AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL

PERVASIVE SOFTWARE SERVICES

Dhaminda B. Abeywickrama and Sita Ramakrishnan

Faculty of Information Technology, Monash University, Clayton Campus, Melbourne, Australia

Keywords:

Pervasive services, Model-driven development, Model checking, Aspect-oriented modeling.

Abstract:

Context-dependent information has several qualities that make pervasive services challenging compared to

conventional Web services. Therefore, sound software engineering practices are needed during their devel-

opment, execution and validation. This paper establishes a framework to evaluate pervasive service-oriented

software architectures. The method of evaluation is based on key features comparison. The framework con-

sists of two views: vertical and horizontal. The vertical evaluation compares several research tools to the

Aspectual FSP Generation tool developed in this research. The tools are compared across the platform-

independent and platform-speciﬁc levels of the model-driven architecture. The horizontal evaluation view is

designed to validate several desired key features that are mainly required at the platform-speciﬁc level of the

service speciﬁcation. These criteria mainly cover two aspects: formal methods and tools employed, and the

context and adaptation dimensions of the customization approach used in the services. The vertical evaluation

has demonstrated that the Aspectual FSP Generation tool has unique features in context-dependent be-

havioral modeling and code generation. The horizontal evaluation has shown that the formal methods and tools

employed, and the customization approach used in the services are effective towards the overall objectives of

this research. The approach is explored using a real-world case study in intelligent transport.

1 INTRODUCTION

A pervasive service is a special type of service that

adapts its behavior or the content it processes to the

context of one or several parameters of a target entity

in a transparent way (e.g. restaurant ﬁnder services,

attractions and activities recommendation services)

(Hegering et al., 2003). With the proliferation of ubiq-

uitous computing devices and Internet, pervasive ser-

vices continue to evolve from simple proof of concept

implementations created in the laboratory to large and

complex real-world services developed in industry.

Context-awareness capabilities in service interfaces

introduce additional challenges to the software engi-

neer. Context information is characterized by sev-

eral qualities that make pervasive services challeng-

ing compared to conventional Web services, such as a

highly dynamic nature, real-time requirements, qual-

ity of context information and automation. The addi-

tional complexities associated with these special ser-

vices necessitate the use of solid software engineer-

ing methodologies during their development, execu-

tion and validation. Most state-of-the-art approaches

to pervasive services relate to the detailed design or

implementation stages (Mandato et al., 2002; Moste-

faoui and Hirsbrunner, 2004) of the software life-

cycle, such as pervasive Web services. Little work

focuses on the early phase of design such as architec-

ture design, which is of a higher level and abstract in

design.

This systematic, architecture-centric approach

(Abeywickrama, 2010; Abeywickrama and Ramakr-

ishnan, 2010; Abeywickrama and Ramakrishnan,

2008b; Abeywickrama and Ramakrishnan, 2008a) for

modeling and verifying pervasive services integrates

the beneﬁts of sound software engineering principles

such as model-driven architecture, aspect-oriented

modeling, and formal model checking. It adopts

model-driven development to represent complex

crosscutting context-dependent functionality in ser-

vice interfaces in a modular manner, and to automate

the generation of state machine-based adaptive behav-

ior. The crosscutting context-dependent information

of the interacting pervasive services is modeled as

aspect-oriented models in UML. Using model trans-

formations (Aspectual FSP Generation tool), we

B. Abeywickrama D. and Ramakrishnan S..

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES.

DOI: 10.5220/0003461200800091

In Proceedings of the 6th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE-2011), pages 80-91

ISBN: 978-989-8425-57-7

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

ensure the correct separation of concerns of the cross-

cutting context-dependent functionality at both semi-

informal UML modeling and formal behavioral spec-

iﬁcation levels. The generated context-dependent

adaptive behavior and the core service behavior for

the pervasive services are rigorously veriﬁed using

formal model checking against speciﬁed system prop-

erties.

This paper establishes a framework to evalu-

ate pervasive service-oriented software architectures.

The method of evaluation is based on key fea-

tures comparison. The evaluation framework con-

sists of two dimensions or views: vertical and hor-

izontal. The vertical evaluation compares several

research tools to the Aspectual FSP Generation

tool developed in the current research. The tools

are compared across the platform-independent model

(PIM) and platform-speciﬁc model (PSM) levels of

the model-driven architecture (MDA). The horizontal

evaluation view is designed to validate several desired

key features that are mainly required at the PSM level

of the aspectual pervasive software services speciﬁ-

cation. These criteria mainly cover two aspects: for-

mal methods and tools employed, and the context and

adaptation dimensions of the customization approach

used in the services. The vertical evaluation has

demonstrated that the Aspectual FSP Generation

tool has unique features in context-dependent behav-

ioral modeling and code generation. The horizontal

evaluation of the approach has shown that the formal

methods and tools employed, and the customization

approach used in the services are indeed effective to-

wards the overall objectives of this research. The ap-

proach is explored using a real-world case study in

intelligent tagging for transport.

The rest of the paper is organized as follows. Sec-

tion 2 provides background information on this study.

An overview of the evaluation framework established

here is provided in Section 3. In Section 4 vertical

evaluation of the research is discussed while Section

5 addresses the horizontal evaluation. Section 6 con-

cludes this paper with a brief analysis of the evalua-

tion results.

2 PRELIMINARIES

This section provides background information on (1)

the overall pervasive services engineering process, (2)

the case study, and (3) the context-dependent adaptive

behavior generation process applied in the research.

Transport

Case Study

Use Cases

MSCs

Component Configuration

Custom UML

Profile with

c-FSP Aspects

Behavioral

Model Synthesis:

MSCs -> FSP

Model

Transformations:

c-FSP Aspects -> FSP

LTSA

Model

Checking

1. Service Specification

2. Architecture Definition

3. Architecture Modularization

Figure 1: Pervasive services engineering.

