Cross-layer Service Adaptation
State-of-the-Art, Shortcoming Analysis, and Future Research Directions
Ameni Meskini
1
,
Yehia Taher
2
, Rafiqul Haque
3
and Yahya Slimani
1
1
University of Carthage, INSAT, LISI research Laboratory, Tunis, Tunisia
2
Laboratoire PRiSM, Universite de Versailles/Saint-Quentin-en-Yvelines, Versailles, France
3
Laboratoire d'InfoRmatique en Image et Systèmes d'information, Université Claude Bernard Lyon 1, Lyon, France
Keywords: Web Service, Service based Application, Service Adaptation, Cross-layer Adaptation.
Abstract: In the past few years several cross-layer monitoring and adaptation technologies have been proposed.
Although these are cross-layer adaptation technologies, however, in practice they focus on a particular layer.
Some solutions involves two layers, yet none of the existing solutions do not consider all the layers during
adaptation process. Furthermore, cross-layer adaptation approaches generate incompatibility problems. This
is an adaptation coordination problem. Incompatibility refers to the situations where the adaptation is
performed in a layer is not compatible with the constraints exposed by the other layers. This survey aims at
studying and analyzing current approaches for web services adaptation, discussing their shortcomings and
proposing research directions on cross-layer web service adaptation.
1 INTRODUCTION
Service adaptation has drawn enormous research
interests in the area of Service Oriented Computing
(SOC) (Geihs et al, 2009). Adaptation from the
functional point of view can be defined as an ability
of a Service Based Application (SBA) (Bucchiarone
et al, 2009) to adapt changes or requirements that are
needed to guarantee fault-tolerance or to
optimize system performances.
While developing an SBA, it may not be possible
to capture all functional and nonfunctional
requirements because many times the requirements
evolve at runtime. Typically, an application is able
to carry out the operations (at runtime) that are
studied and documented during requirement
analysis. The unprecedented requirements that are
evolved at runtime may lead to failure. In other
words, applications are unable to perform the
operations that have not been realized.
This limitation gave rise to the notion of service
adaptation. Many new techniques such as service
replacement or adding new service have been
proposed to build adaptive SBAs. However, there
are several challenges that cannot be dealt with
efficiently by the existing techniques. One major
challenge is handling the impact of adaptation
operations. For instance, service replacement at
different layers of SBAs. It is worth noting that an
SBA has different layers (Papazoglou et al, 2008).
The notion of multi-layer SBAs relies on the Service
Oriented Architecture (SOA) (Liu et al, 2011)
paradigm. These layers interact with each other.
Thus, if a service is adapted in one layer (e.g.,
BPM layer), it may affect the other layers (e.g.,
orchestration layer). This promotes the notion of
cross-layer adaptation which is the main focus of
this paper. Cross-layer adaptation is a process of
adapting a service in different layers of SBAs. It
promotes configuration challenges. Several
techniques have been proposed to tackle these
challenges. This paper aims to investigate all
existing technologies, methodologies, and
techniques related to cross-layer adaptation. It
presents a comprehensive review of the state-of-the-
art, summarizes their strengths and weaknesses, and
identifies future research direction in this area.
This paper is organized as follows. In Section 2,
we present the review of the state-of-the art. We
discuss our findings in Section 3. A conclusion is
drawn in Section 4.
260
Meskini A., Taher Y., Haque R. and Slimani Y..
Cross-layer Service Adaptation - State-of-the-Art, Shortcoming Analysis, and Future Research Directions.
DOI: 10.5220/0005435702600267
In Proceedings of the 5th International Conference on Cloud Computing and Services Science (CLOSER-2015), pages 260-267
ISBN: 978-989-758-104-5
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)
2 SERVICE ADAPTATION
APPROACHES
Although our focus in this paper is cross-layer
adaptation, we cover all the existing technologies
related to service adaptation. The purpose of this
study is to provide a comprehensive understanding
of the strength of service adaptation and also to
outline the limitations why traditional adaptation
technologies are unable to assist in cross-layer
adaptation. It is worth noting that although our study
mainly covers the service based systems, we try to
cover adaptation in the agent based systems as well.
