Measurement Framework for Business Process Outsourcing to the Cloud

Mouna Rekik

1,3

, Khouloud Boukadi

1,3

and Hanene Ben-Abdallah

2,3

University of Sfax, Sfax, Tunisia

King Abdulaziz University, Jeddah, K.S.A.

Mir@cl Laboratory, Sfax, Tunisia

Keywords:

Business Process Performance, BPEL, Monitoring, GQM, IT.

Abstract:

Face to the increasingly stringent business competition, small and medium sized enterprises strive to excel

in the marketplace by adopting different strategies and solutions. Outsourcing their business processes to the

Cloud has been among the most widely adopted strategies. Among others, enterprises outsource their related

business process to improve their performance. However, this strategy is not without inconvenience especially

when the decision is taken without being aware about the business process functional and non functional re-

quirements. We focus in this paper on identifying the business process performance enhancement needs so

to be able to identify requirements when outsourcing business process to the Cloud. This papers major con-

tribution is the presentation of a measurement framework for SOA-based business process performance. The

proposed framework allows ﬁrstly to identify essential metrics to monitor starting from an abstract business

level. Then, identiﬁed metrics are monitored using our Business/Qos (BisQos) listener. The gathered data are

then stored in a database for analysis purpose. The output of the framework speciﬁes whether business process

instances reveal a degradation of their performance caused by business metrics or by Qos metrics, in addition

to the infrastructure properties supporting each web service execution.

1 INTRODUCTION

The number of Small and Medium sized Enterprises

(SMEs) emerging in the competitive markets is in-

creasing considerably. This situation urges enter-

prises to strive for sustaining a good place in such

economic context. Basically, many SMEs are rely-

ing on their business process (BP) to deliver their

goods and services to their clients and to gain added-

value. This reveals the need to deﬁne a well de-

signed environment for the execution of the BP giv-

ing them the ability to perform properly and to attain

their predeﬁned goals. Outsourcing business process

to the Cloud is considered as an emergent solution

thanks to the opportunities it offers. Indeed, accord-

ing to the National Institute of standards and technol-

ogy (NIST) (Mell and Grance, 2011), Cloud comput-

ing is a model for enabling convenient, on-demand

network access to a shared pool of conﬁgurable com-

puting resources that can be rapidly provisioned and

released with minimal management effort or service

provider interaction. Generally, the outsourcing busi-

ness process decision is taken thanks to opportunities

it may offer such as improve business process perfor-

mance (Yang et al., 2007). Assessing whether Cloud

computing may be an opportune environment for out-

sourcing should be preceded by observing and iden-

tifying functional and non functional requirements in

addition to the improvement needs. This paper focus

is to identify the business process performance en-

hancement needs and more speciﬁcally those that can

be fulﬁlled when adopting Cloud computing as an ex-

ecution environment. Usually, business experts deﬁne

their business models. These models are then reﬁned

into executable ones by IT experts and deployed into

process engines. The execution of the business pro-

cess can be made by web service composition follow-

ing the service oriented architecture (SOA) principles

(Josuttis, 2007). To monitor the business process per-

formance, SMEs should be able to depict and monitor

relevant IT/business metrics. In this context, enter-

prises experts should be able to proceed for IT/busi-

ness alignment serving to link between the business

process performance and related inﬂuencing IT/busi-

ness metrics. This in turn will help for enhancing the

business facet of business processes and even the IT

systems supporting them. Generally, assessing the

business process performance is based on Key Per-

Rekik M., Boukadi K. and Ben-Abdallah ..

Measurement Framework for Business Process Outsourcing to the Cloud.

DOI: 10.5220/0005537900490055

In Proceedings of the 12th International Conference on e-Business (ICE-B-2015), pages 49-55

ISBN: 978-989-758-113-7

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

formance Indicator (KPI) by using Business Activ-

ity Monitoring (BAM) tools providing near-real time

monitoring of business processes (zur Muehlen and

Shapiro, 2010). However, these tools are character-

ized by some shortcomings: they aren’t able to pro-

vide customized data corresponding to speciﬁc need

of business processes, reasons behind business pro-

cess performance degradation are not displayed and

ﬁnally they can not identify the infrastructure prop-

erties (CPU, Memory, etc) supporting business pro-

cess so to be able to depict requirements of business

process performance enhancement (Wetzstein et al.,

2011). Our main contribution in this paper is the

presentation of the measurement framework assisting

IT/business experts to:

• Depict relevant KPI, Qos and Business metrics

to monitor for sustaining a satisfactory level of

business process performance by using the GQM

paradigm.

