The computational processes that manifest as systems are getting ever more complex. Interconnecting several similar autonomous systems to a system of systems is frequent, e.g. manufacturers’ autonomous products are integrated to a home automation system. Traditionally, the approach for this problem has been abstraction of the details and thus, chiseling and gluing the bits and pieces together. This has been done by declaring interfaces or by using formal methods. The former declares accessibility whereas the latter may be used to gain a rigorous mathematical-logical view on the complexity and for the ability to reason on this with a set of logical rules. On top of these views, the mediator pattern is defined to provide a reusable solution for a recurring general problem. The mediator pattern encapsulates interaction between a set of autonomous systems with the intension to ease maintenance and refactoring. In this chapter, we formally integrate the mediator pattern in a correct-by-construction manner in the Action Systems formalism. The con- tribution is in introducing the mediator on an abstract level to a contemporary distributed system as a correctness preserving refinement step. In this setting, the mediator may then be further refined to provide an isolated placeholder for introduction of domain specific intelligent resilience addressing possible issues of inconsistency.
|Otsikko||From Action Systems to Distributed Systems - The Refinement Approach|
|Toimittajat||Luigia Petre, Emil Sekerinski|
|DOI - pysyväislinkit|
|Tila||Julkaistu - 2016|
|OKM-julkaisutyyppi||A3 Kirjan osa tai toinen tutkimuskirja|
01/09/13 → 31/08/17