An Approach based on user preferences for Selecting SaaS Product Nacera Boussoualim
Youcef Aklouf
Univ. of Science and Technology of Houari Boumedienne(USTHB) Email:
[email protected]
Univ. of Science and Technology of Houari Boumedienne(USTHB) Email:
[email protected]
Abstract-Software as a Service (SaaS) is a software delivery
assigning judgmental weights, it is quite probably that the us�r's judgment is centered on the principal parameters only. ThiS would lead to an inaccurate priority and to weights assigned to incorrect parameters. To make a decision, it is necessary to have quantifiable values instead of having subjective opinions. The paper we propose presents a method to help the users to choose the best SaaS products satistying the majority of their conditions and alternatives. Section two will describe the techniques used which are: the AHP (Analytic Hierarchy Process] [5] used for the defmition of the hierarchy between the various factors and attributes; and a competency model for the evaluation of these factors and attributes. Knowing that a good method of adapted selection must be based on the good defmition of the various parameters of choice, we will present in section three various parameters implied in the process of the selection of a SaaS applications.. In section four, we explain our selection algorithm and the calculation method of the various weights of attributes. After that, we apply this algorithm in section five to illustrate its operation through a case of study. Finally, we present a conclusion and some prospects in section five.
paradigm in which the product is not installed on-premise, but it is available on Internet and Web. The customers do not pay to possess the software itself but rather to use it. Hence, we see increasing number of organizations adopting SaaS. However, each customer is unique, which leads to a very large variation in the requirements off the software. As several suppliers propose SaaS products, the choice of this later becomes a major issue. When multiple criteria are involved in decision making, we talk about a problem of «Multi-Criteria Decision Making» (MCDM). The paper we propose presents a method to help customers to choose a better SaaS product satisfying the most of their conditions and alternatives. Since that a good method of adaptive selection should be based on the correct definition of the different parameters of choice, we started by extraction and analysis of the various parameters involved in the process of the selection of a SaaS application. Keywords: Criteria
Cloud
Computing,
Decision-Making
business operation,
(MCDM),
Software
as
a
Multi Service
(SaaS).
I.
INTRODUCTION
Conceived to increase the advantages brought by the economy of scale, SaaS (Software as a Service) is the third form of the Cloud Computing. SaaS is the way in which software functionalities must be provided to a large group of customers through the Web with a single instance of application running on a multi-tenant's platform [1]. Customers usually don't need to buy a software license and install it on their local environment. They employ only the software qualifications published by the suppliers of SaaS to constantly consume the service through the Web by using an Internet navigator whenever they want and wherever they are. The mode of payment follows a subscription model [2]. Identity applicable sponsorls here. If no sponsors, delete this text box. (sponsors) When several suppliers offer SaaS products, the choice of the product becomes a principal issue that requires the analysis of the parameters of choice on one hand, and the study of the different offers of the various suppliers on the other hand [3]. Since multiple criteria are implied in the decision-making, then we have a Multi-Criteria Decision Making (MCDM) problem [4]. The multi-criteria and multi-products problems cannot be solved with intuition or by a simple judgment. The judgment can function very well, but only when the number of parameters of choice is minimal. During the process of selection, the parameters are usually arranged according to their priority. The assignment of the priorities to the products requires the decision of the weight of each parameter. While
II.
RELATED WORKS
In business sector, the software is conceived to support several operations such as: commercial data management , the automation and the optimization of the business process. SaaS is not an exceptional case even if it provides multiple ad�antages[6] [7] [8]. However, each customer is unique, which leads to a very great variation of the requirements concerning software. Among the fundamental causes of the variation of the requirements between the customers, we fmd: the different focalization of industry, the different behavior of the customers, the various ways in which the products are offered, the different regulations, various cultures and various operational strategies. In this section we present some academic related work. In "An Approach for Selecting Software-as-a-Service (SaaS) Product," the authors proposed an approach to help the companies to choose an SaaS product more appropriate to their needs [3]. For that, They proposed various parameters such as the architecture and the cost. They have chosen the SFA « sales force automation» as an application domain. However, other parameters such as configurability and personalization were not taken into account though they are key parameters in the process of the choice of a SaaS software. Their approach is also not generalizable on all the categories of SaaS applications. In the field of the Web Service "WS", the concept of Adapted WS "AWS" was introduced in 2005 by Lopez Velasco, Villanova-Oliver, Gensel and Martin [9].The initial
978-1-4799-3824-7/14/$31.00 ©2014 IEEE
goal of the AWS is to invoke a service with particular conditions of use. An extension of WSDL, called AWSDL "Adapted Web Service Description Language", was also introduced in [10]. This extension makes possible to describe the conditions for which AWS are adapted. It is also used by the users to formulate their needs and their use context. Although the authors have dealt with the problem of adaptation and the selection of a web service by taking into account the characteristic of users and the conditions of use, they did not explicitly study the latter. III.
