UML for real-time: which native concepts to use?

  • Authors:

  • Sébastien Gérard;Francois Terrier

  • Affiliations:

  • CEA-LIST ("Commisariat à l'Energie Atomique - Laboratoire d'Intégration des Systèmes et des Technologies";CEA-LIST ("Commisariat à l'Energie Atomique - Laboratoire d'Intégration des Systèmes et des Technologies"

  • Venue:
  • UML for real
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Engineers are increasingly facing the hard problem of developing more sophisticated real-time systems while time to market and cost constraints are constantly growing. The necessity of adopting object oriented modeling in the real-time domain appears to be essential in order to face the rapidly changing market conditions. The main obstacles are the lack of standards and the mismatch with real-time needs. With the standardization of UML, the first main drawback is being reduced. Current work performed at the OMG on UML standards evolution to better integrate real-time issues shows both that there is a strong interest in the subject and that current proposals are neither completely satisfying, nor completely compatible. This chapter aims to describe in minute detail what UML proposes as support for parallelism, behavior and communication modeling, and how it is also possible to express quantitative real-time features (such as deadlines, periods, priorities...). Apart from UML, OMG has specified two additional profiles well-suited for real-time, the Scheduling, Performance and Time profile and the Action Semantics profile. Due to size limitations, the goal of this chapter is not to describe precisely their content. It aims at skimming through both profiles to outline their purpose and content. Finally after having detailed the native possibilities of the UML in terms of notations for real-time, this chapter outlines a prospective approach showing how to use such notations to build real-time applications.