![](bg6.png)
deploy the client in a fifth virtual machine, and the
SHA in a sixth one. For this case, we choose to
centralize the SHA, which centralizes the interaction
management.
In order to test our framework, we inject delays
in the execution time of a Supplier Web service. The
Measurement module computes the new QoS values,
and informs the Diagnosis module about the system
state. The HMM-based diagnosis detects the QoS
degradation and sends a report to the Planner
Module, which generates a recovery plan.
Reconfiguration module will use the plan and
reroute next requests to the new Supplier thanks to
the proxy.
In our experiments, we fix K = 2 for two raisons.
The first is the Chebyshev Inequality (equation 6).
2
1
1]Pr[
KSRTTRTO
ii
−≥+<
σ
(6)
With K=2, we conclude that more than 75% of
responses are valid responses (non-timeout). The
second is on the basis on an already large scale
monitoring experiment on the French Grid’5000
(www.grid5000.fr). The statistical study of the
logged QoS values, shows that 96% of responses are
valid with K=2.
5 CONCLUSIONS AND FURTHER
WORK
In this paper, we presented a self-healing
architecture that manages QoS in a Web service-
based application. The presented approach relays on
different modules for monitoring, diagnosis and
recovering QoS degradation. We illustrate our
approach with the FoodShop application. The first
experiment is achieved while using virtual machines
and we are working on a large-scale experiment
under the Grid5000. We will focus on the
distribution of the SHA while centralizing and
coordinating the diagnosis and the recovery actions,
based on the knowledge of the structural architecture
of the applications and the dependencies between
service invocations.
The recovery action requires a Web service
with offer the same functionalities as the deficient
one. Actually, we deal with a predefined list of
similar and equivalent services. We need to improve
this component while using ontology for specifying
and searching the substitute services.
ACKNOWLEDGEMENTS
This work was supported by CAPES – Brazilian
Council of Research and LAAS-CNRS, France,
through collaboration research project CAPES-
COFECUB.
REFERENCES
Booth, D., Haas , H., McCabe , F., Newcomer , E.,
Champion, M., Ferris , C., Orchard, D. (Eds.) (2004,
February 11). Web Services Architecture. W3C
Working Group Note. Retrieved March 22, 2007,
from http://www.w3.org/TR/2004
/NOTE-ws-arch-20040211/.
Garcia, D. Z. G., & Toledo, M. B. F., A (2006). Web
Service Architecture Providing QoS Management. In
Proceedings of the 12th Brazilian symposium on
Multimedia and the Web, Natal, Rio Grande do Norte,
Brazil, 35–44.
Jacobson, V., & Karels, M. (1988, August). Congestion
Avoidance and Control (revised). In Proc. ACM
SIGCOMM'88, 314-329.
Karn, P., Partridge, C. (1991). Improving Round-Trip
Time Estimates. In Reliable Transport Protocols.
ACM Trans. Comput. Syst. 9(4), 363-373.
Ludwig, H. (2004). Web services QoS: external SLAs and
internal policies or: how do we deliver what we
promise? In Proc. of the Fourth International
Conference on Web Information Systems Engineering
Workshops (WISEW'03), 115-120. Springer.
Mahmoud, Q. H. (2005, April). Service-Oriented
Architecture (SOA) and Web Services: The Road to
Enterprise Application Integration (EAI). Retrieved
November 21, 2007, from http://java.sun.com
/developer/technicalArticles/WebServices/soa/.
Maximilien, E. M., Singh, M.P. (2004, September-
October). A framework and ontology for dynamic
Web services selection, In Internet Computing, IEEE,
8(5), 84- 93.
New to SOA and Web services (n.d.). Retrieved
November 3, 2007, from http://www.ibm.com
/developerworks/webservices/newto/.
Oracle Application Server 10g (n.d.). Retrieved November
12, 2007, from http://www.oracle.com
/technology/products/ias/.
Rud, D., Schmietendorf, A., Dumke, R. (2006).
Performance Modeling of WS-BPEL-Based Web
Service Compositions. IEEE Services Computing
Workshops (SCW'06) 140-147.
Web Services Business Process Execution Language v2.0
(2007, April 11). Retrieved November 20, 2007,
from http://docs.oasis-open.org/wsbpel/2.0/OS
/wsbpel-v2.0-OS.html.
Web Service Level Agreements (WSLA) Project (n.d.).
Retrieved November 12, 2007, from
http://www.research.ibm.com/wsla/.
ICEIS 2008 - International Conference on Enterprise Information Systems
206