BPEL Aided Framework for Constructing Monitoring Rules in

Service Oriented Architectures

Plamen Petkov and Markus Helfert

School of Computing, Dublin City University, Glasnevin, Dublin 9, Ireland

Keywords: Service Oriented Architectures, BPMN, Monitoring Rules, Data Quality, Data Monitor, Business Rules,

Rule Repository.

Abstract: Since the ascension of Service orientation architectures, data quality topic is becoming more and more

extensive. Based on a data quality management approach, we propose a framework for analysing data

produced by the composite service execution. In this paper we focus on the preparation stage, the stage

where the quality is defined. Apart from other methods available which involve extracting or importing

business rules into business processes, we propose technique for constructing monitoring rules which serves

as a data quality arbiter. The process of composing is aided by the BPEL language. More particularly,

service variables are obtained from the BPEL and WSDL repository and sets of operations are suggested to

the business body. Eventually, monitoring rules are stored into a practical oriented repository.

1 INTRODUCTION

Service Oriented Architecture (SOA) has gained

much acceptance in the past few years as the model

for application development that uses groups of

services that are orchestrated around business

processes. SOA strives to provide existing functions

of an information system (IS) as "services" that can

be accessed in a loosely coupled way (Papazoglou,

2007), independently from the technical platform.

Generally, in SOA, workflow or orchestration

processes are fundamental. However, SOA

comprises more than just the manner in which

applications are deployed and cooperate with each

other – it also involves how they interact with data.

Nevertheless, in more complex architectures,

loosely coupled data can cause serious problems to

the whole service composition and thus make it

difficult to handle. In (Petkov, 2012) it was

presented some data issues that mostly occur within

SOA. There is no efficient way to managing such

architectures without having the awareness of the

data, processes and events running within the

enterprise environment. This is why, vital assistant

for proficient and effective deployment and

operation of a SOA-based net-centric system is a

comprehensive monitoring capability. Still, present

monitoring solutions fall short with respect to such

systems because they do not hold the capabilities to

implicitly aggregate metrics, effectively detect

inconsistent or inaccurate data, and so to provide

comprehensive shared situational perception.

In this paper we propose an approach on how

Business Process Execution Language (BPEL) in

SOA can aid the business architects with building

business rules that will eventually facilitate detection

of inaccurate and inconsistent data. The approach

follows the Data Quality Management (DQM)

(Oracle, 2008) model. It focuses on constructing the

business rules which is part of the preparation

stage/defining the quality in DQM.

2 BUSINESS RULES IN IS

The related work in this area is very diverse and

scopes huge area of business rules knowledge.

Business rules deliver knowledge and this

knowledge can be used to achieve different goals

(Hay et al., 2003). For instance, a rule can be used to

provide additional information to a business process

and in this way to alter a data object or change

values in a database. In the other hand, same rule

can be used to verify the correctness of an

informational object. As a result, two groups of rules

can be distinguished: (1) group used to manipulate

141

Petkov P. and Helfert M..

BPEL Aided Framework for Constructing Monitoring Rules in Service Oriented Architectures.

DOI: 10.5220/0004368201410144

In Proceedings of the 9th International Conference on Web Information Systems and Technologies (WEBIST-2013), pages 141-144

ISBN: 978-989-8565-54-9

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

information, usually by bringing extra value to it;

and (2) group of rules that access and monitor

existing information without changing it.

When taking into account the relationship

between business rules and business processes,

literature identifies two research streams – (1) one

which targets deriving and discovering business

rules from business processes (Vasilecas &

Normantas, 2011), (Ramsey & Alpigin, 2002),

(Earls et al., 2002), and (2) another that aims

integrating the BR into the BP. (Ly et al., 2008)

In contrast with the approaches above, our goal

is not concerned to discover or derive business rules,

neither to integrate business rules into business

processes. It aims to assist the business bodies and

architects with building rules, leaving the flexibility

business ultimately to decide quality of data and in

this way achieving the main goal which is to detect

inaccurate data within SOA.

3 CONSTRUCTING QUALITY

RULES

As previously mentioned, business rule can be used

to verify a value of data and in this way to be used as

part of monitoring process. As it was outlined in

(Wand Y, 1996) data must be persistent in order to

be evaluated. An example of monitoring rule in SOA

context is presented by statement (1):

IF NOT (

SrvA.Oper.GetCustStatus = ‘Gold’AND

SrvB.Oper.GetCartQuty > 10 AND

SrvC.Oper.GetShippingCost = 0 )

THEN

GENERATE

RULE

EXEPTION ERROR

(1)

At this stage, the left side (before THEN) of the

aforementioned statement can be brought down to

evaluating sets of simple predicates. The final result

would be evaluated by using logical conjunction

(Hazewinkel, 2001). In logical conjunction, the

result will be true if all of its operands are true,

otherwise the value is false.