2.1 Pervasive Services Engineering

The overall pervasive service-oriented development

process is divided into three stages (Figure 1) (Abey-

wickrama and Ramakrishnan, 2008b; Abeywickrama

and Ramakrishnan, 2008a). First, using the case study

we extract use cases and deﬁne a service speciﬁcation

for the system under consideration using message se-

quence charts. Second, the architecture for the sys-

tem is deﬁned using a component conﬁguration and

an architecture model in Finite State Processes (FSP)

using the LTSA-MSC tool. Third, the architecture

model synthesized from the previous step is modular-

ized with aspect-oriented models in UML called the

contextual-FSP aspects (c-FSP aspects), and

automatically transformed into FSP before applying

model checking using the Labeled Transition System

Analyzer tool (LTSA).

2.2 Case Study

The research approach is explored using a real-world

case study in intelligent tagging for transport known

as the ParcelCall project. ParcelCall (Davie, 2002) is

a European Union project within the Information So-

ciety Technologies program. The case study describes

a scalable, real-time, intelligent, end-to-end tracking

and tracing system using radio frequency identiﬁca-

tion (RFID), sensor networks, and services for trans-

port and logistics. This case study is particularly ap-

pealing to the current research as it provides several

scenarios for representing software services that in-

teroperate in a pervasive, mobile and distributed envi-

ronment. A signiﬁcant subset of the ParcelCall case

study is exception handling that needs to be enforced

when a transport item’s context information violates

acceptable threshold values. The reference scenario

used here describes an awareness monitoring and

notification pervasive service, which alerts

with regards to any exceptional situations that may

arise on transport items, primarily to the vehicle driver

of the transport unit. The threshold values for en-

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES

vironment status (e.g., temperature, pressure, accel-

eration) of transport items and route (location) for

the vehicle are set by the carrier organization in ad-

vance. The service alerts if items’ environment sta-

tus exceeds acceptable levels or if an item is lost or

stolen during transport. The primary context parame-

ters modeled in the study include item identity, loca-

tion, temperature, pressure and acceleration.

2.3 Adaptive Behavior Generation

The notion of context used in this research is based

on a deﬁnition provided by ()analyti for context in

information modeling. They describe context as a

set of objects, each of which is associated with a

set of names and another context called its reference.

Furthermore, they enhance the deﬁnition for context

by stating that each object of a context is either a

simple object or a link object (attribute, instance-of,

ISA) and each object can be related to other objects

through attribute, instance-of or ISA links. (Analyti

et al., 2007) use traditional object-oriented abstraction

mechanisms of attribution, classiﬁcation, generaliza-

tion and encapsulation to structure the contents of a

context.

The model transformation tool created in our

study is called the Aspectual FSP Generation

tool. The transformations have been applied to

the reference scenario in intelligent transport. We

use model transformations to automate the applica-

tion of design patterns and generate infrastructure

code for the c-FSP aspects using FSP semantics.

The current study explores the strengths of both

semi-informal UML meta-level extensions and for-

mal ﬁnite state machines for representing the context-

dependent behavior of software services, and model

transformation techniques are applied as a bridge to

enforce correct separation of concerns between these

two design abstractions. The main beneﬁts of this ap-

proach are: improving the quality and productivity of

service development; easing system maintenance and

evolution; and increasing the portability of the service

design for the pervasive services engineer.

This approach focuses on the application of

model-driven development for engineering pervasive

services at ﬁnite state machine level. An aspect

in FSP can be identiﬁed as an independent ﬁnite

state machine that executes concurrently and syn-

chronizes with its base state machine. In general,

an aspect in FSP needs to contain synchronization

events (transitions) to coordinate with its base state

machine and other aspects. Also, each aspect type

(e.g. context, trigger, and recovery) contains

its unique constructs which can be generated au-

Software Services

(MSCs)

FSP Synthesis

using LTSA-MSC

Core service FSP

c-FSP-UML Profile

c-FSP Aspects

(UML Models)

Model to Text

(UML to FSP)

Aspectual FSP

Model to Model

(UML to EMF)

EMF Intermediate Model

Model to Text

(XML to FSP)

Model Verification

(LTSA Model Checking)

Flow 1

Flow 2

Flow 3

Flow 2

Flow 1

Figure 2: Adaptive behavior generation process (source:

(Abeywickrama and Ramakrishnan, 2010)).

tomatically using model transformation techniques.

For example, a trigger aspect requires constructs

to alert and send notiﬁcations while a recovery

aspect needs constructs to recover from exception-

handling situations. On the other hand, a context

aspect has attribution, instance-of, ISA, and refer-

ence constructs from the notion of context applied

here. In Figure 2, the models and activities of the

development process are represented as ellipses and

square boxes respectively. The development process

is structured into three main ﬂows of activities. Flow

1 and Flow 2 extensively apply model transforma-

tions where Flow 1 uses a model-to-text JET trans-

formation and Flow 2 applies an effective pipeline

of model-to-model and model-to-text JET transfor-

mations. Both Flow 1 and Flow 2 originate from

the c-FSP-UML profile. Flow 3 represents activ-

ities involved for rigorously verifying the context-

dependent adaptive behavior and the core service be-

havior of the pervasive software services using formal

model checking.

3 EVALUATION FRAMEWORK

This evaluation framework (Abeywickrama, 2010)

mainly validates the main contributions or deliver-

ables of this study against several key evaluation cri-

teria. The main tools used in this study include the

Aspectual FSP Generation tool created in this

research, the LTSA model checker and the LTSA-

MSC tool. First, this research has developed a custom

tool (Aspectual FSP Generation tool) applying

an effective pipeline of model-to-model and model-

to-text JET transformations using the IBM Rational

Software Architect to automate the application of de-

ENASE 2011 - 6th International Conference on Evaluation of Novel Software Approaches to Software Engineering

PIM

PSM

Aspectual FSP

Generation tool

Other aspects-based

tools

Key feature analysis of

context and adaptation, formal methods used

Key feature analysis of

context-dependent behavioral modeling, explicit join point model of AOM,

weaving performed, and context-dependent behavioural code generation

Figure 3: Evaluation framework: vertical, horizontal views.

sign patterns and generate infrastructure code for the

aspects. Second, this research has performed rigor-

ous speciﬁcation and veriﬁcation of concurrent mod-

els of pervasive software services and their composi-

tions using the LTSA model checker to assure the cor-

rectness and quality of the pervasive services design.

