Information Technology Reference
In-Depth Information
Bottom-Up Fault Management
in Composite Web Services
Brahim Medjahed
1
and Zaki Malik
2
1
Department of Computer and Information Science,
University of Michigan - Dearborn,
brahim@umich.edu
2
Department of Computer Science,
Wayne State University
zaki@wayne.edu
Abstract.
We propose an approach for managing bottom-up faults in composite
Web services. We define bottom-up faults as abnormal conditions/defects or
changes in component services that may lead to run-time failures in composite
services. The proposed approach uses soft-state signaling to propagate faults
from components to composite services. Soft-state denotes a class of protocols
where state (e.g., whether a service is alive) is constantly refreshed by periodic
messages. Its advantages include implicit error recovery and easier fault man-
agement, resulting in high availability. We introduce a bottom-up fault model
for composite services. Then, we propose a soft-state protocol for bottom-up
fault propagation in composite services. Finally, we present experiments to
assess the performance of our approach.
Keywords:
Service Composition - Fault Management - Bottom-up Fault -
Soft-State - Fault Coordinator.
1 Introduction
Service-oriented architecture (SOA) has recently emerged as a promising approach
for application integration [1,14]. It utilizes services (commonly Web services) as the
building blocks for developing software systems distributed within and across organi-
zations. The primary value of SOA is the ability to (1) reuse pre-developed, autono-
mous, and independently provided resources (e.g., legacy applications, sensors, data-
bases, storage devices, COTS products) as Web services and (2) combine pre-existing
services, called
participants
, into higher level services, called
composite services
,
which perform more complex functions [9,15].
Because of the dynamic and volatile nature of SOAs, Web services are subject to
unavoidable faults during their lifetime. The relationship between a fault and failure
is given in ISO/CD 10303-226 document, where a fault is defined as an abnormal
condition or defect at the component, equipment, or sub-system level which may lead
to a failure. In their seminal work, Avizienis et al. [2] state that faults cycle between
dormant and active states, and a failure occurs when a fault becomes active. In this
