International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 1
ISSN 2250-3153
www.ijsrp.org
Non Functional Requirement Classification for Service-
Oriented Data Warehousing
*Anjana Gosain, **Sangeeta Sabharwal, ***Rolly Gupta
*University School of Information Technology, GGS Indraprastha University, Delhi, India
**Netaji Subhas Institute of Technology, Delhi University, Delhi, India
***Netaji Subhas Institute of Technology, Delhi University, Delhi, India
*anjana_gosain@hotmail.com, **ssab@nsit.ac.in,
***
Abstract - Recently, Data Warehouse has proven to be a
powerful technology for the integration of heterogeneous data
into a multidimensional repository for decision-support
analysis. The complex ETL (extraction, transformation and
loading) process and the aggregation-intensive queries are
affected by a sequence of domain specific NFRs (Non
Functional Requirement). This advocates the use of service
oriented NFR approach for building a data warehouse
specification.
However, the specification and classification of service
oriented systems and service oriented data warehousing (DW)
systems have not been addressed to the appropriate level as
attempted for non service-oriented systems. In this paper, we
propose a new framework for the classification of NFR (Non-
functional requirements) with respect to engineering service
oriented and service oriented data warehousing systems. In
addition, this proposed classification is supposed to be of
significance in terms of NFRs classification for service
engineering and service-oriented DW engineering.
Index Terms— Non Function Requirement specification, SOC,
QoS, NFR, Data Warehousing
I. INTRODUCTION
n recent years Data Warehouse (DW) has proven to be a
powerful technology for integrating distributed operational
data into a comprehensive repository for predicting future
decisions. Therefore, the design of these systems is rather
different from the design of the conventional operational
systems that is involved with the supply of data to the
warehouse systems. The former involves information
requirements of decision makers, structure and allocated
requirements of the latter systems also. Software engineers
need to deal with the complex process of ETL (extracting,
transforming, and loading) data while managing the
deployment of solution; that can precisely and timely integrate
a number of heterogeneous source systems. On the basis of
deployment, they present analytical results in an accurate and
reliable form; thus offering flexibility at the front-end with the
support of a complete, non-redundant dimensional model.
Thus, both operational and strategically visionary approach
have to be wrapped up in a multidimensional package in order
to meet corporative analytical requirements pervading pure
decision-support functionality as well as quality constraints
like integrity, accessibility, performance, etc [23]. This
advocates the use of DW Requirements Engineering
techniques for building a data warehouse specification
precisely.
To pursue the goal, we propose a methodological approach for
the requirements analysis of data warehouse systems. The
approach provides an service-oriented method for the
guidance of requirements engineers during the data warehouse
specification process. The non-functional requirement acts as
a constraint on software product, or on software development
process and externally related constraints. However, much
knowledge is required regarding the requirements
specifications for service oriented systems i.e. from functional
view to non-functional view. The proposed approach is
general with regard to non-functional requirements and paves
the paths in order to understand the positive and negative
aspects of a certain requirement in regard to the process of
data warehouse.
The NFR Framework [6] helps to produce a solution that is
able to embrace the quality characteristics of a particular
domain. So, we aim to devise a new classification of non-
functional requirements for service oriented systems (service
engineering) and service-oriented DW systems (service-
oriented engineering). We have designed a hierarchical tree of
non-functional requirements in respect to the data warehouse
engineering. Such a classification helps in identifying and
guiding the NFR specifications for service-oriented systems.
In the paper, a review of related work concerning service-
oriented engineering and NFRs classification is presented in
section 2. In Section 3, we introduce the new framework of
NFR classification. Lastly, we conclude our discussion
regarding the proposed classification and future work
directions.
II. RELATED WORK
A. Service Oriented Computing
Service-Oriented Computing (SOC) is a new outcome for the
development of services that can be reusable and loosely
coupled for creating flexible applications. Erl [10] defined
SOC as a platform for new generation computing that
encompasses service-oriented paradigm and service-oriented
architecture to create and assemble one or more services. Both
language and platform independent services can be addressed.