Third, in this study the interaction patterns deﬁning

the pervasive software services of the speciﬁcation

have been modeled using the LTSA-MSC tool, from

which a core service model was extracted. The eval-

uation framework established in this paper intends to

validate the aforementioned three main deliverables

against key evaluation criteria.

The method of evaluation used here is based on

key features comparison. The evaluation framework

developed here does not produce additions to the re-

search methodology but instead validates the methods

and tools used in the research as a whole. The eval-

uation framework consists of two views: vertical and

horizontal. The prototype or proof of concept nature

of the methods and tools employed in this research,

however, makes it challenging to compare their bene-

ﬁts with the more advanced industry-based tools. Fig-

ure 3 illustrates the two views of the overall evalua-

tion approach.

4 VERTICAL EVALUATION

This section discusses the vertical view of the evalua-

tion framework (Abeywickrama, 2010). In the verti-

cal view of the evaluation framework, several tools on

aspect-oriented modeling (AOM) are compared with

the Aspectual FSP Generation tool as a whole

at the PIM and PSM levels of model-driven devel-

opment. As the Aspectual FSP Generation tool

covers several modeling layers of the MDA stack such

as PIM and PSM levels, the evaluation process is es-

sentially vertical in nature. The vertical evaluation es-

sentially provides an analysis of several features of

the Aspectual FSP Generation tool against sev-

eral aspect-oriented modeling-based tools. This eval-

uation is based on the following criteria: context-

dependent behavioral modeling at the PIM level, ex-

plicit joinpoint model of AOM at the PIM level, weav-

ing performed at the PIM level or PSM level, and

context-dependent behavioral code generation from

the PIM level to PSM level. The results of this evalu-

ation are presented in Tables 1- 4.

4.1 Aspectual FSP Generation Tool

The Aspectual FSP Generation tool created in

this research (Abeywickrama, 2010; Abeywickrama

and Ramakrishnan, 2010; Abeywickrama and Ra-

makrishnan, 2008b; VIDE, 2009) using IBM Rational

Software Architect provides for context-dependent

behavioral modeling at the PIM level, and context-

dependent behavioral code generation from the PIM

level to the PSM level of model-driven develop-

ment. This tool effectively applies model-driven de-

velopment in pervasive services engineering at the

state machine level. Context-dependent behavior

at the service interface level has been modeled us-

ing a custom UML proﬁle called the c-FSP-UML

profile, and aspect-oriented UML class models

called the c-FSP aspects. The proﬁle supports

an explicit joinpoint model of AOM at the PIM

level, and towards this the proﬁle deﬁnes sev-

eral stereotypes such as Aspect, ContextAspect,

TriggerAspect, RecoveryAspect, Advice and

Pointcut. The Aspectual FSP Generation tool

can be employed to generate PSMs in formal be-

havioral speciﬁcation level in FSP using an effective

chain of model transformations. Model transforma-

tions are employed here to automate the application

of design patterns and generate infrastructure code for

the c-FSP aspects using FSP semantics. Also, us-

ing transformations the correct separation of concerns

both at UML modeling and FSP behavioral speciﬁca-

tion levels is ensured. The main beneﬁts of this ap-

proach are: improving the quality and productivity of

service development; easing system maintenance and

evolution; and increasing the portability of the ser-

vice design. Weaving between an aspect and a base

state machine is performed using an explicit weav-

ing mechanism at the executable state machine level

in FSP. The context modeling and transformations

features of the Aspectual FSP Generation tool

have been explored using the reference scenario on

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES

intelligent transport.

4.2 Groher and Schulze’s Approach

(Groher and Schulze, 2003) present an approach

for specifying crosscutting concerns using aspect-

oriented modeling and discuss the seamless integra-

tion of those models to implementation. In their ap-

proach, UML has been customized for supporting

aspect-oriented modeling using UML’s standard ex-

tension mechanisms, such as stereotypes, tagged val-

ues and constraints. Their design notation for aspect-

oriented modeling provides a base package for mod-

eling the business logic, an aspect package for model-

ing the crosscutting concerns, and a connector (weav-

ing rules) for linking the aspect and the base ele-

ments. The connector includes program execution

points (pointcuts), and actions to be executed at those

points (advices). The weaving in their approach is

essentially performed at the PIM level and not at

the PSM level as in the current study. The authors

have implemented an AspectJ code generator using

the CASE tool Together from Borland. In their tool

aspect-oriented validation and code generation in As-

pectJ have been implemented as modules. The au-

thors’ work (Groher and Schulze, 2003) is not based

on pervasive services, which is a key difference to the

current research. Also, the code generation is at the

implementation level with AspectJ whereas in the cur-

rent research it is at the architectural level with FSP.

4.3 Whittle and Jayaraman’s Approach

(Whittle and Jayaraman, 2008) present a UML-based

aspect-oriented modeling tool called MATA that ap-

plies graph transformations for specifying and com-

posing aspect models. Their work is different to

most other approaches on aspect-oriented modeling

in three respects. First, there is no support for ex-

plicit joinpoints and composition is considered as a

special form of model transformations. Second, the

use of graph transformations for aspect composition,

and third, support for statically analyzing aspect in-

teractions using critical pair analysis, make their ap-

proach different to other approaches. In their ap-

proach (Whittle and Jayaraman, 2008), the compo-

sition of a base model and an aspect model (weav-

ing) is speciﬁed using a graph rule. A main differ-

ence in the authors’ approach is that graph rules have

been deﬁned over the concrete syntax of the modeling

language and not at the meta-level as in most known

approaches on model transformations. The authors

use a cell phone example to demonstrate the valid-

ity of their method and tool. Tool support for MATA

has been built using IBM Rational Software Mod-

eler. The tool uses graph transformation as its un-

derlying theory for aspect composition. MATA uses

AGG as its graph rule execution tool as it supports

critical pair analysis. Similar to (Groher and Schulze,

2003)’s approach, MATA is not based on pervasive

services. The authors’ work (Whittle and Jayaraman,

2008) does not support explicit joinpoints as provided

in the current research. Also, no code generation of

the models has been provided in their work.

