International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 6June 2013

ISSN: 2231-2803 http://www.ijcttjournal.org Page 1845

Survey on SOA Fault Tolerance Techniques

Anushka
M.Tech (IT)Student, Dept. of ISE, R.V.College of Engineering, Bangalore,India


Abstract: The collection of discrete software modules
forming software architecture design pattern to deliver
services. These services collectively provide the
complete functionality of a large software application.
SOA allows performing the operations of a large
number of computers that are connected over a
network with ease. SOA exhibits features loose
coupling, black box, dynamism, etc. Sometimes, SOA
faces problem in delivering messages between different
services and faults in the SOA components, etc. These
faults are diagnosed with various methods such as
actuators, different service routing methods, MAPE,
etc.

Key Words: SOA, Fault Management, MAPE, Double
Redundancy, Service Routing.

I. INTRODUCTION

Service-Oriented Architecture (SOA) is an effective
model for developing applications. SOA discloses
features which are not commonly found in
conventional computing paradigms [7]. The features
of services, loose coupling, dynamism, blackbox,
and heterogeneity, make service management more
challenging than conventional system management.
Consequently, SOA service management face
problems of increased management effort, decreased
management effectiveness, and irresolvable service
faults. There are many resolutions to these ill
operations and this paper will be discussing some of
them [1]. This paper will discuss the faults, problems
that has been identified by authors occurring in SOA
and their different technique used to resolve them.
Authors of An Engineering Process for Autonomous
Fault Management in Service Oriented Systems
promises an approach to manage faults in SOA is to
apply the principle of autonomous computing (AC)
[12], on service management, where AC is known as
an efficient approach to manage systems in
autonomous manner with minimum human
intervention. To apply AC to managing faults in
SOA, 1) the whole life-cycle model and 2)
instructions for carrying out each phase are both
required. 3) Also, a key phase which requires a
mathematically sound foundation in this life-cycle is
to diagnose the given symptoms/faults and determine
their underlying causes. Moreover, it is required that
the process supports to make the autonomous
management agent/system more intelligent, by 4)
learning new facts and rules from the experiences of
applying each cycle. This is an essential requirement
to be autonomous since human intervention is not an
integral part of AC and can be called as engineering
process. Hence, in engineering process it evolves
following steps. First, define a whole life-cycle
process for managing faults in SOA as well as the
essential artifacts of each phase. Then, present a set
of instructions for five phases; Preparation Phase,
Monitoring Phase, Diagnosis Phase, Healing Phase,
and Learning Phase.
Author of Double Redundant Fault-Tolerance
Service Routing Model in ESB [2],[6] believes that
Enterprise Service Bus (ESB) is a backbone for
service integration. Usually, it is regarded ESB as the
kernel of SOA, and it can manage mass services for
different purposes. An ESB [19] acts as a standards-
based integration platform which syndicates
messaging, web services and data transformation in a
highly distributed environment. The Author
concentrates more on how to provide a reliable
service routing so that despite of any faults in routing
the messages has to be reached to the destination.
Hence, they introduced a double service routing
method wherein there will be replica of each service
routing so that if one service routing fails to deliver
the message then there will be another to deliver it.
Authors of A Practical Framework of Realizing
Actuators for Autonomous Fault Management in
SOA [3] aim to resolve the faults that occur in the
components of SOA system. They came up with an
approach applying the principles of autonomic
computing (AC) [10], [12], which will manage the
services in consistent and dynamic manner. The first
two phases of MAPE [3],[13] are to monitor target
systems and diagnose faults to determine underlying
cause. The other two phases of it are to plan
healing/actuation methods and to execute them. They
introduced Actuators the software modules which
will act as a healing mechanism to the faulty
components based on their symptoms. They
presented a practical framework to design actuators
which can be invoked autonomously in SOA
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 6June 2013

ISSN: 2231-2803 http://www.ijcttjournal.org Page 1846

environment, by considering the relationship among
fault, source, and actuator.



II. LITERATURE REVIEW

In An Engineering Process for Autonomous Fault
Management In Service Oriented Systems Service-
Oriented Computing exposes features such; loose
coupling, black box, dynamism and heterogeneity
[1][7]. Thus, service-oriented systems often face
problems of augmented cost-effort, reduced
effectiveness, and irresolvable service faults. Du Wan
Cheun and Soo Dong Kim exhibit a method to
manage faults in autonomous manner [11]. It
includes a complete life-cycle process to manage
faults in SOA and the important artifacts of each
stage, which carries out some instructions. Each stage
consists of set of instructions - Preparation Phase,
Monitoring Phase, Diagnosis Phase, Healing Phase,
and Learning Phase [8]. They suggest methods for
diagnosing faults which is devised with two inference
mechanisms, i.e. Model-Based Reasoning (MBR)
[15],[17], - Probability-based reasoning (PBR) and
reasoning knowledgebase updates by learning
mechanism in detailed. The process is carried out on
current service-oriented architecture standards.