In the general case the complete rule expression

for monitoring purposes could be exemplified in the

following way:

IF NOT (

Predicate_1 BITWISE_OPERATION

Predicate_2 BITWISE_OPERATION

. . .

Predicate_N )

THEN

GENERATE

RULE

EXEPTION ERROR

(3)

where

Predicate_1 . . . Predicate_N are

predicates linked with Bitwise Operations. Each

predicate is composed of three objects -

[returned_value] comparison_operation

[expected_value] where [returned_value] is the

value delivered by the target service operation

(usually fetch function e.g.

getOperation() );

comparison_operation is any of the following

logical operations: ‘<’(less), ‘>’(greater), ‘=’equal,

‘!=’(not equal), ‘<=’(less than or equal) ‘>=’(greater

than or equal); [expected_value] is expected value,

the one to be conditioned, usually defined by the

user;

BITWISE_OPERATION is any of the Boolean

algebra operations (Koppelberg, 1989): ‘∧’ (AND,

conjunction), ‘∨’ (OR, disjunction), ‘¬’ (NOT,

negation) . The whole predicate set must be inverted

using negating Boolean logical operation (NOT), if

the logic is positive. If negative logic is used, negate

operation must be omitted after predicates are been

evaluated.

4 FRAMEWORK FOR

COMPOSING BUSINESS

RULES AID BY BPEL

LANGUAGE

Framework for composing quality rules is part of

overall Data Quality Management (DQM) process

which identifies four main stages in data quality

cycle - Quality Assessment, Quality Design, Quality

Transformation and Quality monitoring. Since DQM

provides us only with general guidance how to

manage quality of information, the framework we

propose aims to show in detail how to define quality

of information in SOA. The framework is focused

particularly on the ‘Quality Design’ of DQM. We

name this stage ‘Preparation stage’. This is the stage

were the quality is defined. It is comprised of four

steps: (i) selecting BP; (ii) extracting the services

and variables involved into the chosen business

process or sub-process; (iii) constructing business

rules based on the services and variables involved in

the chosen BP; (iv) mapping and saving created

rules into repository; The preparation stage is

depicted on Figure 1.

The approach we propose aim to construct

quality rules which serve as data quality arbiter, and

that construction is aided by BPEL language. In

order to achieve the goal, we incorporate into our

framework some of the vital building blocks in SoA

such as Business Process Engine (BPE), Enterprise

WEBIST2013-9thInternationalConferenceonWebInformationSystemsandTechnologies

142

Service Bus (ESB), WSDL Repository and Business

Processes’ Repository (part of BPE).

Selecting a Business Process.

In this step (Figure 1

– (i)) a business process or sub-process is selected

by an authorised person. Based on the selection

made, later in the preparation process, objects and

function names are extracted in order to aid

constructing of the business rules. Hence, it is

necessity business person to be familiar with the

selected process its activities. It is important that the

scope of the process is well defined as this will

affect construction of the rules later in preparation

stage.

Extracting Services’ Names, Operation and

Variables. In extraction phase, the business process

chosen from previous step undergoes series of

analysis performed by business process reader.

Ultimately specific meta-data (data about the

process) is obtained and temporary stored by the rule

builder (Figure 1(ii)). This meta-data includes

‘service name’, ‘operation’, ‘input’ and ‘output’

variables. In Service-oriented Architectures business

process usually is described with Business Process

Execution Language (BPEL). It is a powerful tool

build on XML technology stack that provides a rich

vocabulary for description of business processes.

The vocabulary includes tags for describing

primitive activities – <invoke>, <receive> as well as

structural activities <sequence>, <flow>.

For the purpose of the extraction, BPEL file

containing the prescription of the selected process is

opened for examination. The process of analysing

consists of scanning the file and searching for BPEL

attributes that correspond to the metadata that need

to be collected. The data needed for next step are

‘service name’, ‘operation’, ‘input’ and ‘output’.

The representative BPEL attributes are as follow:

‘partnerLink’, ‘operation’, ‘inputVariable’ and

‘outputvariable’. The implementation of the BP

Reader is intentionally omitted due to the scope of

this paper.