4.4 Cottenier et al. Approach

()cottenierMotorolaWEAVRAspect present Motorola

WEAVR, which provides aspect-oriented weaving for

UML state charts that include action semantics. Mo-

torola WEAVR is an industrial-strength aspect weaver

for UML 2.0, which is implemented as an add-in to

the Telelogic TAU G2. The tool essentially performs

four main functions. First, the tool’s proﬁle allows

engineers to deﬁne aspects in UML 2.0. Second, it

presents a joinpoint visualization engine for visualiz-

ing and validating the effects of the aspects. Third,

the tool provides full aspect weaving at the model-

ing level. Finally, the tool’s simulation engine allows

the simulation of the aspect models without breaking

their modular structure. The authors have provided

several UML stereotype classes for identifying vari-

ous constructs of an aspect at the PIM level. Motorola

WEAVR introduces two fundamental language con-

structs to support aspect-oriented modeling. First, the

authors provide constructs to specify the locations or

joinpoints in the models where crosscutting behavior

emerges. In the approach, action joinpoints capture

call expressions, timer set actions or constructor calls,

whereas transition joinpoints capture sets of execu-

tion paths within a state machine. Second, the authors

provide constructs to specify the actual behavior of

the crosscutting concerns using the connector stereo-

type. Weaving process consists of two phases: advice

instantiation and advice instance binding. As weav-

ing is performed at the PIM level with aspects wo-

ven into executable UML models, it allows PSMs and

source code to be generated automatically. Source

code, which can be for example in C or C++, is gener-

ated using an optimizing code generator. Several ex-

amples on exception handling and recovery have been

developed by the authors to demonstrate the validity

of their approach. However, the authors’ work (Cot-

tenier et al., 2007) is not in the pervasive computing

domain as in the current research. Also, their PSMs

are not at the formal behavioral speciﬁcation level as

in our study.

ENASE 2011 - 6th International Conference on Evaluation of Novel Software Approaches to Software Engineering

4.5 Fuentes et al. Approach

()fuentes present an aspect-oriented executable mod-

eling UML 2.0 proﬁle called AOEM for design-

ing pervasive applications. Using the AOEM pro-

ﬁle, (Fuentes et al., 2008) demonstrate how aspect-

oriented executable design models for context-aware

pervasive applications can be constructed and exe-

cuted. These models are run and tested using Pop-

ulo, which is an eclipse plug-in created by the au-

thors for interpreting executable UML models. The

AOEM proﬁle aims at addressing two challenges in

pervasive applications. They are the crosscutting na-

ture of context-awareness, and the complexity associ-

ated with reasoning about the correctness of the de-

sign model. In their approach, weaving of aspects to

core components is performed at the PIM level. Spe-

cial composition rules expressed as pointcuts in the

AOEM proﬁle describe how aspects are applied to the

core components. An aspect-oriented model weaver,

which is a type of compiler or preprocessor, has been

developed for weaving. The weaver essentially cre-

ates the design model by injecting the crosscutting

behavior of the aspects into the core modules that the

aspects crosscut. The authors illustrate their approach

using a location-aware intelligent transportation sys-

tem. Although (Fuentes et al., 2008) provide for mod-

eling of adaptive behavior at the PIM level, they do

not provide any model transformations to generate

PSMs. Also, in general, their context models are at

the context-aware application and middleware levels

and not at the service interface level (state machine

level) as in the current study.

Tables 1- 4 show that the Aspectual FSP

Generation tool satisﬁes all the criteria as

opposed to the other tools.

Table 1: Comparison matrix: PIM level support for context-

dependent behavioral modeling.

Evaluation Criteria

(PIM level)

4.1 4.2 4.3 4.4 4.5

Context-dependent

behavioral modeling

Table 2: Comparison matrix: PIM level support for explicit

joinpoint model of aspect-oriented modeling.

Evaluation Criteria

(PIM level)

4.1 4.2 4.3 4.4 4.5

Explicit joinpoint

model of AOM

Table 3: Comparison matrix: PIM level or PSM level sup-

port for weaving.

Evaluation Criteria

(PIM level or PSM

level)

4.1 4.2 4.3 4.4 4.5

Weaving

Table 4: Comparison matrix: PIM level and PSM level sup-

port for context-dependent behavioral code generation.

Evaluation Criteria

(PIM level & PSM

level)

4.1 4.2 4.3 4.4 4.5

Context-dependent

behavioral code

generation

Complete

Partial

No cover

5 HORIZONTAL EVALUATION

This section describes the horizontal dimension of

the framework (Abeywickrama, 2010). In contrast to

vertical evaluation discussed above, horizontal eval-

uation validates only particular features at a speciﬁc

level of abstraction of the MDA stack. The horizontal

evaluation view is designed to validate several desired

key features that are mainly required at the PSM level

of the aspectual pervasive software services speciﬁ-

cation. These evaluation criteria cover two aspects of

the study. They are the formal methods and tools em-

ployed in the study, and the context and adaptation

dimensions of the customization approach used in the

services.

5.1 Formal Methods and Tools Used

In this subsection, the formal methods and tools used

in the current study at the PSM level of abstraction are

evaluated. (Clarke et al., 1996) provide several crite-

ria that formal methods-based approaches and tools

need to support. According to ()clarkeFormal, al-

though some of these criteria are ideals, it is still con-

sidered good to aim for them. As provided in Section

2.1, the research methodology of the current study

contains three stages: service speciﬁcation, architec-

ture deﬁnition and architecture modularization. In the

present study, formal methods and tools have been

applied during the service speciﬁcation and architec-

ture deﬁnition stages of the research methodology,

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES

and ﬁnally for model checking the aspectual pervasive

software services speciﬁcation. The LTSA-MSC tool

has been used for specifying the software services of

the service speciﬁcation and generating a behavioral

model in FSP. Using this generated behavioral model,

a core service model was extracted. The formal model

checker, the LTSA, has been used to verify the aspec-

tual pervasive software service speciﬁcation against

speciﬁed system requirements. This subsection eval-

uates the application of the aforementioned formal

methods and tools in the current research against the

criteria provided by (Clarke et al., 1996).