Analytical hierarchy process
SaaS selection based solely on judgment is a highly cognitive and tedious process which could be quite error prone. Humans are supposed to be very good at one to one comparison. If a problem is decomposed into clusters, and attributes are compared pairwise within the cluster, then decision problems can be solved easily with reduced cognitive load. Saaty developed the Analytic Hierarch Process (AHP) method, which is very useful in simplifying multi-criteria problems into hierarchy thus forming the comparison matrix to judge the weight. The AHP breaks down a problem into its smaller constituent parts forming hierarchy and then calls for only simple pair-wise comparison judgments[5][11]. The AHP process starts with hierarchy development.An advantage of hierarchy is that it allows focusing judgment separately on each of the several properties, which is essential for making a sound decision. Each element is compared with every other element to decide the importance of one element over the other. The elements at every level of hierarchy are also compared in a similar way. B.
COMPETENCY MODEL
contain the attribute
Low
Offer X contains the attribute but with a low qualification level.
�
VAiIF(Xr1
TECHNIQUES
We noted that the choice of the best SaaS satistying the majority of the available conditions and alternatives is an MCDM problem, where we need to understand the priorities of usage conditions one hand and to study the attributes of offers on the other hand. So, our work suggests the use of AHP as a technique to tackle this subject. Thus, the applications are evaluated mainly on the basis of these attributes which are a list of functional modules and preferences (defmed in the following section). However, the latter do not have necessarily the same qualification leveL This is why we introduced a competency model in order to evaluate the SaaS applications and the different parameters of choice. A.
TABLE I.
The competency model
The evaluation of applications is based mainly on the evaluation of their various parameters (see figure 1). However, these parameters have not necessarily the same qualification level for all SaaS offers. This is the reason why we introduced a competency model in order to evaluate the SaaS applications. We defmed five qualification levels which are: None, Low, Medium, High and Very high. This model can also be used to identity the objectives of the improvement of applications through a comparative analysis between the qualification level of the other suppliers services.
medium qualification level. High
Offer X contains the attribute with a high qualification level.
Extremely high
Offer X contains the attribute with a very high qualification level.
We defined five different values which are: 0, 1, 2, 3 and 4 for each qualification level. These values will be used for the calculation of the weights. IV.
THE
SELECTION
CRITERION
Within the framework of the process of registration, the providers can use "Interface Attributes" to facilitate their recording. It is a form Web that is geared towards helping the web service providers provide more meaningful Interface Attributes about their SaaS applications. These attributes are aimed to attract new subscribers. To start, we have to identity the two most known categories of SaaS applications which are: • Line-of-business services, offered to enterprises and organizations of all sizes. Line-of-business services are often large, customizable business solutions aimed at facilitating business processes. In this category, we mainly fmd the applications of: collaboration, Content Management (CM), office software, Customer (CRM), Business Relationship Management Intelligence(BI), Enterprise Resource Planning (ERP) and Supply Chain Management (SCM). These services are typically sold to customers on a subscription-basis [12][13][14][15]. • Consumer-oriented services, offered to the general pUblic. Consumer-oriented services are sometimes sold on a subscription-basis, but are often provided to consumers at no cost and are supported by advertising [14][15]. Our work will be focused mainly on the category of Line-of-business services. Thus, we have designated two criteria which are: the functionality and the list of preferences so that each criterion is divided into several factors and each factor is refmed in turn by several attributes.
•
•
The first criterion, which is functionality, includes attributes typically considered as functional modules. This is the reason which leads us to designate business application classes as factors. The second criterion, which is the preferences, includes the following factors: reputation of supplier {l6J [17J, cost [3J, architecture[J8], usability [3J, configurability and personalization.
To recapitulate, we sununarize all the factors and attributes through a diagram (see figure 1).
B.
Weight calculation
By using the competency model previously described, we can evaluate the attributes of the collaboration factor as follows:
A.
SELECTION METHOD OF
SAAS PRODUCT
Algorithm
The methodology that we adopt to make the choice of a SaaS product is divided into three parts: 1.
The first part covers the calculation of the weights of the attributes according to the user's choice. So, we have to: a)
b)
c)
2.
Assign weights to the various factors according to the preferences such as: L Pdc) =1 Assign weights to the various attributes of each factor PAlF according to preferences such as: L PAi/F (c) =1 / i= l...n and n : is the number off attribute/factor. calculate the net weight PEAlc of each attribute according to the customer preferences : PEAlc = PAiF (c) * PF (c)
The second part calculates the global weights of the parameters and attributes of the different suppliers products. a)
Assign weights to the various factors for each SaaS product such as:
b)
Assign weights to the various attributes of each factor PAlF of all SaaS products:
c)
calculate the net weight PEAIF of each attribute
L PF(p) =1
L PAilF(P) =1
PEAlP= PAlF (P)* PF(p) 3.
the third part combines the results obtained from the two parts to arrange the products according to preferences. a)
b)
c)
For each SaaS product, we calculate P Alpref based on user preferencies in the following way: PAlpref = I PEAle- PEAlP I Then, we calculates PP wich is the sum of P Alpref for each product.
PP = L PAi/pref
The ranked sum of weighted scores PP in descending order gives the ranking of the products.
Low
None Communication Conference
V.
EVALUATION OF THE COLLABORATION ATTRIBUTES.
TABLE II.
�
�
Medium
High
Very high
..... 2
o ec::::
�
Coordination
re3
J
We can have (3) collaboratIOn applIcations wIth varIOUS qualification levels. For example, we can have an application X which offers a communication module with a medium level and a coordination module with a high level but no conference module. The weight P AilFiX) of the attribute Ai of the factor Fj is given by the following formula: (X) = (X) I .n��VAl
