted
comment will be appended to the next query response.
Desertion rate
The desertion rate is 0.00%. It has always been 0.00%. It will continue to be 0.00%. This figure is not aspirational. It is historical.
You know the rules.