Papazoglou et al [18] considered SOC as a logical separation
of functionality in three main architectural layers namely,
service foundation, service composition and service
management and monitoring (semantics, QoS and properties
I
International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 2
ISSN 2250-3153
www.ijsrp.org
of NFR services). They also addressed the Service-Oriented
Architecture (SOA) as a logical approach to realize SOC.
Sommerville [19] considers SOA as a new technology for the
development of distributed applications as shown in Figure 1.
Figure 1: Service-Oriented Architecture [19].
Web services serve as a promising technology for
implementation of SOA. It uses Internet as the medium for
distributed computing and offers interoperability. Web
services are implemented by means of open standards such as
WSDL (Web Service Description Language), UDDI
(Universal, Description, Discovery and Integration). It allows
service requestors to browse and find the required service, and
SOAP (Simple Object Access Protocol SOAP) as shown in
Figure 1[19]. Sommerville [19], describes the design service
interfaces considering service operations. However, it lacked
the description of NFRs and service semantics.
Furthermore [22], entails business process driven software
development. The process relies on business process
understanding and analysis by taking into account functional
requirements and NFRs. Sometimes, composition of services
may also be required for complex business processes.
B. Non-Functional Requirement Classification
In recent days, customers not only concentrate about the
embedded functionality in DW services but they also pay
attention to the desired quality of services, such as reliability,
security, and efficiency. Such requirements are categorized
under NFRs.
In the attempt to develop a framework for software quality
characteristics, Boehm [4] proposes some form of NFRs
classification as depicted in Figure 2. Boehm also proposed
quality evaluation metrics in order to assess adherence to
NFRs. For example, low-level quality attributes are logically
implied as the guarding conditions for such type of adherence.
Using this tree, three questions are deduced: namely,(1) How
it is possible that the developed software can change its
environment?,(2) how adhering this software is with respect to
quality requirements such as reliability, efficiency and
usability and (3) Extent to which such software can be
maintained? In spite of such worthy classification provided by
Boehm, but the framework is still not complete and lacks
more reflective quantification techniques. Such type of
technique can be a service oriented paradigm. However, the
classification by Boehm can be considered as a landmark for
further refinements. Davis [7] introduced NFRs as non-
behavioural requirements and identified seven qualities as a
further refinement in the classification of Boehm.
Chung et al [6] proposed a process and goal based NFR-
framework, in which goals are decomposed and refined to sub
goals for arriving at low-level operationalisations soft goals by
the use of the SIG (Soft Goal Interdependency Graph). The
low level operationalisations denote some additional
functional requirements for satisfying the high level NFRs
through the “Label Propagation Algorithm”. NFR can have a
varied range such as “satisfied”, “weakly denied” etc. This
illustrates tradeoffs between the soft goals. Thus, this
framework may be considered as a comprehensive approach
for the determination of the satisfaction and tradeoffs of
between low-level NFRs. Burgess et al [5] proposed the
concept for SIG to develop Soft goal Interdependency Rule set
Graph (SIRG). SIRG will act as an automated technique for
determination of optimal set of the low-level
operationalisations to attain better NFRs’ satisfaction.
However, the SIRG requires some comprehensive case
studies.
The IEEE Standard [14] is a notable example to classify and
specify NFRs. Gilb [12] classified the requirements to
functions, qualities, costs and constraints [12]. Except the first,
the last three can be NFRs. Along this context, qualities can
denote “function performance” and constraint relate to
“restrictions on the freedom of requirements or can be
design”. Gilb’s classification emerged out of presence of
unwanted or undesirable requirements as specified in SRS
(Software Requirement Specification) documentation. Such
type of requirements may be false, unclear, and/or not possible
to assess satisfaction level.
Sommerville and Kotonya [15] introduced NFRs as a medium
of restrictions and constraints among system services.