In Double Redundant Fault-Tolerance Service
Routing Model in ESB the development of the
Service Oriented Architecture (SOA), Enterprise
Service Bus (ESB)[2] [6] is becoming more
important in the management of mass services. The
main feature is to deliver the messages between
different services by using service of routing.
Presently static configuration service routing is being
used widely. If any one message delivering service
fails, the system could not identify this fault. Bin Wu,
Shijun Liu and Wei Cui is resolve this problem, they
present a double redundant fault tolerant service
routing model. They propose an algorithm to achieve
the double redundant fault tolerant mechanism. If the
original service fails, an alternate replica service will
be activated that has the same function and it will
return the response message automatically [8]. The
service requester receives the response message
evidently without any knowledge that where it comes
from. Thus, the response message service will be
delivered even there be any problem ensuring the
reliability. The service that has failed to perform its
job will be recorded by service management. Also,
the performance of double redundant fault tolerant
service routing model are evaluated. By introducing
double redundant fault tolerance by implementing
dynamic routing, and ensure the reliability of
messaging in SOA.


In A Practical Framework of Realizing Actuators for
Autonomous Fault Management in SOA the faults
in SOA can be diagnosed by applying the principles
of autonomic computing (AC)[9], [10],[12], of which
process is specified in Monitor, Analyze, Plan,
Execute (MAPE).MAPE comprises of 2 stage[3],
firstly to monitor target systems and diagnose faults
to determine underlying cause. Secondly it plans the
healing or actuation methods and to execute actuators
applying on faults. Hyun J ung La and Soo Dong
Kim, present a practical framework to design
actuators which can be invoked autonomously. By
seeing the relationships between fault, cause, and
actuator, they derive the abstract and concrete
actuators [9]. For these actuators, they derive
algorithms which can be implemented in practice on
faults. Thus, the faults can be analyzed by
implementing respective actuators.


In SOAR: An Extended Model-Based Reasoning for
Diagnosing Faults in Service-Oriented Architecture.
Service-Oriented Architecture (SOA) is a cost
effective approach to build enterprise applications
[4]. SOA reveals non-conventional challenges such
as fault diagnosis at runtime is challenging due to the
SOA features. Model-Based Reasoning (MBR) [17]
is a formal approach to analyze faults, which is built
on predicate calculus and term resolution. Soo Dong
Kim and Soo Ho Chang present SOAR [9]
(Service-Oriented Abductive Reasoning) which
covers the basic MBR [16], [18] to diagnose faults in
copious SOA components. SOAR delivers an
enhanced inference capability by comparing the
monitored state based and QoS-based reasoning in
addition to the basic setting or observation-based
reasoning [13]. They proposed a method to formally
represent system description i.e. system component
description, components normal behaviour [9],[10],
[12], fault model and observations, and reasoning
methods to analyze faults and to identify their causes.

In The Sender Released Pattern: An SOA design
pattern for inter-service message Exchange it
describes modelling a complex interactions and
integrations between distributed and heterogeneous
applications Service-oriented architectures are widely
being used [5]. But, these architectures possess some
failures (e.g. reliability, availability, and performance
problems). Design patterns, can be used as tested
solutions to specific problems. The Reliable
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 6June 2013

ISSN: 2231-2803 http://www.ijcttjournal.org Page 1847

Messaging" pattern is proposed by the SOA design
pattern community, its reliable system is subjected to
a great number of failures that can crash the
dependable functionality [14]. Imen TOUNSI,
Mohamed Hadj Kacem, proposed a service based
approach for dynamic management of service-
oriented architectures using SOA design patterns
[20]. They introduced a new pattern that extends the
Reliable Messaging" pattern in order to resolve its
fault and also introduced message-oriented SOA
design patterns model with the SoaML [21], standard
language through Sender Released" pattern.
III. CRITICAL EVALUATION

TABLE I
COMPARISON OF VARIOUS SOA FAULT TECHNIQUES

Literature
Reference
Problem
discussed
Technique Used

1

Service Faults
Two inference
mechanisms; MBR
(model based reasoning)
and PBR (Probability
based reasoning).

2

Service routing
A double redundant fault
tolerant service routing
model

3

Faults in the SOA
components
Deriving theabstract
and
concrete actuators.

