| 摘要 |
A technique to replace components adaptively to a sign of imminent failure or actual failure without allowing error or deterioration in performance to occur at the time of replacing components is described. A technique to have new requests directed to the replaced components executed by the new components is also described. A substitution component for a component and the condition presaging the occurrence of a failure in a component as defined; the condition is checked for each component on the basis of the state of each component and the state of the execution platform; if the condition for a certain component A is met, the instance of a substitutive component As is created on the application execution platform; a pair of the interface of that instance and the name As identifing that instance is registered into a name server; the swap function of the name server causes the interface parts of the name A of the component currently in operation and the name As of the newly registered substitutive component to be swapped; thereafter the interface of the instance of the substitution component As is enabled and named A; and the communication destination is switched to cause communication addressed to the component A to be sent to the substitution component As.
|