• Early Payback. Early payback is one of the key

beneﬁts of the current study. This study is focused

on the architectural level of the software life-

cycle. This architecture-centric approach builds

models of pervasive software services and their

compositions and veriﬁes their behavior against

speciﬁed system properties. Building architec-

tural models of pervasive software services allows

the software engineers to validate the actual cor-

rectness of the services before the services are im-

plemented later in the software life-cycle. Thus, it

provides early payback or feedback to the service

engineer on the validity of the services.

• Incremental Gain for Incremental Effort. In the

study, the PSMs of the aspectual pervasive soft-

ware services speciﬁcation have been derived fol-

lowing the three incremental stages of the re-

search methodology: service speciﬁcation, archi-

tecture deﬁnition and architecture modularization.

Each of these stages has its own deliverables such

as an message sequence chart speciﬁcation for

software services, architecture model for the soft-

ware services in FSP, and a modularized archi-

tecture with aspect-oriented models to represent

context-dependent behavior at the service inter-

face level. This demonstrates that the service en-

gineer can receive the beneﬁts of the methodology

in an incremental manner.

• Multiple Use. The pervasive services engineer-

ing methodology established in this research in

general covers the requirements and architecture

design stages of the software life-cycle. There-

fore, the beneﬁts of this methodology can be

seen in multiple stages of the software life-

cycle. This design methodology effectively fa-

cilitates the transition from requirements-oriented

scenario descriptions of pervasive software ser-

vices to architecture-centric behavioral models of

aspectual pervasive software services.

• Integrated Use. The formal methods and tools

used in this research are widely known in both

academia and industry. First, this research has

modeled the pervasive software services using

message sequence charts provided by the LTSA-

MSC tool. Message sequence charts are one of

the most widely used sequence chart notations for

describing system behavior. Second, the model

checking tool employed in this study (LTSA tool)

is widely used for behavior modeling and analysis

and is well supported with documentation (Magee

and Kramer, 2006).

• Ease of Use. This research applies three auto-

mated tools in the pervasive services engineering

process: the LTSA-MSC tool to generate the ar-

chitecture model in FSP which is later used to ex-

tract the core service model, the Aspectual FSP

Generation tool to generate context-dependent

behavior in FSP, and the LTSA tool for simulat-

ing, animating and model checking pervasive ser-

vices. The use of automated tools such as the

LTSA makes the engineering process much easier

to understand and adopt for the service engineer.

• Efﬁciency. One of the limitations of the current

study is efﬁciency in terms of time and space.

This is mainly attributed to the formal model

checking technique used for verifying the perva-

sive services speciﬁcation. One of the main chal-

lenges associated with model checking technique

is the state space explosion problem. Neverthe-

less, by using action hiding and minimization fea-

tures of the LTSA model checker the present study

has proven sufﬁciently efﬁcient in modeling and

verifying the case study subset of this research.

• Ease of Learning. The graphical interfaces pro-

vided in the LTSA-MSC tool and the Aspectual

FSP Generation tool and the automated nature

of these tools effectively reduce the need to know

formal FSP to start realizing the beneﬁts of this

research. The use of basic and high-level message

sequence charts in the service speciﬁcation stage

of the methodology are widely used and under-

stood in both academia and industry. Also, the se-

quence chart notion and semantics applied in the

LTSA-MSC tool are restricted to basic features. It

does not include complex constructs such as mes-

sage queues, gates, parametric messages, or dy-

namic creation and termination of instances.

• Orientation Toward Error Detection. The ap-

proach presented in this paper is oriented to-

wards detecting errors in the aspectual pervasive

software services speciﬁcation using the formal

model checking technique. Aspects can intro-

duce an additional correctness problem in soft-

ware speciﬁcations because of their crosscutting

ENASE 2011 - 6th International Conference on Evaluation of Novel Software Approaches to Software Engineering

effect and obliviousness nature. Therefore, tool

support if possible automatic, is highly desirable

to ensure the correctness of the speciﬁcation. This

research employs formal model checking to ﬁnd

errors concerning safety, progress, and ﬂuent lin-

ear temporal logic assertions in the service spec-

iﬁcation, which can be hidden behind the indi-

vidual components and aspects, or in the woven

model. In the study, two techniques - counterex-

amples and witness executions - have been em-

ployed to point out any errors in the speciﬁcation,

which can be used by the service engineer to im-

prove the state models or the system properties for

the aspectual pervasive software services.

• Focused Analysis. This research is explored using

a modiﬁed subset of a real-world case study

called the ParcelCall project. The case study

subset focuses on exception handling that needs

to be enforced when a transport item’s context

information violates acceptable threshold values.

The reference scenario used in the research

describes an awareness monitoring and

notification pervasive service, which

alerts with regards to any exceptional situations

that may arise on transport items primarily to

the vehicle driver of the transport unit. This is

an example where the research is focused on

analyzing only a particular aspect of the system

and not the entire system. Also, at the PSM level

only temperature and pressure context properties

have been modeled and veriﬁed. Similarly, other

context properties such as item identity, location

and acceleration can also be supported.

• Evolutionary Development. The incremental and

iterative nature of the activities performed in each

stage of the methodology essentially facilitates

evolutionary system development. This can be

demonstrated by the fact that the engineering pro-

cess, which is initiated as a scenario-based spec-

iﬁcation expressed as message sequence charts,

has evolved into a modularized service architec-

ture where complex context-dependent informa-

tion has been separated from the core service be-

havior as aspect-oriented models.

5.2 Context and Adaptation of the

Customization Approach

This subsection evaluates the customization approach

used in the pervasive services of the current study.

This evaluation focuses mainly on the PSM level of

abstraction. ()kappel and ()schwinger present a com-

prehensive and uniform evaluation framework, which

can be used to compare customization capabilities

of approaches originating from the mobile computing

and the personalization domains. The notion of cus-

tomization refers to the adaptation of an application’s

services towards the current context. Their frame-

work has two orthogonal dimensions, which are con-

text and adaptation, and the mapping between con-

text and adaptation represented by the notion of cus-

tomization. (Kappel et al., 2003) and (Schwinger

et al., 2005) provide detailed criteria for both the

context and adaptation dimensions of the framework.

Their evaluation framework aims at providing three

main beneﬁts. First, it provides a structured and uni-

form view of the various aspects of customization.