• Monitor the depicted business/Qos metrics and

the business process performance by observing

their KPI value.

• Analyze gathered data and distinguish between

Qos and business metrics that inﬂuence the busi-

ness process performance degradation. This step

will be helpful in the context of our research, to

help enterprises to identify business process re-

quirements when outsourcing to the cloud. More-

over, the framework allows for the business/IT

alignment as its process starts from a high-level

namely the business process goals, to the low level

referring the properties of infrastructure support-

ing it.

The remainder of this paper is structured as follows.

Section 2 describes the related work. Section 3

presents in detail our proposed measurement frame-

work. In section 4 we illustrate an application case

to present the usefulness of our framework. Section 5

summarizes the work status and highlights its exten-

sions.

2 RELATED WORK

Monitoring the business process performance and

their execution as web service composition is know-

ing an exponential interest. Basically, several works

try to generate and analyze two categories of data re-

lated to business processes such as business and Qos

metrics. Qos are non functional requirements con-

sidered important for the satisfaction of costumers

and their objective function such as response time,

throughput, and availability (Zhang et al., 2009). In

the other side, business metrics are deﬁned by busi-

ness people based on business goals (e.g., comput-

ing process rentability, service reliability) (Wetzstein

et al., 2009) and are the basis for monitoring, con-

trolling, analyzing and improvement of the business

processes execution(Ruijie and Hong, 2010). In this

context, (Grati et al., 2012) and (Comes et al., 2009)

present techniques and solutions to monitor differ-

ent Qos dimensions of web services composing busi-

ness processes. Besides, assessing the continuous im-

provement of business process performance is tackled

by (Delgado et al., 2014) based on executionmeasure-

ment model containing a set of execution measures.

Observing and assessing whether business processes

are attaining their goals is done essentially by moni-

toring their KPI (Pan and Wei, 2012). KPIs are de-

ﬁned by experts over each business process to mea-

sure their performance. They are associated with a

target value as an objective to be attained in an anal-

ysis period. KPIs may be inﬂuenced by business

factors named process performance metrics (PPMs)

or/and Qos metrics. Different research works focus

on ﬁnding out how to model and monitor PPMs and

KPIs. In this context, (del Ro-Ortega et al., 2010)

presents an ontology to deﬁne process performance

metrics and explicitly depict relations between them

and elements of BPMN. Authors in (del Ro-Ortega

et al., 2012) propose templates and ontology based

linguistic pattern to facilitate the task of deﬁning PPM

over business processes. Generally, monitoring busi-

ness processes performance through KPIs is done by

business activity monitoring (BAM) tools. The BAM

allows for the supervision of business processes by

providing near real time information about their sta-

tus. It resorts essentially on dashboards where to

display needed information to accomplish enterprises

objectives. In this context, authors in (Friedenstab

et al., 2012) extends the BPMN with fundamental

concepts of BAM and the way to analyze and dis-

play its value. Although enterprises rely basically

on BAM tools to detect possible degradation of busi-

ness process performance, depicting possible inﬂuen-

tial factors on KPI isn’t provided. Overall, the re-

ported researches present solutions to model, mon-

itor and analyze either business or Qos metrics re-

lated to the execution of business processes. In the

best of our knowledge, there is no strategy neither a

complete tool proposed as a measurement framework

allowing: to depict appropriate Qos/business metrics

that may inﬂuence the business process performance,

monitor and analyze these metrics to differentiate be-

tween most inﬂuencing metrics (business/Qos) and ﬁ-

nally identify the infrastructure properties supporting

the business process execution. These three aspects

ICE-B2015-InternationalConferenceone-Business

Figure 1: The proposed measurement framework.

allow for deﬁning business process requirements for

performance enhancement when migration to another

executing environment is decided.

3 PROPOSED MEASUREMENT

FRAMEWORK

We proposed a framework based on standards for a

ﬁne and logic IT/business alignment contributing to

a pertinent collaboration between IT and business ex-

perts within enterprises. Next sections present in de-

tail each phase in this framework illustrated in ﬁgure

3.1 The Design Time Phase

The ﬁrst step consists on the depiction of essential

metrics to monitor the business process performance,

namely the KPI in addition to metrics that may in-

ﬂuence its value (business and Qos metrics) which is

accomplished using the GQM approach. The GQM

(Aversano et al., 2004) is an analytic goal-orientedap-

proach starting with a goal, reﬁned then into questions

from which users can identify relevant metrics.