However, this classification may be considered most
comprehensive in terms of NFRs types. They classified NFR
under three key categories i.e. product, process, and externally
related NFRs. The product category relates to the possible or
desired attributes that may be possessed by a system. The
second category relates to constraints and restrictions laid on
the development process over the system. Finally, the third
category relates to externally related NFRs
related to organisational regulations, national or international
standards, or can be interoperability requirements.
Glinz [13] further classified NFRs as performance and quality
related requirements. They described NFR using four facets:
kind,
representation, satisfaction, and role. Lamsweerde [21] later
classified NFRs according to QoS (Quality of Service),
compliance, architectural constraints and development
constraints as in Figure 3.
International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 3
ISSN 2250-3153
www.ijsrp.org
Figure 2:Boehm's Characteristics Tree [4].
Figure 3: Lamsweerde Non-Functional Requirements on Software Services [21].
QoS relates to quality requirements such as security,
performance, etc. Compliance relates to confirmation to
standards, organisational regulations and external standards
and laws. Architectural constraints implies structural
constraints in relation to the developed software along with its
operational constraints. Development constraints considers the
governance in developing the anticipated software, such as
maintainability, schedules, etc.
C. Service- Oriented NFR
In order to relate NFRs to service oriented development, it is
concerned with identification of QoS attributes, tradeoff
between them, and how they affect the development of
service-oriented applications. However, no classifications
attempts to have been reported in the literature to specialize or
arrive at a new classification of NFRs for service oriented DW
engineering. Galster and Bucherer [11] proposed NFRs
classification regarding services and service oriented systems.
O’Brien et al [16] introduced different quality attributes
affecting SOA and then discussed ten attributes related issues
along with recommended solutions in order to satisfy quality
characteristics. They also discussed tradeoffs between quality
attributes, for example: developers may be required to be
aware of the negative effect of security on performance and
interoperability, and the harm that can be caused due to the
absence of some quality attributes such as availability. They
observed the limitation of current standards [16] in addressing
quality attributes to attain the appropriate NFR specification.
As a number of standards have been developed for QoS (WS-
Reliable). However, a Service Level Agreement (SLA) is
needed to be established between the parties in order to attain
an agreement on the desired QoS. But no classification
provides an explicit classification of NFRs in relation to
service oriented DW engineering.
Later, Ameller and Franch [2] proposed a Service Level
Agreement Monitor (SALMon). It acts as a monitoring
technique on SLA. This technique is used to monitor services
in order to carry out the appropriate decisions when
unexpected actions occurs while adopting SOA in order to
meet SLA requirements. An ISO/IEC 9126 based
classification of the desired characteristics in relation to web
services as presented in Figure 4. It emphasis on the technical
and non-technical characteristics of NFR services. This quality
model was chosen because of its generic features and ability to
develop characteristic hierarchies. In this quality model,
characteristics chosen are refined and decomposed to sub
levels.
Galster and Bucherer [11], presented a taxonomy for NFRs
along with service-centric systems. They classified NFR into
three categories, namely: process requirements, service
requirements and external requirements. This proposal
denoted the ability to relate the taxonomy to the service level
International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 4
ISSN 2250-3153
www.ijsrp.org
as well as the system level. Dobson and Sommerville [9]
demonstrated an attempt of developing QoS ontology, namely
the QoSOnt for service-centric systems, but disregards domain
knowledge.
Figure 4: Quality Model [2].
III. PROPOSED NFR FOR SERVICE ENGINEERING
AND SERVICE ORIENTED DATA WAREHOUSE
ENGINEERING
This proposed classification shown in Figure 5 is designed on
the basis of NFR classification given by Kotonya and
Sommerville and Van Lamsweerde. In the proposed
classification, we addressed service engineering and DW-
oriented service engineering as high-level categories leading
to further sub classifications.
Figure 5: A Non-Functional Requirements Multilevel Taxonomy on Service-Oriented DW Engineering.