2.1 Interaction-based Adaptation
Approaches
Interaction-based adaptation approaches deals with
interactions between web services such as, actions
required to mediate communications between web
services, and adapting the compositions of web
services in case of failure of a component. Unlike
the approaches that focus on QoS adaptation (of
instance composition), interaction-based adaptation
approaches are concerned with the changes and the
adaptation of interactions within a service
composition.
For example, re-engineering services to ensure
that they can integrate new components by
guaranteeing interoperability. Sometime adaptation
operations such as substitute to repair or to optimize
QoS are not sufficient for efficient adaptation. The
main reason is the emergence of new requirements
and additional constraints which may not be possible
to handle efficiently by he selected services. For
instance, a service may not be able to meet user
needs or may fail to handle heterogeneity of the
interfaces between services or communication
protocols. The objective of adaptation in this case is
not only to manage the QoS adaptation, but also to
ensure that the adaptation measures do not lead to
interaction failures. The composition and the
mediation are the most common solutions in such
situations.
In interactions-based adaptation, existing
approaches realize exchanging messages in service
compositions based on pre-defined policies such as
the policies proposed in (Baresi et al, 2007). This
involves the business processes which are essentially
composition of services. We found several services
composition-based adaptation approaches which we
discuss in this section. Baresi et al. (Baresi et al,
2007) address the problem of substitution of services
and the dynamic binding of the service providers in
order to repair failures. Their work targets the
adaptation of the workflows (defined using BPEL
(IBM, BEA Systems, Microsoft, SAP AG, Siebel
Systems, 2003) at runtime to select between
available alternatives based on nonfunctional
requirements, or to retry a service following in the
first choice. To enable the deployment and the
reconfiguration of service compositions during its
execution, the authors used a specification of BPEL
process which is enriched by a set of rules and
constraints for the discovery or dynamic service
binding until the time to execute.
The choice depends on the criteria defined by the
user during the establishment process. The proposed
framework can also exchange services based on the
events collected during the monitoring phase. It
relies on three actors: the registry service (DIRE)
(Baresi et al, 2007) that can be distributed between
the service providers, the runtime environment
(SCENE) (Baresi et al, 2007) with the rules of
discovery and binding, and monitoring features
(Dynamo) (Baresi et al, 2005) and (Baresi et al,
2007) that produce events to reconfigure the
processes.
The main limitation of this approach is the web
service composition language. The authors in
(Ardagna et al, 2007) propose an implicit approach
for adaptive composition of services within the
flexible processes. This approach is implemented in
business process management layer. The main
objective is to select the best set of services available
at run-time by taking the constraints of business
process, users preferences, and execution contexts
into account.
The authors introduce a new approach to model
the problem of service selection. This approach is
effective for large process and in the case of QoS
constraints are at extreme. In the proposed model,
the problem of service selection is formalized as a
mixed problem of linear programming, the loop
peeling is adopted in optimization, and the
constraints posed by the stateful web services are
considered.
2.2 Mediation-based Adaptation
Approaches
While composing interactions, services may
encounter heterogeneity problems. For instance,
interaction types can be different, incompatible
communication protocols; different semantics of
interactions promote the heterogeneity problem.
These problems may occur in different steps of
Cross-layerServiceAdaptation-State-of-the-Art,ShortcomingAnalysis,andFutureResearchDirections
261
composition. Also, the problems may occur while
adaptation actions carried out such as, during
substitution of a WS by another WS.
The solution of the heterogeneity problem is
called mediation which is critical to achieve
adaptation and composition of services. However,
additional mechanisms are needed for successful
interactions between web services and to perform
different adaptation actions (e.g., substitution, re-
selection, composition, etc.). Adaptation (in this
case “mediation) is an important functionality which
enables integration of business services. Generally
speaking, mediation resolves conflicts between two
actors. In the context of web services, mediation
aims to resolve heterogeneity between web services
in order to enable successful interactions (Chafle et
al., 2006).
