Abstract
This chapter resumes our survey of application-level fault-tolerance protocols considering approaches based on aspect-oriented programming. Aspect-compliant programming languages allow a source code to be regarded as a pliable web that the designer can weave so as to specialize or optimize towards a certain goal without having to recode it. This useful property keeps concerns separated, bounds complexity, and enhances maintainability. Aspect programs may be used for different objectives, including non-functional properties such as dependability. To date, it is not known whether aspect-orientation will actually provide satisfactory solutions for fault-tolerance in the application layer. Some researchers believe this is not the case (Kienzle & Guerraou, 2002)—at least for some fault-tolerance paradigm. Some preliminary studies have been carried out (for instance in (Lippert & Videira Lopes, 2000)), but no definitive word has been said on the matter. It is our belief that, at least for some paradigms, aspects may reveal themselves as invaluable tools to engineer the application-level of fault-tolerance services. For this reason their approach is described in this chapter.
TopIntroduction And Objectives
This chapter resumes our survey of application-level fault-tolerance protocols considering approaches based on aspect-oriented programming.
Aspect-compliant programming languages allow a source code to be regarded as a pliable web that the designer can weave so as to specialize or optimize towards a certain goal without having to recode it. This useful property keeps concerns separated, bounds complexity, and enhances maintainability. Aspect programs may be used for different objectives, including non-functional properties such as dependability. To date, it is not known whether aspect-orientation will actually provide satisfactory solutions for fault-tolerance in the application layer. Some researchers believe this is not the case (Kienzle & Guerraou, 2002)—at least for some fault-tolerance paradigm. Some preliminary studies have been carried out (for instance in (Lippert & Videira Lopes, 2000)), but no definitive word has been said on the matter. It is our belief that, at least for some paradigms, aspects may reveal themselves as invaluable tools to engineer the application-level of fault-tolerance services. For this reason their approach is described in this chapter.