Building up the Rules. This stage focuses on actual

composing of the rules. The actual construction of

the rules is done by business body – architect or

administrator. The metadata stored temporary form

BP reader in extraction stage then is referred to

WSDL repository and files containing the relevant

variables are opened (iii). Then a list of services

along with their data/read functions is presented to

the user. Next the business administrator/architect

composes the rules using the methodology we

described in detail in Section 3. In this way business

body is allowed to compose simple rules –

composed only of one statement or complex rules –

combination of statements linked with logical

operators in specific order.

Mapping Rules with Processes and Storing them

into Repository. The composed quality rules are

stored (iv) in an external repository. In order to

maintain their execution in loosely coupled way later

in the execution stage, we introduce a holistic way to

store the rules into XML files. Hence we propose a

suitable template. The template is presented below

along with an example:



Figure 1: BPEL aided framework.

BPELAidedFrameworkforConstructingMonitoringRulesinServiceOrientedArchitectures

143

<ruleSet id = “23”

name = “free_delivery”

processId = “12” >

<description>Check if customer status

is gold</description>

<serviceName>custSRV</serviceName>

<portType>customerDb</portType>

<serviceOperation>getCustomerStatus

</serviceOperation>

<expectedVariable>'GOLD'

</expectedVariable>

AND

</ruleLink>

</ruleLinks>

...

</ruleSet>

</ruleList>

It is necessary to notice few important features

that this template offers: First by storing

identification number of the process into the rule

record will avoid data ‘dead locking’ later in the

execution step. Second, introducing ‘

<ruleLinks>’

tag will allow building complex set of rules by

storing information about their relations e.g. logical

AND, OR , XOR operations.

5 CONCLUSIONS

Inspired from research in the area of data quality and

service oriented architectures, in this paper we have

presented a BPEL aided framework for constructing

monitoring rules in web service composition and

execution. More specifically, we have divided the

framework into four phases, namely ‘selecting

business process’ and ‘extracting services’ names

operation and variables’, ‘constructing the rules’ and

‘mapping and saving the rules into external

repository’.

Apart from other well-known methods for

directly extracting or integrating business rules form

and into business process, our approach differs in the

way by providing the business bodies with a holistic

solution that will aid them with building rules which

will serve as data quality arbiter. The approach uses

BPEL language to extract services names and

operation and then with backend by WSDL library

offers a list of operations that aid the business

architect witch composition of the rules.

ACKNOWLEDGEMENTS

This work was supported by the Irish Research

Council (IRC) under Postgraduate Scholarship

scheme.

REFERENCES

Earls, A., Embury, S. & Turner, N., 2002. A method for

the manual extraction of business. BT technology

journal, 20(4), pp.127-45.

Hay, J., Kolber, A. & Anderson, K., 2003. Defining

Business Rules - what are they really. http://

www.businessrulesgroup.org/first_paper/BRG-

whatisBR_3ed.pdf [Accessed November 2012].

Hazewinkel, M.., 2001. Conjunction. In Encyclopedia of

Mathematics. Springer.

Koppelberg, S., 1989. General Theory of Boolean

Algebras. In J.D. Monk & J.D. Bonnet, eds. Handbook

of Boolean Algebras. Amsterdam.

Ly, L., Rinderle, S. & Dadam, P., 2008. Integration and

verification of semantic. Journal of Data &

Knowledge , 64, pp.2-23.

Oracle, I., 2008. Understanding Data Quality Manage-

ment. In Oracle® Warehouse Builder User's Guide.

Oracle Inc.

Papazoglou, M. P., Traverso, P., Dustdar, S. & Leymann,

F., November 2007. Service-Oriented Comput-

ing:State of the Art and Research Challenges. IEEE

Computer.

Petkov, P. H. M., 2012. Data oriented challenges of

service architectures a data quality perspective. In

CompSysTech. New York, NY, 2012.

Ramsey, F. & Alpigin, J.i., 2002. A simple mathematically

based framework for rule. In 26th Annual

International Conference In Computer Software and

Applications., 2002. IEEE.

Vasilecas, O. & Normantas, K., 2011. Deriving business

rules from the models of existing information systems.

In 12th International Conference on Computer

Systems and Technologies. New York, NY, 2011.

ACM.

Wand Y, W. R., 1996. Anchoring Data Quality

Dimensions in Ontological Foundations. Communica-

tions of the ACM, pp.86-95.

WEBIST2013-9thInternationalConferenceonWebInformationSystemsandTechnologies

144