One needs to generate a service that ensures
interactions between the two services with two
signatures, different protocols or interfaces, in order
to guarantee interoperability. The requirements of an
adaptation in these approaches stem from two
sources: (i) the level of heterogeneity in the upper
stack of interoperability (e.g., business level,
infrastructure protocols.), and (ii) the diversity of
customers, each one of them supports different
protocols and interfaces. Mediation can be automatic
(Williams et al., 2006) or semi-automatic (Reza et
al, 2007).
Taher et al, 2009 (Taher et al, 2009) propose a
multilayer software architecture. They propose a
framework for transparent and flexible substitution
of a service provider by another with respect to an
existed consumer. A framework for automatic
generation of adapters and service interfaces
modelling using automata was adopted to solve the
problem of incompatibility in the interaction
between two services: a consumer and a new
provider. If incompatibilities between these services
are detected, an adapter is generated automatically
based on the incompatibilities. The generation of the
adapter relies on the automata model. The generated
adapter contains a sufficient detail of the projected
technology called CEP (Complex Event Processing)
engine (Luckham et al, 2001).
However, unfortunately, the complex
incompatibilities were not considered in this tool.
For example, the implementation of several different
operations of customer service and a supplier service
is not possible by this tool.
The solution proposed in (Hau, 2003) uses OWL
(Dean, 2002) to annotate interfaces too. Both
solutions (proposed by Syu and Hau) have an
abstraction layer called meta-data space. Semantic
annotation is used to describe the methods of
services.
Meta-services use these annotations to find
appropriate matches between needs and
implementations. These solutions differ from the
other adaptation approaches. Two distinguished
aspects of these approaches are as follows:
Their locations are dependent on
architectures in which they are embedded, and an
adoption concerning with the interfaces of web
service is often the responsibility of the service
provider.
These approaches are platform dependent
such as they are dependent on languages
and composition engines
2.3 Cross-layer Adaptation
Approaches
The cross-layer adaptation refers to a process of
adapting a general system consisting of several
layers, where the technology and processes of each
layer are integrated and controlled by the same
adapter frame. In the context of SOA, this denotes a
consistent adaptation through the service interface of
different layers and applying a SOA system while
maintaining the characteristics such as loose
coupling and service autonomy.
The problem of monitoring and adaptation of
different types of software systems has gained
interests in both the research community and
industry. In recent years, these issues have promoted
interest in the area of SOA. However, the results and
directions are still insufficient. One of the key issues
here is that the proposed approaches are very
fragmented. They deal only with the problems which
are specific to a particular aspect of web service and
a particular functional layer, such as business
process management layer, service composition and
coordination layer, or service infrastructure layer.
However, the implementation of various layers of
web service can be nested in different artifacts. A
layer may contain objects that reside in another
layer. However, such cases are ignored by traditional
monitoring and adaptation solutions.
Consequently, there is a possibility that these
solutions will detect the problems incorrectly which
will lead to inaccurate decisions concerning
adaptation. This shortcoming of existing solution
promotes the need of cross-layer adaptation. In this
section, we study the most recent solutions which
have been proposed to provide a monitoring and
adaptation tools that covers multiple layers. We
found that in these solutions, controlling and
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
262
adaptation are developed by using various
techniques such as, monitoring and event logging,
detecting the patterns of events, and correlation and
mapping between events and appropriate adaptation
strategies, etc.. The solutions proposed in (Gjrven et
al, 2008), (Popescu et al., 2010), (Popescu et al.,
2012), (Zengin et al., 2011), (Zengin et al., 2011)
and (Zeginis et al. 2011) are based on the situation-
action mechanism. The situations correspond to a set
of events and disparities while the actions are
defined as templates for adaptation. These
approaches combine the taxonomies of adaptation
problems and mechanisms based on the events for
guiding the selection process of the adaptation
models based on the degree of correspondence
between events and disparities of adaptation.
In (Gjrven et al, 2008), a middleware called QuA
is presented that provides a multilayer adaptation
coordinated by incorporating multiple mechanisms
of adaptation in the interface and application layers.
