TRANSLATIONS OF SERVICE LEVEL AGREEMENT IN SYSTEMS BASED ON SERVICE-ORIENTED ARCHITECTURES

  • Authors:
  • Adam Grzech;Piotr Rygielski;Paweł Swiątek

  • Affiliations:
  • Institute of Computer Science, Wroclaw University of Technology, Wroclaw, Poland;Institute of Computer Science, Wroclaw University of Technology, Wroclaw, Poland;Institute of Computer Science, Wroclaw University of Technology, Wroclaw, Poland

  • Venue:
  • Cybernetics and Systems - SMART MODELING SUPPORT FOR MANAGING COMPLEXITIES AND DYNAMICS OF KNOWLEDGE-BASED SYSTEMS—PART 1
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The gain of the article is to introduce and to discuss a formal specification of a computer system's service level agreement (SLA) and its translation into the structure of complex services (composed of atomic services) delivering required functionalities and nonfunctionalities in a distributed environment. It is assumed that the SLA is composed of two parts specifying quantitative and qualitative requirements. The former requirements define the structure of the adequate complex services in form of a directed graph, where potential parallelism of atomic services performance may be taken into account. The latter—qualitative requirements—are applied to select the optimal complex service realization scenario; it is based on assumption that various atomic services distinguished in the complex services structures are available at the considered distributed environment in different versions and locations. Different versions of atomic services are atomic services delivering the required functionalities and satisfy nonfunctionalities at various levels. Different locations (installation place) of available atomic services means that the cost of atomic services delivery (communication and calculation) depends on parts of the distributed systems where the services are performed. The proposed model of SLA translation into complex services structures and variants may be applied—among others—to calculate upper and lower complex services' delivery times and to estimate the validity of possible parallelism in complex services.