The file or files in which the data sets are stored can be located on the hard disk(s) of the same computer or of another computer in the network (e.g. on a file server).
Files can be accessed, i.e. data can be transferred from the hard disk to the main memory, much more rapidly if a powerful file server is used, thus temporarily lightening the load on a monolithically structured application.
However, this type of application continues to have a structural disadvantage: The search and update procedures, which make up the database functionality and place the highest demands on the operating-system and hardware environment, must continue to be executed on the workstation on which the user is working in order to achieve the fastest possible response times. This means that these procedures can never match the throughput possible when the application program and database function are divided among several workstations and a dedicated server.
Naturally, some of the problems that arise in monolithic applications are the result of inferior programming, i.e. they cannot be corrected even by upsizing. However, when these applications have reached their upper limits with regard to the number of users simultaneously accessing the database and the size of the database, their usefulness has by no means come to an end. By changing the database technology on which they are based, you can maintain such applications without modification or with only minor modifications.