However, the proposed middleware is lacking
theflexibility because the adaptation logic is
predefined and static. A multi-layer adaptation
framework is proposed in (Popescu et al., 2010). The
authors use taxonomy and adaptation models
(patterns) which are created during the design phase
to represent the possible solutions to adaptation
problems. In this framework, they designed adaptive
predefined templates to provide a means for
dynamic multi-layers adaptation.
These models define the behaviour of the
adaptation processes. However, this approach does
not consider the infrastructure layer and the authors
do not provide the mechanism for detection
disparities. An adaptation manager called CLAM is
proposed in (Zengin et al., 2011) to handle adaptive
inter-layer and multilayer problems. The authors
have classified a group of adaptation paths of an
adaptation tree which can be built in any layer of
SBAs. The limitation of these approaches is the
execution control which is performed in an isolated
manner. This does not allow an effective analysis of
monitoring data and detected events because events
are analysed and processed independently of each
other and the critical information are not propagated
between layers. This can lead to an incorrect
identification of the original source of the problems.
Also, some approaches do not realize monitoring in
all the layers which affects the final step of
adaptations. For example, the actual problem can
occur in the infrastructure layer, while it is detected
in the composition layer and therefore, it cannot be
properly diagnosed.
Additionally, in (Guinea et al, 2011), it is also
argued that monitoring of the web services is not
sufficient to allow proper and effective adaptation at
runtime. The authors present a framework which
uses various techniques for monitoring different
layers. Also, it uses a centralized agent of adaptation
to collect the events and analyse the violations of
KPIs.
Although the cross-layer adaptation approaches
designed to identify the sources of problems through
analysis and diagnosis that take several layers into
account, the works presented in this section have
some limitations. Based on our analysis, these
approaches can be improved to be more efficient.
For instance, since the adaptation approaches do not
consider the characteristics and requirements of all
the layers of SBAs rather they focus on a specific
layer, the activities of adaptation may fail to achieve
the desired effects. Furthermore, these approaches
may lead to incompatibility problems.
2.4 Adaptation in Agent based Systems
From architectural point of view, there is a similarity
between agent and service based systems. This is
one of the main reasons we studied the adaptation
solutions proposed in this domain. The notion of
agent based system is relatively new. We found a
few research works on adaptive agent based system.
Qureshi and Perini (Nauman et al., 2008)
proposed a methodology called TProcess for
seamless self-adaptation in agent based system. The
methodology is shaped a triangle that includes three
elements include requirement-time, design-time
time, and runtime. The authors argue that adaptation
should be built on the top of these mutual dependent
elements. The critical components of TProcess are
goal models which are defined at requirement-time
step. The goal models contain QoS parameters, their
values and conditions. These are mapped to the
implementation platform in the design-time step.
In (Bernon et al., 2003), the author proposed a
methodology called ADELFE to guide developers to
develop adaptive multi-agent systems. The
methodology is based on object-oriented
methodologies, follows rational unified process and
uses Unified Modeling. In (Ibrahim, 2004), the
author proposed a framework for developing
intelligent adaptive agents.
In the proposed framework, the agents are
defined as systems or machines that utilize
inferential or complex computational methodologies
to modify or change control parameters, knowledge
bases, task plans, problem-solving, methodologies,
course of actions, or other objects in order to
Cross-layerServiceAdaptation-State-of-the-Art,ShortcomingAnalysis,andFutureResearchDirections
263
successfully accomplish a set of tasks that are of
interest to the user. The intelligent adaptive agents
are classified into three based on the agent’s
capabilities on performing external and internal.
These categories are listed below:
Internal adaptation: In this criterion, the
internal systems of the agent are adaptive;
however, its external actions do not reflect
adaptive behaviour.
External adaptation: It is simply the
opposite of internal adaptation. In this the
internal systems of agents do not reflect
adaptive behaviour.
Complete adaptation: Internal systems are
adaptive and external actions reflect
adaptive behavior.
There are a few significant differences between
adaptive SBAs and adaptive agents. In SBAs,
adaptations are performed in different layers, as
these applications rely on multilayer architecture.