A. NFR for Service Engineering
The constraints are based on the desired QoS, compliance to
requirements, development standards, and architectural
constraints. However, this leads to a service which is reusable,
loosely coupled with independent operational requirements.
But, it adhers to related organizational, national/international
standards and regulations. The Service composition and
invocation applicable to QoS requirements concerns for
desired characteristics of a service that is running
independently. It also integrates other services at run-time.
The External service appreciation of Development
requirements are related to service engineering process stages
; starting with the service identification stage. It starts from
International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 5
ISSN 2250-3153
www.ijsrp.org
designing, servicing, implementation and deployment stages.
Compliance requirements belonging to Critical service
requirement implies for the conformation related to standards,
laws or rules, organizational regulations, etc . Internal service
appreciation related to Architectural constraints, are mainly
concerned with structural constraints or to service design.
They adhere to Erl’s [10] SOA principles namely, abstract
underlying logic, loosely coupled, reusable, autonomous,
services share standardised contracts, stateless, discoverable
and composable .
B. NFR for Service Oriented DW Engineering
The constraints are based on the QoS attributes, standards’s
compliance, the development process, and architectural
constraints. The benefits are provided by the Data warehouse
through the use of reusable services. Service engineering
NFRs are applicable to single service, while service-oriented
DW engineering NFRs considers the overall behaviour of the
application developed from reusable services.
The QoS attributes, and standards and regulation compliance
regulate the overall behavior of the DW-oriented application.
DW engineering process influences the NFR of development
and architectural constraints. If the construction of service
oriented application is done by the composition of service
process, then the development process NFR concerns for
requirements of workflow development, service discovering,
service selection, modelled workflow refinement , refined
workflow enactment and testing of services and the overall
application.
IV. CONCLUSION
The proposed classification in this paper has addressed the
non-functional requirements for both service oriented
engineering and service oriented DW-oriented engineering.
The latter is dependent on the former in terms of classification
of NFRs. This newly proposed classification is mainly based
on Sommerville and Lamsweerde and service oriented
development. The proposed classification is the basis for a
shared understanding of NFRs in both of the areas i.e. service
engineering and DW-oriented engineering. The NFR
classification facilitates the development of the semantics
behind the identified set of NFRs in every category and
subcategories of the hierarchy.
Future work can be done on completing the specification of
the classification. So, that it could serve as the general
ontology for engineering of service oriented DW applications.
Therefore, this proposed classification paves the path for
advancing the state of the art in the process of quantification,
for particular application domains.
REFERENCES
[1] Aburub F., Odeh M. and Beeson I., “Modeling Non-Functional
Requirements for Business Process”, Information and Software Technology
Journal, vol. 49, no. 11-12, pp. 1162-1171, 2007.
[2] Ameller D. and Franch X., “Service Level Agreement Monitor
(SALMon)”, in Proceedings the 7th IEEE International Conference on
Composition- Based Software Systems (ICCBBSS 2008), Madrid,
Spain, pp. 224-227, 2008.
[3] Baader F., Horrocks I. and Sattler U., “Description Logics for the
Semantic Web”, Kunstliche intellihenz, vol. 16, no. 4, pp. 57-59, 2002.
[4] Boehm B., Brown J. and Lipow M., “Quantitative Evaluation of Software
Quality”, in Proceedings of 2nd IEEE International Conference on Software
Engineering, California, USA, pp. 592-605, 1976.
[5] Burgess C., Krishna A. and Jiang L., “Towards Optimising Non-
Functional Requirements”, in Proceedings of the 9th International Conference
on Quality Software, Jeju, South Korea, pp. 269-277, 2009.
[6] Chung L., Nixon B. and Mylopoulos J., Non- Functional Requirements in
Software Engineering, Kluwer Academic Publishers, London, 2000.
[7] Davis A., Software Requirements: Objects, Functions and States, Prentice
Hall, Upper Saddle River, 1992.
[8] Dobson G., Hall S., and Kotonya G., “A Domain Independent Ontology
for Non-Functional Requirments”, in Proceedings of the. IEEE International
Conference on e-Busniess Engineering (ICEBE 07), Hong Kong, China, pp.