Second, the framework can be effectively used as

a conceptual framework for evaluating existing cus-

tomization approaches. Finally, the framework can be

effective for developing any future customization ap-

proaches. The criteria for context and adaptation from

(Kappel et al., 2003; Schwinger et al., 2005) are out-

lined next before applying them to validate the cus-

tomization approach used in the pervasive software

services of the current research.

• Context. Context characteristics are important for

the management of context data. These charac-

teristics can be categorized as scope of context,

representation, acquisition, and the access mech-

anism used.

– Scope. The scope of context comprises several

characteristics as outlined next. First, the scope

includes different context properties (C.P.) sup-

ported by the system, such as location, time,

device, network and user. Second, it describes

the ability to extend (C.E.) the properties in or-

der to handle any unforseen requirements. Fi-

nally, for each of the context properties the

time dimension needs to be considered with the

chronology (C.C.), validity (C.V.), and avail-

ability (C.Av.) of context data.

– Representation. The representation of con-

text is characterized by two issues: reusability

(C.R.) and abstraction (C.Ab.). First, the rep-

resentation of context needs to provide mech-

anisms to reuse already deﬁned context from

various context sources, such as geographic in-

formation systems or device proﬁles. Second,

the level of abstraction used for representing

context can be physical, such as sensed context

data, or logical, which is derived context.

– Acquisition. Acquisition is characterized by

two key issues: automation (C.Au.) and dy-

namicity (C.D.). First, the degree of automa-

tion identiﬁes who is responsible for acquiring

context, which can be either a human (manual)

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES

or the system (automatic) or a combination of

both (semi-automatic). Second, the degree of

dynamicity indicates when the context is ac-

quired, which can be statically at system start-

up or dynamically at run-time.

– Mechanism. This characteristic identiﬁes

the mechanism (C.M.) used for acquiring

the context information and made accessi-

ble to services, which can be pull-based or

push-based. Pull-based approaches request

for context information while push-based ap-

proaches provide context information when

context changes.

• Adaptation. The second dimension of the evalua-

tion framework is provided by the notion of adap-

tation. Adaptation can be categorized as the kind

of adaptation, the subject of adaptation, and the

process of adaptation.

– Kind. The kind of adaptation can be a built-

in adaptation operation (A.O.) such as ﬁlter

content, add links, or an extension mechanism

(A.Ex.), which introduces a new user-deﬁned

adaptation operation. The adaptation can af-

fect (A.Ef.) the system when certain parts of

the system are added, removed or transformed.

Complex adaptations (A.C.) can be formed by

combining a series of adaptation operations.

– Subject. The subject of adaptation is character-

ized by several issues as discussed next. First,

the level (A.L.) of the Web application that is

affected by the adaptation can be content level,

hyperbase level, and presentation level. Sec-

ond, these layers contain several application el-

ements (A.El.) that can be adapted, such as

links, pages, access structures and input ﬁelds.

Third, the granularity of adaptation (A.G.) can

be micro or macro adaptation depending on the

number of application elements affected in the

adaptation.

– Process. The process of adaption comprises

several characteristics such as task (A.T.), au-

tomation (A.A.), dynamicity (A.D.) and incre-

mentality (A.I.). The process of adaptation can

contain several tasks such as initiation, pro-

posal, selection, production, presentation and

reversion of the adaptation. This separation ef-

fectively allows ﬁne-grained control about the

automation and dynamicity of the tasks. As

in the context dimension, automation identi-

ﬁes who is responsible for performing the tasks

which can be performed automatically, manu-

ally or semi-automatically. Dynamicity of the

adaptation indicates when the adaptation tasks

are performed, which can be at design time

(static) or at run-time (dynamic). Incremen-

tality signiﬁes whether the adaptation is per-

formed from scratch or incrementally. Perform-

ing the adaptation incrementally means that the

subsequent adaptations are done based on the

results of the pervious adaptations.

Having presented the evaluation criteria provided

in (Kappel et al., 2003; Schwinger et al., 2005), next

the context and adaptation dimensions of the cus-

tomization approach used in the services are evalu-

ated using those criteria. The results of this evaluation

are summarized in two tables respectively: Table 5

and Table 6. This evaluation is part of the horizontal

evaluation dimension of the framework, and mainly

covers the PSM level of the MDA stack. However,

several examples on the PIM level are also provided.

This evaluation is presented in three logical sections

as suggested in (Kappel et al., 2003; Schwinger et al.,

2005). First, general details of the customization ap-

proach are provided covering issues on origin, ma-

jor focus, technology, architecture and implementa-

tion (if applicable) of the customization approach.

Second, the context dimension of the approach is de-

scribed, and third, the adaptation dimension of the ap-

proach is discussed.

The current research originates from the perva-

sive computing domain in general, and speciﬁcally

from the pervasive services domain. This research

is at the software architectural level, and no imple-

mentation of services such as pervasive Web services,

has been considered in the study. The main focus

of the awareness monitoring and notification

pervasive service is to alert on any exceptional

situations that may arise on transport items primarily

to the vehicle driver of the transport unit. The compo-

nent conﬁguration of the architecture deﬁned for the

pervasive software services speciﬁcation is based on

an event-control-action architecture pattern. The re-

search approach has been applied to a modiﬁed sub-

set of a real-world case study in intelligent transport

called the ParcelCall project. As stated previously

in this paper, three automated tools have been used

in the research: LTSA-MSC tool, Aspectual FSP

Generation tool, and the LTSA model checker.

The customization of the architecture can be consid-

ered internal as the system is not aware of the cus-

tomization in terms of knowing about context or adap-

tation.

The context dimension of the customization ap-

proach is described next (Table 5). The primary

context parameters modeled at the PIM level of ab-

straction comprise location, temperature and pres-

sure. However, at the PSM level only temperature

and pressure primary context properties have been

ENASE 2011 - 6th International Conference on Evaluation of Novel Software Approaches to Software Engineering

Table 5: Current study’s context characteristics.

Scope of Context Representation of Context Acquisition of Context Access of Context

Property (C.P.)

Extensibility (C.E.)

Chronology (C.C.)

Validity (C.V.)

Reusability (C.R.)

Abstraction (C.Ab.)

Automation (C.Au.)

Dynamicity (C.D.)