However, multilayer adaptation is of the scope of
agent based systems. Additionally, none of the
adaptive agent based solutions is aware of cross-
layer adaption. However, evidently, the service
based systems can be benefited by using the
approaches used in adaptive agent based systems.
Particularly, the notion of context-awareness and
self-adaption can be efficacious for adaptive service
based systems.
3 ANALYSIS AND RESEARCH
DIRECTIONS
In this section, we summarise our findings and
propose a few potential extensions specifically in the
area of cross-layer adaptation. We studied various
solutions published in the literature. It is worth
noting that in this section we limit our discussion in
the context of service based systems which is the
main focus of this study.
3.1 Analysis
We studied different research initiatives that focus
on adaptation problems concerning service
interaction in the service composition layer.
Specifically, we studied the heterogeneity problems
regarding interactions which can be found in the
service interface layer. The heterogeneity problem
may lead to inconsistency with respect to data
exchanged between the services. We found that the
main reason for heterogeneity problem is different
formats of the messages exchanged between
services. For an effective and adaptation
heterogeneity between web services must be dealt
with efficiently.
We found mediation-based adaptation
approaches deals with heterogeneity. They enable
exchanging consistent data between Web services.
However, these approaches have limitations. They
lack of flexibility and the automation needs to be
efficient for a complete and effective adaptation.
Moreover, they are limited to technical and
structural aspects of a system. They do not cover
other aspects. In addition, due to the highly dynamic
and evolving nature of the environment and different
requirements of service users (infrastructure
protocols, and behavior), a manual intervention is
required, especially to define the management tasks
to handle disparities or to specify or adjust the
composition diagram. This is certainly a limitation
to carry out adaptation operations efficiently.
In addition, the adaptation mechanisms are not
rich enough and deals only with the specific
adaptation situations and actions, which does not
cover multiple anomalies that may occur in
execution environments. The cross-layer adaptation
approaches are fragmented and isolated. They do not
consider the effects of changes and modifications on
all the functional layers of the SBAs. The existing
cross-layer, adaptation solutions are designed to
adapt a particular functional layer, namely, the
business layer, the service composition layer, or
infrastructure layer. The realization of different
layers of web service can be nested such as different
artifacts of a layer can refer to the same objects
reside in another layer, while these relationships are
ignored by the current monitoring and cross-layer
adaptation solutions.
Also, these mechanisms are designed to support
quality assurance for adaptation. They deal with the
analysis of adaptation activities against the system
model, and adaptation measures. Table 1 presents a
synthetic summary of the cross-layer adaptation
solutions which we studied in this paper. We
consider three factors, defined by (Reza et al.2007)
adaptation objectives that involves adaptation
requirements (repair, optimization, mediation, etc..),
adaptation methodology, and the layers covered by
the solutions. Also, these mechanisms are designed
to support quality assurance for adaptation.
They deal with the analysis of adaptation
activities against the system model, adaptation
measures, and other adaptations.
Table 1 presents a summary of the approaches
found in the literature. We consider three factors
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
264
Table 1: Classification of cross-layer adaptation approaches.
Approach Adaptation
Objectives
Methodology of
adaptation
Layer affected
( Reza et al, 2007) Fault tolerance Proactive BPM, SCC
(Popescu et al., 2010) Mediation Reactive BPM
(Popescu et al., 2011) Reparation Reactive BPM
(Guinea et al, 2011) Reparation Reactive BPM, SI
(Mos et al. 2009) Monitoring Reactive SI
(Schmieders et al., 2011) Reparation Reactive SCC, SI
(Vidackovic et al., 2009) Optimistaion Reactive BPM
(Gjrven et al, 2008) Configuration Proactive BPM, SCC
(Syu et al, 2004) Mediation Reactive SI
defined by (Reza et al. 2007): (i) adaptation
objectives involves adaptation requirements (repair,
optimization, mediation, etc.., (ii) Adaptation
methodology, and (iii) affected SBA layers which
concerns with the change of locations and adaptation
progress. From the comparison (shown in the above
table) we conclude that none of the current
approaches cover all the layers of service based
systems. The solutions proposed by Reza et al.,
Guinea et al., Schmieders et al., Gjerven et al. are
relatively more efficient as they cover two layers.