The elaboration of the GQM approach in our

framework passes through three essential steps pre-

sented in table 1. This metric identiﬁcation step is

elaborated by business experts. The second step de-

ﬁnes goal using the depicted KPI. This in turn will

help IT/business experts to identify inﬂuential metrics

that may lead to business process performance degra-

dation namely business/Qos metrics. The ﬁnal step

helps IT experts to deﬁne implied IT properties that

may inﬂuence KPI values such as CPU, RAM, mem-

ory. We focus in the context of this paper on response

time Qod metric and IT properties that may inﬂuence

it.

3.2 The Runtime Phase

We develop a Business/Qos (Bis/Qos) listener for as-

sessing business process performance. It uses as in-

put metrics identiﬁed when applying GQM approach.

Our proposed listener accesses to the database where

metrics and formula used to calculate each one of

them are stored. Details about techniques used to

access the database and how to manipulate data are

non mentioned in this paper for space limitation. The

BisQos subscribes to the process engine allowing to

generate different data related to the business process

instances, such as the KPI value and its relative inﬂu-

ential metrics (PPM/Qos). Moreover, more detailed

data may be provided such as the IT properties sup-

porting the business process. The calculator compo-

nent uses the ﬁltered event data on which it applies

formula described on metrics database to calculate

PPM and Qos values. Furthermore, the BisQos lis-

tener calls the IT monitor responsible for gathering

IT properties and characteristic supporting BPEL in-

stances when execution begins, which is in our case

the Nagios monitor tool. The Nagios is an open

source tool commonly used by different projects and

researches (Barth, 2006). Finally, the output of the

execution data is stored in a speciﬁed database.

3.3 The Analysis Time Phase

The analysis step of the business/Qos data is based

on data mining methods offered by the WEKA data

mining software (Hall et al., 2009). Principally, the

goal of the data mining is to extract information

from a considerable number of data and transform it

into an understandable structure for further use. We

choose data mining as it enables studying and depict-

ing knowledge from historical business process ex-

ecution. This helps then to identify, based on this

new knowledge, reasons behind business process per-

formance degradation. Different methods are pro-

posed in the data mining. However, we used the de-

cision tree method as it can respond to our needs in

the analysis step (Quinlan, 1986). A decision tree is

a ﬂowchart-like structure in which each node corre-

MeasurementFrameworkforBusinessProcessOutsourcingtotheCloud

Table 1: Identifying metrics using GQM.

First step

Team Goal Question Metric

Business enterprise objectives Questions raised to de-

pict metrics to achieve

the objective

Metrics and KPI are de-

picted

Second step

Business enterprise objectives Questions raised to de-

pict metrics to achieve

the objective

Metrics and KPI are de-

picted

Business and

KPI to fulﬁll the objec-

tive

Questions raised to de-

pict IT and business in-

ﬂuential metrics

Inﬂuential metrics of

KPI

Third step

IT IT components affecting

Qos inﬂuential metrics

Questions raised to de-

pict implied IT compo-

nents

Implied IT components

sponds to an attribute and each branch corresponds to

a test on this attribute. Finally leaf node corresponds

to the class which is in our case yes/no. Classiﬁca-

tion rules are identiﬁed starting from the root to leaf.

Our focus is to identify whether Qos or business met-

rics inﬂuence most the KPI violation (business pro-

cess performance degradation). We proceed for the

analysis as follows: We ﬁrstly identify classiﬁcation

rules corresponding to the class yes (KPI violated).

Next, we calculate the number of occurrence of each

type of metric (business/Qos). Whenever Qos met-

rics inﬂuence most the KPI violation (number of Qos

metric occurrence is greater), we proceed for the anal-

ysis of IT properties of machines supporting the web

services executions using similarly, the decision tree

method. Solutions to enhance the business facet of

business process performance are not the focus of this

paper as we aim to depict the non functional require-

ments of business process to outsource it to the cloud

computing.

4 APPLICATION CASE

In this section we will present an example to illus-

trate in more details the utilization of our measure-

ment framework.

