A method for, from a server, controlling transfers of jobs from clients to shared resources in the network comprises that a request for sending a job to a shared resource selected by a client is received from the client and it is checked continuously whether the resource is available and has capacity by the moment for reception of jobs. A go-ahead is sent immediately to the client that the client can send the job to the resource if the resource is available and has capacity by the moment. The request is placed in a queue for the resource if the resource is available, but for the moment lacks capacity, the queue is updated continuously and a go-ahead is sent to the client that the client can send the job to the resource when the request has advanced to the first position in the queue and the resource has capacity. The step of receiving a request is followed by assigning an identity to the job, which is sent to the client, and the identity is comprised in the go-ahead sent to the client so that the client from said identity can identify the job. A confirmation that the job has been completed successfully by the resource or an indication that the job has not been completed successfully by the resource is received from the client, after which the request is removed from the queue.