However, cross-layer adaptation solution must
cover all three layers of SBAs to deal with various
runtime challenges efficient that evolve in current
service based system such as cloud service based
applications. Remarkably, most of these approaches
cover BPM layer, however, to the best of our
understanding if an event adapted in the BPM layer,
yet it the adaptation has not been propagated to the
bottom layers implies that the adaptation has not be
realized automatically and may not have done
efficiently. This is an important limitation. The
current solutions focus on specific layers (e.g.,
infrastructure layer or Business Process
Management layer). One might think of building a
hybrid solution which can combine two or more of
the existing solutions. However, it will promote a
huge complexity. Developing a hybrid solution
needs a list of complex tasks include the following:
Analysis of the affected layer,
Identification of adaptation actions,
Aggregation of these actions to check their
effects on different layers,
Launching a coordination system to
coordinate adaptation actions,
Checking whether the adjustment
performed at one layer is compatible with
the constraints posed by other layers, etc..)
which can be costly in terms of response
time.
3.2 Research Directions
We identified four critical aspects: context
awareness, self-adaptation, completeness,
performance, which should be focused in the topic
of cross layer adaptation.
Context aware adaptation and self-adaptation
have already been studied in agent oriented system.
It is worth noting that context awareness and self-
adaptation are complementary because self-adaptive
system should be aware of the context. Otherwise,
self-adaptation can be difficult.
The Table 1 shown in the previous section
unearthed a very important shortcoming of cross
layer service adaptation technologies. Although
these technologies are known as cross-layer
adaptation solution, to the best of our understanding,
these solutions are complete. These approaches lack
the ability to trace incompatibilities that can be
triggered through adaptation. Therefore, a solution is
needed which can create adaptation loop which runs
adaptation process until new requirements or
changes are adapted by resolving incompatibilities
or conflicts. Adaptation promotes performance
challenge. In other words, the system performance
can be challenged enormously by adaptation. We
found literature reported trade-off between
adaptation and performance. An extensive research
is necessary to develop a solution that can process
adaptation by guaranteeing high efficiency (with
respect to processing time).
We plan to develop an intelligent and fault-
tolerant solution for cross-layer adaptation that can
address the requirements discussed in the above. The
proposed solution will enable to perform adaptation
process by guaranteeing efficiency and
effectiveness. It will be able to perform adaptation in
all the layers of SBAs without any incompatibilities
or conflicts. The solution will be context-aware and
will support self-adaptiveness. This will ensure the
autonomic execution of adaptation operations across
Cross-layerServiceAdaptation-State-of-the-Art,ShortcomingAnalysis,andFutureResearchDirections
265
the SBA layers.
We strongly believe that the genetic algorithms
are potential for our solution especially to optimize
the adaptation process. Genetic algorithms are
widely used to handle cases such as requirement
evolution and performance optimization which are
the two most critical issues.
4 CONCLUSIONS
In this paper we studied adaptation technologies
particularly the cross-layer adaptation technologies.
We discussed the outcomes of our analysis. In
particular, we discussed the limitations of different
approaches of cross-layer service adaptations.
The major limitation we found is the lack of
coordination between adaptation activities that may
lead to conflicts or incompatibilities. According to
our study, the current solutions do not consider the
fact that adaptation in a layer may affect adversely
the other layers of service based systems. According
to our study, current cross-layer adaptation
approaches lack efficient coordination which leads
to conflict and incompatibilities. We believe that
these problem must be addressed for an efficient
cross-layer service adaptation. We presented the
results of a brief study on adaptive agent based
systems. We found in our study that the agent based
adaptive systems have some advanced , features
such as context-awareness, self-adaptation, etc.. The
adaptive SBAs can be benefited by these features
especially, the service based adaptive systems can be
more intelligent and autonomous.
Additionally, based on our understanding we
presented some research directions in the area of
cross layer service adaptations. We strongly believe
that the research in this area should focus on context
awareness, self-adaptation, and performance etc. to
develop highly high-performance solutions. We also
presented a proposal of a solution which are
currently working on.
