Active Customization-oriented Adaptive SOA with
Runtime Exception Handling
Bin Wen, Ziqiang Luo
School of Information Science and Technology,
Hainan Normal University,
Haikou, China
Email: binwenwebb@gmail.com
Yazhi Wen
School of Computer Science and Technology,
Huazhong University of Science and Technology,
Wuhan, China
Email: wyzhust@163.com
Abstract—Service has become the basic way to access and
enlarge the capability of all kinds of infrastructure. But now,
service-based software system based on SOA lacks of runtime
iterative re-aggregation process management and exception han-
dling. And current approaches can’t align services aggregation
with requirements to overcome the runtime exceptions and deal
with services resource insufficiency and context changes. This pa-
per concentrates on the adaptability of service-based system, i.e.
runtime exception handling mechanism driven by personalized
requirements. The following progresses have been achieved. (1)
Service requester-centric adaptive SOA with runtime exception
handling has been investigated. (2) Mathematical basis and
related algorithm for adaptive SOA have been built and verified.
Combination of theoretical research and empirical validation, the
paper tries to provide a technical operational and cost-effective
solution for SOA with adaptive mechanism and promoting the
adaptive ability for runtime requirements evolution.
Keywords—personalized customization; adaptive SOA; excep-
tion handling; runtime
I. INTRODUCTION
Now the general process for service-based software pro-
duction is as follow: producing services resource from ser-
vice providers→publishing services→selecting services from
service consumers→services aggregation (services resource
binding and combination on demand). It is similar to the
waterfall model or mixed alignment mode between top-down
and bottom-top approach. But the process lacks of runtime
exception handling and iterative services re-aggregation con-
sidering without starting from the requirements for services
system adaptive mechanism to overcome the exceptions, ser-
vices resource insufficiency and context changes.
Therefore, establishing the direct feedback channel for user
requirements, implementing SOA runtime adaptive adjustment
and improving the production for service-based system to
build adaptive SOA are the motivations of the studying and
significance.
A. The importance of support adaptive SOA
Resource is a very board concept. The difference of its
connotation and abstraction granularity is very big including
many other forms such as computing resource, network re-
source, storage resource, data resource and services resource.
Services resource presented as a set of well-defined and
interoperable Web software components. Its basic entity unit
can be understand as Web service, namely autonomous, open-
ness, self-descripted networked component. Service resource
provisioning refers to production mode, providing method and
source structure of services. If personalized services resource
does not exist, and also cannot be achieved through service
composition, then customized services resource production
problem based on captured personalized requirements will be
involved. How to adjust the runtime architecture to adapt to the
requirements and context changes? How to embody the cus-
tomized demand and instant production to deal with runtime
exception? How to let the personalized requirements segment
associated with the back-end services resource production?
This is what this paper focuses on the runtime self-adaptive
implementation mechanism for service-based system (SBS).
The importance of services resource supply mainly re-
flects in: (1) Adequate and personalized requirements satisfied
services resource is the material basis for services aggrega-
tion and software production. (2) Effective services resource
provisioning method insures successful and effective SBS
production.
Services resource provisioning has become the consensus
of SBS development for industrialization. It has been used in
software development and acquired partial support from the
mainstream tools [1, 2].
The current SOA is producer-centric style. The basic idea
lies in that the services provider release their services for
consumers to search available resource. But the user’s prefer-
ences and needs are continuously changing, it is very difficult
to build the application only relying on the current available
services resource.
Services computing and virtualization are two support tech-
nology of cloud computing. As the basis of cloud computing
and software production technology, services computing has
been partially resolve the description, registration, and assem-
bly of services. Now its knowledge body mainly includes
lifecycle planning, resource production, publishing, billing
and management of SBS. But for the ultimate purpose of
computing, i.e. on-demand service and its related problem,
it is lack of consideration and further research. Industry and
academic community generally focus on the services discovery
and services composition, the starting point of research is to
2016 13th Web Information Systems and Applications Conference
978-1-5090-5437-4/16 $31.00 © 2016 IEEE
DOI 10.1109/WISA.2016.1
103
2016 13th Web Information Systems and Applications Conference
978-1-5090-5437-4/16 $31.00 © 2016 IEEE
DOI 10.1109/WISA.2016.30
103