Mechanism (C.M.)

location

temperature

pressure

time

device

network

user

application

history

future

manual

semi-automatic

automatic

static

dynamic

push

pull

Explicitly supported

Not explicitly supported

Not applicable

Table 6: Current study’s adaptation characteristics.

Kind of Adaptation Subject of Adaptation Process of Adaptation

Operation (A.O.)

Extensibility (A.Ex.)

Effect (A.Ef.)

Complexity (A.C.)

Level (A.L.)

Element (A.El.)

Granularity (A.G.)

Tasks (A.T.)

Automation (A.A.)

Dynamicity (A.D.)

Incrementality (A.I.)

add

remove

transform

simple

complex

content

hyperbase

presentation

others

text

audio

image

video

link

others

micro

macro

automatic

semi-automatic

manual

static

dynamic

Explicitly supported

Not explicitly supported

Not applicable

modeled (C.P.). Nevertheless, at the PSM level other

context properties such as item identity, location and

acceleration can be supported as well (C.E.). These

context parameters as a whole constitute the phys-

ical context of the study. Context modeling at the

PIM level is provided by the c-FSP-UML profile

and the c-FSP aspects. The proﬁle essentially pro-

vides several stereotypes to represent the core service

behavior, the context-dependent behavior, and the de-

pendencies between the core service behavior and the

context-dependent behavior at the service interface

level. The use of stereotypes essentially supports the

notion of extensibility (C.E.). The object-oriented no-

tions used in the proﬁle such as generalization fur-

ther support the notion of extensibility. At the PSM

level, the notions of attribution, classiﬁcation, gener-

alization and encapsulation from the context deﬁni-

tion have been modeled to structure and link the ob-

jects deﬁned in the aspects (C.E.). In the study, va-

lidity period (C.V.), chronology (C.C.) or availability

(C.Av) of context are not supported as the time di-

mension of the context properties have not been con-

sidered. The explicit support provided for attribu-

tion, instance-of and ISA notions at the PSM level,

facilitates reusability of context (C.R.). The approach

provides a high-level inference mechanism to auto-

matically derive higher-level logical context (C.Ab).

Both primary and logical context are modularized

into aspects as atomic context aspects and compos-

ite context aspects respectively. Thus, the study sup-

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES

ports an explicit separation between physical and log-

ical context (C.Ab). For example, low-level temper-

ature readings from the RFID tags are inferred as

low temperature or high temperature during the re-

ﬁnement step of the pervasive service. At the PIM

level, the proﬁle is maintained manually by the ser-

vice engineer (C.Au.). At the PSM level, both physi-

cal and logical context information are acquired auto-

matically (C.Au.) and at run-time (C.D.). The perva-

sive service engineer using the LTSA animator can

select values for the temperature or pressure read-

ings from a range of values at run-time. The mech-

anism (C.M.) used to acquire context and made ac-

cessible to the pervasive service can be considered

push-based as context readings from the RFID tags

are provided based on context changes and not on

requests. The second dimension of the customiza-

tion approach is provided by the notion of adapta-

tion (Table 6). In this study, there are two types of

adaptation operations: triggering and recovery opera-

tions (A.O.). At the PSM level, both these operations

have been supported by the Trigger and Recovery

c-FSP aspects (A.O.). Trigger aspects, for

example Trigger Aspect Adverse Environment

Status, effectively send notiﬁcations or alert mes-

sages to the vehicle driver when a transport

item’s context information violates acceptable levels.

Recovery aspects, for example Recovery Aspect

Adverse Environment Status, model any recov-

ery actions that need to be enforced after an excep-

tion situation is raised by a Trigger aspect. Both

these aspects have been modeled as independent state

machines at PSM level, which synchronize with their

base state machines using explicit synchronization

events. The adaptation operations provided by the

aspects are associated with the core service model

through weaving, and the behavior of these aspects

can affect the core service behavior depending on the

context information (A.Ef.). This can be, for exam-

ple, executing or modifying a service based on con-

text information (A.Ef.). At the PIM level, adapta-

tion operations are speciﬁed using the stereotypes:

TriggerAspect and RecoveryAspect. As the adap-

tation process contains a series of stages, it can be

considered a complex process (A.C.). Also, a dis-

tinct separation can be identiﬁed between the differ-

ent tasks of the adaptation process (A.T.). When an

item’s context information is violated, ﬁrst, the per-

vasive service alerts the vehicle driver by sending an

SMS. Second, the service can perform any appropri-

ate recovery actions to remedy the situation, such as

control the refrigerator’s temperature (A.C.). Third,

the service adaptation can be extended as follows

(A.Ex.). If the environment status of items is critical

the service can alert the goods tracing server and

eventually the customer being affected through the

goods information server (A.Ex.). In the current

study, context information is represented at the ser-

vice interface level (A.L.) that essentially consists of

operation invocations and the exchange of respective

input/output parameters. The core service elements

represented at the modeling level include states, tran-

sitions, processes, and services (A.El.). Any adapta-

tion operation, which can be a triggering operation or

a recovery operation, is bound to the transitions of the

core service model. Therefore, the adaptation level

and adaptation elements of this study are the service

interface level (A.L.) and transitions (A.El.) respec-

tively. As a result, different Web application levels

such as content, hyperbase and presentation are not

applicable in this study nor the Web adaptation ele-

ments of text, audio, image, video and link provided

in (Kappel et al., 2003; Schwinger et al., 2005). The

adaptation granularity (A.G.) can be considered mi-

cro considering the number of elements affected by

the adaptation process in the study. Also, it is per-

formed automatically (A.A.) and at run-time (A.D.).

The adaptation process can be considered incremen-

tal (A.I.) as recovery operations are dependent on trig-

gering operations at both PIM and PSM levels.