There are a few limitations of our study. Firstly,
this is merely a literature review. However, the state
of the art could be better reviewed or understood by
benchmarking the existing solutions. A comparison
of adaption technologies in different contexts can be
done by following a set of rigorous protocols. This
paper is missing such an comparison. In our future
work, we plan to conduct an empirical study with
the current cross-layer adaptation technologies.
Also, we plan to conduct a study by covering more
contexts.
REFERENCES
Ardagna, D. , Pernici, B.(2007). Adaptive Service Com-
position in Flexible Processes, IEEE Transactions on
Software Engineering, vol. 33, no. 6, pp. 369-384,
June 2007.
Baresi , L. , Guinea, S. , (2005), Dynamo: Dynamic
Monitoring of WS-BPEL Processes. In 5th
International Conference on Service Oriented
Computing, pages 478–483, 2005.
Baresi, L., Di Nitto, E. , Ghezzi, C., and Guinea, S. ,
(2007) . A Framework for the Deployment of
Adaptable Web Service Compositions, Service
Oriented Computing and Applications, vol. 1, no. 1,
pp.75-91, 2007.
Bernon, C., Gleizes, M. P., Peyruqueou, S., Picard, G.
(2003). ADELFE: a methodology for adaptive multi-
agent systems engineering. In Engineering Societies in
the Agents World III (pp. 156-169). Springer Berlin
Heidelberg.
Bucchiarone, A., Cappiello, C., Di Nitto, E.,
Kazhamiakin, R., Mazza, V., and Pistore, M., (2009).
Design for Adaptation of Service-Based Applications:
Main Issues and Requirements ,ICSOC/ServiceWave
in page 467–476.
Chafle, G., Dasgupta, K. , Kumar, A. , Mittal, S.,and
Srivastava, B. , (2006). Adaptation in Web Service
Composition and Execution, Proc. IEEE Int'l Conf.
Web Services (ICWS '06), pp. 549-557, 2006.
Dean, M., Connolly, D., Harmelen, F., Hendler, J.,
Horrocks, I., Debo-rah L., McGuinness, Peter F.
Patel-Schneider, Andrea Stein, L., (2002). Web
ontology language (OWL) reference version 1.0.
Technical report, www.w3c.org, 2002.
Geihs, K., Reichle, R., Wagner, M., Khan ,M., (2009).
Service-Oriented Adaptation in Ubiquitous
Computing Environments , International Conference
on Computational Science and Engineering.
Gjrven, E., Rouvoy, R., Eliassen, F. , (2008). Cross-layer
self-adaptation of service-oriented architectures,
Proceedings of the 3rd workshop on Middleware for
service oriented computing 2008.
Guinea, S. , Kecskemeti, G. , Marconi, A. , Wetzstein, B. ,
( 2011). Multi-layered monitoring and adaptation,
Hau, J., Lee, W. , Newhouse, S. , (2003) . The ICENI
Semantic Service Adaptation Framework. In: UK e-
Science All Hands Meeting (2003).
Ibrahim F. Imam: Adaptive applications of intelligent
agents. ISCC 2004: 7-12.
IBM, BEA Systems, Microsoft, SAP AG, Siebel Systems,
(2003). Business Process Execution Language for
Web Services version 1.1.
http://www.ibm.com/developerworks/library/specificat
ion/wsbpel/. 2004 Imam, I. F. (2004, June). Adaptive
applications of intelligent agents. In Computers and
Communications, 2004. Proceedings. ISCC 2004.
Ninth International Symposium on (Vol. 1, pp. 7-12).
IEEE.
Liu, F., Tong, J., Mao, J., Bohn, R., Messina, J., Badger,
L., and Leaf, D., (2011). NIST Cloud Computing
CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience
266
Reference Architecture: Recommendations of the
National Institute of Standards and Technology. NIST
Special Publication 500-292. pp. 10.
Luckham, D., (2001), The Power of Events: An
Introduction to Complex Event Processing in
Distributed Enterprise Systems. Addison-Wesley
Longman (2001).
