Capacity after a host failure

A fictional design must serve nine hundred simultaneous sessions under an agreed response-time target. A representative test of the intended image establishes one hundred and eighty usable sessions per host at that target. The proposed deployment has five identical hosts, and the manager calls it resilient because every host is configured the same way. The requirement also states that the service must continue at the full demand after one host is lost. No extra capacity has been included. The numerical model treats every host as independently available, assumes sessions can be placed on surviving capacity as required by the design and does not establish automatic migration of existing sessions. Separate teams still need to validate storage, Controller, network, identity, profile and application-service capacity. The architect must explain the calculation and its limits.