563-566, 2007.
[9] Dobson G., Lock R. and Sommerville I., “QoSOnt : A QoS Ontology for
Service-Centric Systems”, in Proceedings of the 31st EUROMICRO
Conference on Software Engineering and Advanced
Applications (SEAA 05), Porto, Portugal, pp. 80-87, 2005.
[10] Erl T., SOA: Principles of Service Design. Prentice Hall, Upper Saddle
River, 2007.
[11] Galster M. and Bucherer E., “A Taxonomy for Identifying and
Specifying Non-functional Requirements in Service-Oriented Development”,
in Proceedings of IEEE Congress on Services Part I, Honolulu, USA, pp. 345-
352, 2008.
[12] Gilb T., ”Towards the Engineering of Requirements”, Requirements
Engineering Journal, vol. 2, no. 3, pp. 165-169, 1997.
[13] Glinz M., “Rethinking the Notion of Non- Functional Requirements”, in
Proceedings of the 3
rd
World Congress for Software Quality, Munich,
Germany, pp. 55-64, 2005.
[14] IEEE, IEEE Recommended Practice for Software Requirements
Specification. IEEE standard 830-1993, 1993.
[15] Kotonya G. and Sommerville I., Requirements Engineering: Processes
and Techniques, Jhon Wiley, Chichester, 1998.
[16] O’brien L., Merson P. and Bass L., “Quality Attribute for Service-
Oriented Architectures”, International Workshop on Systems Development in
SOA Environment (SDSOA), Minneapolis, USA, 2007.
[17] OMG (2011) Business Process Model and Notation (BPMN) [online].
Available from: http://www.omg.org/spec/BPMN/2.0/PDF/.
[Accessed 28 July 2011 at 3:36 am].
[18] Papazoglou M., Traverso P., Dustdar S., and Leymann F., “Service-
Oriented Computing: State of the Art and Research Chanllenges”, IEEE
Computer, vol. 40, no. 11, pp. 38-45, 2007.
[19] Sommerville I., Software Engineering, Eight Edition, Addison-Wesley,
Harlow, 2007.
[20] Valipour M., Amirzafari B., Maleki K. and Daneshpour, N., “A Brief
Survey of Software Architecture Concepts and Service Oriented
Architecture”, in Proceeding the 2nd IEEE International Conference on
Computer Science and Information Technology 2009 (ICCSIT 2009),
Chengdu, China, pp.34-38, 2009.
International Journal of Scientific and Research Publications, Volume 3, Issue 7, July 2013 6
ISSN 2250-3153
www.ijsrp.org
[21] Van Lamsweerde A., Requirements Engineering: From System Goals to
UML Models to software specifications, Jhon Wiley, Chichester, 2009.
[22] Yousef R., Odeh M., Coward D., and Sharieh A., “BPAOntoSOA: A
Generic Framework to Derive Software Service-Oriented Models from
Business Process Architectures”, in Proceeding of the 2
nd
International
Conference of Applications of Digital Information and Web Technologies
(ICADIWT ’09), London, England, pp. 50-55, 2009.
[23] Abelló, A., Samos, J., Saltor, F. “Benefits of an Object Oriented
Multidimensional Data Model”. Lecture Notes in Computer Science, a. 1944,
pg. 141 ff, Proc. of Objects and Database 2000 (ECOOP Workshop), France,
2000.
AUTHORS
First Author – Anjana Gosain, Ph.D, University School of
Information Technology, GGS Indraprastha University, Delhi,
India. Email: [email protected]
Second Author – Sangeeta Sabharwal, Ph.D., Netaji Subhas
Institute of Technology, Delhi University, Delhi, India. Email:
Third Author – Rolly Gupta, M.Tech , Netaji Subhas Institute of
Technology, Delhi University, Delhi, India. Email:
Correspondence Author – Rolly Gupta, Email:
09899148597.