4

Diagnose fault at
runtime in SOA
components
SOAR
(Service-Oriented
Abduvtive Reasoning)
which covers thebasic
MBR

5
Faults in the
interactions and
integrations
between
distributed and
heterogeneous
applications
Service-oriented
architectures
Design patterns :
Sender Released
pattern, the modeling of
the Sender
Released" pattern with
the help of SoaML
language.

IV CONCLUSION
This paper discusses about the various SOA faults,
techniques to identify, analyse and resolve these
faults.
REFERENCES
[1] Du Wan Cheun, Soo Dong Kim, An Engineering Process
for Autonomous Fault Management in Service-Oriented
Systems, 9th IEEE/ACIS International Conference on
Computer and Information Science.
[2] Bin Wu, Shijun Liu, Wei Cui Double Redundant Fault-
Tolerance Service Routing Model in ESB, 2009.
[3] Hyun J ung La and Soo Dong Kim, A Practical Framework
of Realizing Actuators for Autonomous Fault Management
in SOA,2009.
[4] Soo Dong Kim and Soo Ho Chang, SOAR: An Extended
Model-Based Reasoning for Diagnosing Faults in Service-
Oriented Architecture, 2009 Congress on Services I.
[5] Imen TOUNSI, Mohamed HADJ KACEM, Ahmed HADJ
KACEM, Khalil DRIRA The Sender Released Pattern: An
SOA design pattern for inter-service message Exchange,
2012.
[6] MartinKeen, Amit Acharya, Susan Bishop, et.a1, Patterns:
Implementing an SOA Using an Enterprise Service Bus
,Redbooks, IBM Press, J uly 2004.
[7] Erl, T., SOA Principles of Service Design, Prentice Hall,
J uly, 2007.
[8] Gregor Hohpe, Enterprise Integration Patterns: Designing,
Building, and Deploying Messaging Solutions, Addison-
Wesley Professional, 2003.
[9] P. Brittenham, R. Cutlip, C. Draper, B. Miller, S.
Choudhary, and M. Perazolo, IT service management
architecture and autonomic computing, IBM Systems
Journal, Vol. 46, No, 3, 2007, pp. 565-581.
[10] R. Cutlip and C. Zabeu, Autonomic Computing:
Strengthening Manageability for SOA Implementations,
Autonomic Computing White Paper, Dec., 2006.
[11] B. Pernici, and A.M. Rosati., Automatic Learning of Repair
Strategies for Web Services, Proc. 5th IEEE EuropeanConf.
Web Services (ECOWS 2007), IEEE Computer Society,
Nov. 2007. pp. 119-128.
[12] Kephart, O. and Chess, M., The Vision of Autonomic
Computing, IEEE Computer, IEEE Computer Society Press,
Vol. 36, No. 1, J anuary 2003, pp. 41-50.
[13] Bocconi, S, et al, Model-based Diagnosability Analysis for
Web Services, LNAI 4733, pp. 24-35,
2007.
[14] Gomaa, H., Hashimoto, K., Kim, M., Malek, S.,Menasc,
D.A.: Software adaptation patterns for service-oriented
architectures. In: SAC 10: Proceedings of the 2010
ACMSymposiumon Applied Computing, New York, NY,
USA, ACM (2010) 462469.
[15] Ardissono, L., et al., Cooperative Model-Based Diagnosis
of Web Services, In Proceedings of the 16th International
Workshop .On Principles of Diagnosis (DX 2005), 2005,
pp. 125-132.
[16] Peischl, B. and Wotawa, F., Model-Based Diagnosis or
Reasoning from First Principles, Intelligent System, IEEE
Computer Society, Volume 18, May 2003, pp. 32-37.
[17] Ardissono,L., et al., Cooperative Model-Based Diagnosis
of Web Services, In Proceedings of the 16
th
International
Workshop on Principles of Diagnosis (DX
2005),2005,pp.125-132.
[18] Peischl, B. and Wotawa, F., Model-Based Diagnosis or
Reasoning from First Principles, Intelligent System, IEEE
Computer Society, Volume 18, May 2003, pp. 32-37.
[19] Xi aoyi ng Bai , J i hui Xi e, Bi n Chen, Si nan Xi ao ,
Dynami c routi ng i n Enterpri se Servi ce Bus, I EEE
i nternat i onal Conf erence on e- busi ness
Engi neeri ng, 2007.
[20] Marcano-Kamenoff, R., Lvy, N., Losavio, F.: spcification
et spcialisation de patterns enUML et B. In: LMO. (2000)
245260.
[21] Object Management Group, O.M.G.: Service oriented
architecture Modeling Language (SoaML)- Specification
UML Profile and metamodel for services (UPMS). (2008)