Nauman,A., Qureshi, Anna Perini:
An Agent-Based Middleware for Adaptive
Systems. QSIC 2008: 423-428.
Papazoglou, M. P., (2008) . Web Services - Principles and
Technology. Prentice Hall. ISBN 978-0-321-15555-9.
pp. 1-752.
Popescu, R. , Staikopoulos, A. , Liu, P., Brogi, A. ,
Clarke, S. , ( 2010) . Taxonomy-Driven Adaptation of
Multilayer Applications Using Templates,” saso,
pp.213-222, 2010 Fourth IEEE International
Conference on Self-Adaptive and Self-Organizing
Systems, 2010.
Popescu, R., Staikopoulos, A. , Liu, P. , Brogi, A. , Clarke.
S. (2011). A Formalised, TaxonomyDriven Approach
to Cross-Layer Application Adaptation ACM
Transactions on Autonomous and Adaptive Systems,
ICSOC.
Popescu, R., Staikopoulos, A. , Liu, P. , Brogi, A. ,
Clarke. S., 2012. A Formalised, Taxonomy-Driven
Approach to Cross-Layer Application Adaptation” in
ACM Transactions on Autonomous and Adaptive
Systems, Proceedings of the 9th international
conference on Service-Oriented Computing,
December 05-08, 2011, Paphos, Cyprus.
Reza, H. , Nezhad , M. , Benatallah , B. , Martens , A. ,
Curbera , F. , Casati, F. , ( 2007). Semi-automated
adaptation of service interactions, Proceedings of the
16th international conference on World Wide Web,
May 08-12, 2007, Banff, Alberta, Canada.
Schmieders, E., Micsik, A., Oriol, M., Mahbub, K., and
Kazhamiakin R., “Combining SLA prediction and
cross layer adaptation for preventing SLA violations”.
In Proceedings of the 2nd Workshop on Software
Services: Cloud Computing and Applications based on
Software Services, Timisoara, Romania, June 2011.
Syu, J.-Y., (2004). An Ontology-Based Approach to
Automatic Adaptation of Web Services”, Department
of Information Management National Taiwan
University, 2004. (http://www.im.ntu.edu.tw/
IM/Theses/r92/R91725051.pdf).
Taher, C., Aït-Bachir, .A, Fauvet, .M, Benslimane,
.M,:Diagnosing Incompatibilities in Web Service
Interactions for Automatic Generation of
Adapters. AINA 2009: 652-659.
Vidackovic, K., Weiner, N., Kett, H., Renner, T.
“Towards business-oriented monitoring and adaptation
of distributed service-based applications from a
process owner's viewpoint”. In: ICSOC/ServiceWave
Workshops. pp. 385394, 2009.
Williams , S.K., Battle , S. A., Cuadrado, J. E. , (2006).
Protocol mediation for adaptation in semantic web
services, Proceedings of the 3rd European conference
on The Semantic Web: research and applications, June
11-14, 2006, Budva, Montenegro.
Zeginis, C., Konsolaki, K. , Kritikos, K. , and Plexousakis,
D. , (2011). Ecmaf: An event-based cross-layer service
monitoring and adaptation framework, In 5th
Workshop on Non-Functional Properties and SLA
Management in Service-Oriented Computing
(NFPSLAM-SOC’11) co-located with ICSOC 2011.
Springer, 2011.
Zeginis, C., Plexousakis, D., (2010). Web Service
Adaptation: State of the art and Research Challenges,
Technical Report 410, ICS-FORTH, October 2010.
Zengin, A. , Marconi, A. , Baresi, L. , Pistore, M. , (
2011). CLAM: Managing cross-layer adaptation in
service based systems, soca, pp.1-8, 2011 IEEE
International Conference on Service-Oriented
Computing.
Zengin, A., Kazhamiakin, R. , Pistore, M.: “CLAM:
Cross-Layer Management of Adaptation Decisions for
Service-Based Applications,” icws, pp.698-699, 2011
IEEE International Conference on Web Services,
2011.
Cross-layerServiceAdaptation-State-of-the-Art,ShortcomingAnalysis,andFutureResearchDirections
267
