Authors:
Norman Ahmed
1
and
Bharat Bhargava
2
Affiliations:
1
Purdue University and Air Force Research Laboratory/RIS, United States
;
2
Purdue University, United States
Keyword(s):
Service Oriented Architecture, Web Services, Aspect Oriented Programming, Web Services Security, Cloud Computing, Quality of Service.
Related
Ontology
Subjects/Areas/Topics:
Cloud Computing
;
Fundamentals
;
QoS for Applications on Clouds
Abstract:
Service Oriented Architecture (SOA) is an architectural style that provides agility to align technical solutions to modular business Web Services (WS) that are well decoupled from their consumers. This agility is established by interconnecting WS family of standards specification protocols (commonly referred to as WS-* (WS-star)) to enable security, ease of service interoperability and orchestration complexities when extending services across organizational boundaries. While orchestrating services or chaining services in varying ways to satisfy different business needs, on highly scalable cloud platforms is undeniably useful, it is increasingly challenging to effectively monitor Quality of Service (QoS), especially, service response time. This is due to a) lack of proper formulation of the WS-star interconnections mechanisms, and b) the transient performance behaviour intrinsic to the heterogeneity of the hardware and shared virtualized network and IO resources built on the cloud pla
tforms. We present an analysis of WS-star standards, classifying and discussing their inter-dependencies to provide a basis for QoS monitoring context on protocol formulation. We then illustrate a practical implementation of a dynamic QoS monitoring mechanism using runtime service instrumentation with Aspect Oriented Programming (AOP). Preliminary evaluations show the efficiency of computing QoS on a transient performance cloud platform.
(More)