4.1 Identifying the Enterprise Vision

and Goals

We take as an example a Tunisian enterprise which

sells its products over the Internet. The enterprise

vision is to be the leader in Tunisia for online prod-

uct selling, depicted goals should insure the achieve-

ment of this vision. These long-term (strategic) goals

are SG1 ”Have a good relationship with costumers”,

and SG2 ”Be known in all Tunisia cities”. Objec-

tives quantiﬁes goals by presenting measures and tar-

get values to attain. In the studied business strategies

of the Tunisian enterprise, objectives are OB1 ”In the

end of December 2015, the number of adherents in

the enterprise web site increase to 2000” and the sec-

ond objective is OB2” In the end of 2016, number

of purchase order achieved in time and in full attains

60% against the number corresponding to 2015”. The

second cited objective quantiﬁes the strategic goal

SG1. Experts collaborate to decide about business

processes that follows each strategic goal. This step

allows to restrain the number of business processes to

be monitored in ulterior steps.

We will only focus in this paper, on the business

process ”Purchase Order” (PO) for space limitation.

The PO process follows basically the SG1. Next step

aims to depict essential metrics to monitor based on

the GQM approach.

4.2 Applying the GQM Paradigm

This phase encompasses three essential steps allow-

ing to identify ﬁrstly KPI to assess performance of

business processes, then potential inﬂuential metrics

and ﬁnally implied IT components.

4.2.1 Identifying KPI and Related Metrics

In this phase we aim to extricate for each business

process, the KPI to monitor its performance and its

behavior towards the achievement of its related strate-

gic goal. Table 2 presents the process used to this

purpose. It should be noted that we focus in this pa-

ICE-B2015-InternationalConferenceone-Business

per on KPI that can be calculated and derived from

runtime data of executable business processes. Busi-

ness experts are concerned in this phase. In this sec-

tion we will present an excerpt of metrics monitored

by the BisQos tool. In this step, several metrics are

depicted, only essential ones are designed to be the

KPI. In the presented example, the ”order conﬁrma-

tion lead time” is designed by business experts to be

the KPI reﬂecting the performance of the PO business

process.

4.2.2 Identifying Inﬂuential Metrics of the KPI

This step is intended to identify Qos and business

metrics that can deviate the KPI from its target value.

Table 3 presents in detail the process used to identify

inﬂuential metrics of the studied KPI. Next step fo-

cus on the IT properties and characteristics inﬂuenc-

ing the depicted Qos metrics.

4.2.3 Identifying Implied IT Properties

This phase helps to identify IT properties require-

ments for the enhancement of the business process

performance. The goal is identiﬁed from the Qos re-

quirements elicited previously. The excerpt of metrics

extracted when applying the GQM approach should

serve as input for our monitoring tool. Further details

are given in table 4. Next section presents the result

generated by our monitoring tool for the PO business

process.

4.3 Applying the Monitoring Tool for

the ”Purchase order” Business

Process

After identifying the KPI, inﬂuencing business and

Qos metrics, in addition to the related IT properties,

our tool monitor this metrics and display results for

experts for ulterior analysis steps. To do so, we con-

sider the simpliﬁed business process modeled with

BPMN ((OMG), 2011) presented in ﬁgure 2. In the

execution phase, BPMN is mapped to an executable

business process (BPEL process), and then an Apache

ODE is used as BPEL engine. The implemented

BisQos listener access to the repository where pre-

deﬁned metrics are stored. Essentially, the tool gath-

ers the monitoring data from speciﬁed repository that

encompasses different relevant data related to: the

target value of KPI and formula necessary to calcu-

late its value, and inﬂuential metrics identiﬁed using

the GQM. To properly monitor and calculate inﬂu-

ential metrics, we store in addition to their ID, the

allowed range of value and the attachments relating

Table 2: Extraction of KPI relative to PO process.

First step (Business team): Identifying

KPI to monitor and its related metrics

Goal: (OB2) In the end of 2016, number

of purchase order achieved in time and in

full attains 60%.

Question Metric Formula

Q1: What is

the number

of POs of the

current year?

M1: number

of PO

∑

Q2: What

is the per-

centage of

approved

and accorded

orders?

M2.1: num-

ber of Ac-

corded

Orders in

Time (AOT)

M2.2:number

of Total

Orders re-

quests(TO)

∑

Q3: What

is the Order

Conﬁrmation

lead Time

(OCT)?

M3.1: time

stamp (TR)

correspond-

ing to the

reception of

PO request

M3.2 time

stamp (TA)

correspond-

ing of the PO

conﬁrmation

OCT=TA-TR

each BPEL element with its needed events Our devel-

oped listener ﬁlters unnecessary metrics from repos-

itory and tackles only metrics identiﬁed by GQM as

input for the monitoring tool.

4.4 Analyzing Generated Data

Different decision tree methods are used to generate

pertinent results, namely, J48,CART, ADTree, and

REPTree. Following, we will give an example of

classiﬁcation rule generated when applying the J48

method on the business/Qos data.

