Fault tolerance for complex operations distributed computing

A method of allowing fault tolerance in a client process (306) running on a system of distributed computing (200), the method comprising: (a) instantiating a runtime environment (316) in response to a call a client process (306), the runtime environment (316) associated with the client process (306) running on distributed computing system (200); (B) executing instructions in said execution environment (316), causing said instructions to said execution environment (316) to issue new instructions to the system distributed computing (200), said additional instructions relating to actions to be taken respect to data stored in said system distributed computing (200); and (c) receiving said additional instructions via a proxy interface objects (320), wherein at least one of said additional instruction is an instruction to save the state which causes said interface proxy object (320) store a current state of the execution environment (316) in a data warehouse.