Methodologies for Design of Service-Based Systems - Intentional Perspectives on Information Systems Engineering - page 305

guidelines for service design, while later approaches such as the P2S methodology
[ 2] provide a systematic and algorithmic approach to measure service cohesion and
coupling and to derive appropriate granularity levels based on process modelling
techniques and the analysis of service interactions to identify candidate services.
However, these approaches, while they focus on specific aspects of service com-
position, they still lack an analysis at the design time of the dynamic properties
which are characteristics of the service approach.
While the first attempts to provide dynamic service composition at run time have
been based on managing service invocations to guarantee quality constraints or to
heal failures, a more systematic approach to designing service-based application
based on an adaptivity paradigm at run time has been proposed within the European
S-Cube network on Software Services and Systems. S-Cube is focusing not only
on service composition at design time, but also on the adaptive characteristics of
service-based approaches at rune time. One of the results of the project is to dis-
tinguish clearly among adaptation needs, strategies, and enactment, and to include
their consideration during design. In this way the service life-cycle can be extended
separating the traditional design phases and the run time service management and
adaptation, creating the links between them (see Fig. 2) .
In particular, during requirements engineering and design and service construc-
tion, the adaptation contexts and strategies need to be defined in conjunction with
the service design as described in previous approaches, to enable systematic service
adaptation at run time [ 3] . At run time, to be able to enact adaptation, first adaptation
needs are identified, then adaptation strategies are applied.
In the following of the chapter, we focus mainly on the phases which go from the
early requirements engineering to service design, considering in particular how the
adaptation aspects can be considered at design time to enable run-time adaptation.

Fig. 2 S-Cube life-cycle
of adaptable service-based
applications [ 3]

Early Requirements
Engineering

Requirements
Engineering
& Design

Identify
adaptation
need

Operation &
management

Identify
adaptation
strategy

Construction

Deployment &
provisioning

Enact
adaptation

3 Requirements

In most of the proposed service-oriented design approaches, requirements engineer-
ing is the first phase in service design and development. However, in the literature,
most authors refer to traditional approaches for this phase.