Rule1 : i f ( check shipment

a v a i l a b i l i t y RT > 97)

and

( or der in stock = t ru e )

and

( check warehouse a v a i l a b i l i t y

RT > 101)

and ( check with s u pp l ier

RT >10)

MeasurementFrameworkforBusinessProcessOutsourcingtotheCloud

Table 3: Identifying inﬂuential metrics.

Second step (Business & IT team): Identifying inﬂuential factors of the KPI

Goal: The Duration of the PO process is lower than 3 days.

Question Metric Formula Type

Q1: What is the du-

ration of a web ser-

vice i corresponding

to REC process ?

M1.1: time stamp

TEi relative to the end

of web service i M1.2

time stamp relative to

the start of web ser-

vice TSi

duration=TEi-TSi Qos

Q2: What is the iden-

tity of costumers that

almost all their PO re-

quests lead to a re-

jected response?

M2: identity of cos-

tumers

PPM

Q3: Is the product al-

ways available?

M3: availability of

the product

PPM

Figure 2: The simpliﬁed purchase order process.

Table 4: Identifying implied IT components.

Third step (IT team): Identifying implied IT

components

Goal:Ameliorate duration of web services

composing the business process

Question Metric

Q1: What is CPU prop-

erties of server on which

web service i is running

corresponding to a BPEL

process j?

M1: CPU rate of us-

age related to server

supporting web ser-

vice i corresponding

to a BPEL process j

Q2: What is the System

disk properties of server

on which web service i is

running corresponding to

a BPEL process j?

M2: System disk

rate of usage related

to server supporting

web service i corre-

sponding to a BPEL

process j

then v i ola t ed = tru e ;

Table 5 presents the results corresponding to each

applied decision tree method, more speciﬁcally, the

decision tree method name, the percentage of cor-

rectly classiﬁed instances and the number of occur-

rence of the business/Qos metrics.

When calculating the occurrence of each attribute

(business and Qos), we observe that the number of

Qos metrics is greater then business metric. More

speciﬁcally, we can precise which web services Qos

Table 5: Experimental results: the application of decision

tree methods on business/Qos data.

Decision

tree methods

Percentage

of correctly

classiﬁed

instances

Number

of Qos

metrics

Number

of PPM

metrics

J48 85,48% 3 1

CART 87,36% 2 0

ADTree 86,26% 4 1

REPTree 86,26% 4 1

inﬂuence the most the KPI violation which are in our

case ”check warehouse availability RT” and ”check

shipment availability RT”. To enhance the Qos met-

rics, we go further in the analysis phase by depicting

the IT properties of the violated business instances

generated using the Nagios monitor. Monitored IT

properties in our case are CPU usage, memory usage,

system disk, virtual memory, page ﬁle, and total pro-

cesses. We choose to monitor these IT properties as

they are components that inﬂuence most the response

time Qos the focus of our paper. Herein an example

of classiﬁcation rule generated when we apply the J48

method on the IT properties data:

i f ( MemoryUsage1 > 82.5 and

MemoryUsage6> 90)

then v i ola t ed = tr u e ;

ICE-B2015-InternationalConferenceone-Business

MemoryUsage1 and MemoryUsage6 refers re-

spectively to the MemoryUsage of the machines sup-

porting respectively the web service ”check ware-

house availability” and ”check shipment availability”.

Regarding the huge number and the variety of com-

pute instances offered by Cloud provider, the gener-

ated classiﬁcation rule presents a means to limit the

range of choices. Referring to our example, enter-

prises should focus only on ﬁnding computing in-

stances providing better memory usage for the exe-

cution of the identiﬁed web services

5 CONCLUSION

Monitoring SOA based business processes presents

an interesting means for SMEs to analyze, interpret

and ameliorate the business process performance. De-

spite the evolving number of researches focusing on

that issue, realizing a method for identifying essen-

tial processes to monitor, identifying and monitoring

business and Qos metrics is not tackled in the best of

our knowledge. Thus, the framework presents a perti-

nent solution to assist business/IT experts for achiev-

ing enterprise goals. The tool we presented in this

paper is characterized by its top down aspect. In fact,

it links between business and IT levels helping ex-

perts to depict the relation between eventual KPI vi-

olation and the concerned IT properties. Moreover,

the framework identify the outsourcing to the Cloud

requirements. The presented framework is beneﬁcial

for both IT and business levels for displaying, analyz-

ing and enhancing different business processes data.