6 CONCLUSIONS

To summarize this paper has discussed the evalua-

tion framework developed to validate the main meth-

ods and tools employed in this study for engineer-

ing pervasive software services. The method of eval-

uation used here is based on key features compari-

son. The evaluation framework consists of two di-

mensions or views: vertical and horizontal. The

vertical evaluation of the research compared several

research tools to the Aspectual FSP Generation

tool developed in the current research. The tools

were compared across the PIM and PSM levels of

the MDA stack. This evaluation was based on sev-

eral criteria: context-dependent behavioral modeling

at the PIM level, explicit joinpoint model of AOM at

the PIM level, weaving performed at PIM or PSM

levels, and context-dependent behavioral code gen-

eration from the PIM level to the PSM level. The

horizontal evaluation view was designed to validate

several desired key features required mainly at the

PSM level (i.e. FSP) of the aspectual pervasive soft-

ware services speciﬁcation. These evaluation criteria

mainly cover two aspects. They are the formal meth-

ods and tools employed in the study and the context

and adaptation dimensions of the customization ap-

ENASE 2011 - 6th International Conference on Evaluation of Novel Software Approaches to Software Engineering

proach used in the pervasive services.

The results of the evaluation are assuring.

The vertical evaluation has demonstrated that the

Aspectual FSP Generation tool has unique fea-

tures in context-dependent behavioral modeling and

context-dependent behavioral code generation. The

horizontal evaluation of the approach has shown that

the formal methods and tools employed in the re-

search, and the customization approach used in the

services are indeed effective towards the overall ob-

jectives of this research. However, the prototype or

proof of concept nature of the methods and tools em-

ployed in this research makes it challenging to com-

pare their beneﬁts with more advanced industry-based

tools, which is a limitation.

REFERENCES

Abeywickrama, D. B. (2010). Pervasive Services Engineer-

ing for SOAs. Ph.D Thesis, Faculty of IT, Clayton

Campus, Monash University, Australia.

Abeywickrama, D. B. and Ramakrishnan, S. (2008a). A

Model-Based Approach for Engineering Pervasive

Services in SOAs. In Proc. 5th International Confer-

ence on Pervasive Services (ICPS’08), pages 57–60,

Sorrento, Italy. ACM.

Abeywickrama, D. B. and Ramakrishnan, S. (2008b). To-

wards Engineering Models of Aspectual Pervasive

Software Services. In Proc. 3rd Workshop on Soft-

ware Engineering for Pervasive Services (SEPS’08),

pages 3–8, Sorrento, Italy. ACM.

Abeywickrama, D. B. and Ramakrishnan, S. (2010).

Model-Driven Development of Aspectual Pervasive

Software Services. In Proc. 14th IEEE International

Enterprise Distributed Object Computing Conference

Workshops, pages 49–59, Vitoria, Brazil. IEEE.

Analyti, A., Theodorakis, M., Spyratos, N., and Constan-

topoulos, P. (2007). Contextualization as an Indepen-

dent Abstraction Mechanism for Conceptual Model-

ing. Information Systems Journal, 32(1):24–60. Else-

vier Science Ltd., Oxford, UK.

Clarke, E. M., Wing, J. M., and Alur, R. (1996). Formal

Methods: State of the Art and Future Directions. ACM

Computing Surveys, 28(4):626–643. ACM.

Cottenier, T., van den Berg, A., and Elrad, T. (2007).

Motorola WEAVR: Aspect Orientation and Model-

Driven Engineering. Journal of Object Technology,

6(7):51–88. Chair of Software Engineering, ETH

Zurich, Switzerland.

Davie, A. (2002). Intelligent Tagging for Transport and

Logistics: The ParcelCall Approach. Electronics &

Communication Engineering Journal, 14(3):122–128.

Institution of Electrical Engineers, London, UK.

Fuentes, L., Gamez, N., and Sanchez, P. (2008). Aspect-

Oriented Executable UML Models for Context-Aware

Pervasive Applications. In Proc. 2008 5th Interna-

tional Workshop on Model-Based Methodologies for

Pervasive and Embedded Software (MOMPES’08),

pages 34–43, Budapest, Hungary. IEEE.

Groher, I. and Schulze, S. (2003). Generating Aspect

Code from UML Models. In Proc. 3rd Interna-

tional Workshop on Aspect-Oriented Modeling co-

located with 2nd International Conference on Aspect-

Oriented Software Development (AOSD’03), Boston,

USA.

Hegering, H.-G., K

upper, A., Linnhoff-Popien, C., and

Reiser, H. (2003). Management Challenges of

Context-Aware Services in Ubiquitous Environments.

In Self-Managing Distributed Systems, volume 2867

of Lecture Notes in Computer Science, pages 321–

339. Springer Berlin / Heidelberg.

Kappel, G., Proll, B., Retschitzegger, W., and Schwinger,

W. (2003). Customisation for Ubiquitous Web Appli-

cations: A Comparison of Approaches. International

Journal of Web Engineering and Technology, 1(1):79–

111. Inderscience Publishers, Geneva, Switzerland.

Magee, J. and Kramer, J. (2006). Concurrency: State Mod-

els and Java Programs. John Wiley and Sons, second

edition.

Mandato, D., Kovacs, E., Hohl, F., and Amir-Alikhani, H.

(2002). CAMP: a Context-Aware Mobile Portal. IEEE

Communications Magazine, 40(1):90–97. IEEE.

Mostefaoui, S. K. and Hirsbrunner, B. (2004). Context-

Aware Service Provisioning. In Proc. IEEE/ACS

International Conference on Pervasive Services

(ICPS’04), pages 71–80, Beirut, Lebanon. IEEE.

Schwinger, W., Gr

un, C., Pr

oll, B., Retschitzegger, W.,

and Schauerhuber, A. (2005). Context-Awareness in

Mobile Tourism Guides - A Comprehensive Survey.

Technical report, Johannes Kepler University, Linz,

Austria.

VIDE (2009). VIsualize all moDel drivEn pro-

gramming (VIDE), WP 11, Deliverable number

D11.3 (European Commission supported Speciﬁc

Targeted Research Project, Information Society

Technologies). WWW page. http://www.vide-

ist.eu/download/VIDE D11.3.pdf (Last accessed on

02/04/2011).

Whittle, J. and Jayaraman, P. (2008). MATA: A Tool for

Aspect-Oriented Modeling based on Graph Transfor-

mation. In Models in Software Engineering, volume

5002 of Lecture Notes in Computer Science, pages

16–27. Springer Berlin / Heidelberg.

AN EVALUATION FRAMEWORK FOR VALIDATING ASPECTUAL PERVASIVE SOFTWARE SERVICES