We are working on deﬁning an outsourcing algo-

rithm allowing to depict most relevant business pro-

cess parts to outsource and the evaluation the perfor-

mance of business processes executed in the Cloud.

REFERENCES

Barth, W. (2006). Nagios. System and Network Monitoring.

No Starch Press, u.s. ed edition.

Comes, D., Bleul, S., Weise, T., and Geihs, K. (2009). A

ﬂexible approach for business processes monitoring.

In Senivongse, T. and Oliveira, R., editors, Distributed

Applications and Interoperable Systems, volume 5523

of Lecture Notes in Computer Science, pages 116–

128. Springer Berlin Heidelberg.

del Ro-Ortega, A., Resinas, M., and Ruiz-Corts, A. (2010).

Deﬁning process performance indicators: An ontolog-

ical approach. In Meersman, R., Dillon, T., and Her-

rero, P., editors, On the Move to Meaningful Internet

Systems: OTM 2010, volume 6426 of Lecture Notes

in Computer Science, pages 555–572. Springer Berlin

Heidelberg.

del Ro-Ortega, A., Resinas Arias de Reyna, M., Durn Toro,

A., and Ruiz-Corts, A. (2012). Deﬁning process per-

formance indicators by using templates and patterns.

In Barros, A., Gal, A., and Kindler, E., editors, Busi-

ness Process Management, volume 7481 of Lecture

Notes in Computer Science, pages 223–228. Springer

Berlin Heidelberg.

Delgado, A., Weber, B., Ruiz, F., de Guzmn, I. G.-R., and

Piattini, M. (2014). An integrated approach based on

execution measures for the continuous improvement

of business processes realized by services. Informa-

tion and Software Technology, 56(2):134 – 162.

Friedenstab, J., Janiesch, C., Matzner, M., and Muller, O.

(2012). Extending bpmn for business activity moni-

toring. In 2012 45th Hawaii International Conference

on System Science (HICSS), pages 4158–4167.

Grati, R., Boukadi, K., and Ben-Abdallah, H. (2012). A qos

monitoring framework for composite web services in

the cloud. In In The Sixth International Conference on

Advanced Engineering Computing and Applications

in Sciences (Advcomp12).

Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann,

P., and Witten, I. H. (2009). The weka data min-

ing software: An update. SIGKDD Explor. Newsl.,

11(1):10–18.

Josuttis, N. (2007). Soa in Practice: The Art of Distributed

System Design. O’Reilly Media, Inc.

Mell, P. M. and Grance, T. (2011). Sp 800-145. the nist deﬁ-

nition of cloud computing. Technical report, Gaithers-

burg, MD, United States.

(OMG), O. M. G. (2011). Business process model and no-

tation (bpmn) version 2.0. Technical report.

Pan, W. and Wei, H. (2012). Research on key performance

indicator (kpi) of business process. In 2012 Second

International Conference on Business Computing and

Global Informatization (BCGIN), pages 151–154.

Quinlan, J. R. (1986). Induction of decision trees. Mach.

Learn., 1(1):81–106.

Ruijie, L. and Hong, H. (2010). Business process metrics

modeling based on model-driven. In 2010 IEEE In-

ternational Conference on Software Engineering and

Service Sciences (ICSESS), pages 327–330.

Wetzstein, B., Leitner, P., Rosenberg, F., Dustdar, S., and

Leymann, F. (2011). Identifying inﬂuential factors of

business process performance using dependency anal-

ysis. Enterprise IS, 5(1):79–98.

Wetzstein, B., Strauch, S., and Leymann, F. (2009). Mea-

suring performance metrics of ws-bpel service com-

positions. International conference on Networking

and Services (ICNS’06), 0:49–56.

Yang, D.-H., Kim, S., Nam, C., and Min, J.-W. (2007).

Developing a decision model for business process

outsourcing. Computers and Operations Research,

34(12):3769 – 3778. Operations Research and Out-

sourcing.

Zhang, J., Nie, W., Panahi, M., Chang, Y., and Lin, K.-J.

(2009). Business process composition with qos op-

timization. In Commerce and Enterprise Computing,

2009. CEC ’09. IEEE Conference on, pages 499–502.

zur Muehlen, M. and Shapiro, R. (2010). Business process

analytics. In Handbook on Business P rocess Manage-

ment 2, pages 137–157. Springer Berlin Heidelberg.

MeasurementFrameworkforBusinessProcessOutsourcingtotheCloud