Decision Support System for Cloud Computing Service Selection Using the Weighted Product Method

. The selection of cloud computing services requires careful consideration and review. Some aspects of the criteria that must be considered such as direct selection between services that take a long time and need to be done repeatedly. Decision Support System with Weighted Product (WP) method is an effective method because the time needed for calculation is much shorter. The purpose of this research is to apply Weighted Product (WP) method in the decision support system to choose cloud computing services where as a case study is PT Deptech Digital Indonesia, so that it can make it easier for companies to make decisions according to their needs. The calculation results using the WP method give preference to the top 3 services: Google Cloud, Amazon Web Services and Microsoft Azure. This proves that research with the WP method can be applied to various services that will be used in the future according to predetermined criteria. Based on these results, the system can recommend cloud computing services according to the needs and a good level of accuracy.


INTRODUCTION
Cloud computing service is a set of technologies that serves to empower the internet as a server center used to manage data as well as user applications [1]. Utilization of cloud computing does not require special expertise because the entire infrastructure or information system is controlled by cloud computing devices [2], to manage digital-based services over the internet network. Making users feel comfortable in accessing our website or digital services on the internet is a difficult problem, therefore the appropriate infrastructure will certainly be one of the main keys in providing a good user experience to users. Because according to many experts say that the user experience offered is an indicator of the success of a At this stage what is done is to find classification, decomposition, and ownership of problems. At this stage, there is also a process of tracing and detection of the scope of the problem as well as the process of problem recognition. The data that has been obtained is then processed and tested in order to identify the problem.

Design Phase
The design stage has several processes including the creation, development, There are two types of approaches at the election stage, namely: a. Analytical technical uses mathematical formulation b. Algorithm that describes the process step by step 4. Implementation Phase This stage is the process of implementing the system design made at the stage of design and implementation of alternative actions selected at the selection stage. Weighted Product (WP) method requires a normalization process because this method multiplies the results of the assessment of each attribute [13] [14] i . According to D. Khairina in [11] Weighted Product is one of the methods used to complete Multi Attribute Decision Making (MADM), by using multiplication technique to connect rating attribute, which needs to be done lifting with the corresponding weight attribute for each attribute rating Here are some steps to take to troubleshoot using weighted product: 1. Normalize or improve weights. (1) Normalize to produce the value wj = 1 where j = 1, 2,..n is a lot of alternatives and ∑wj is the total value of the weight.
2. Determine vector values by multiplying all criteria by alternative normalization or improvement of weights of positive rank (benefit) and negative rank for cost criteria . (2) With I = 1, 2,....,m and j = 1, 2,....,n. Description: Si : Score / value of each alternative X ij : Alternate value to i to attribute to j In determining vector value (S) can be done by multiplying alternative normalization results with all criteria with or improvement of weights that rank positive for benefit criteria and negative rank for cost criteria. Where (S) is the prefence of criteria, (x) is the value of criteria and (n) is the number of criteria.
3. Determine the vector value, where this vector is an alternative preference used for warkingan.
( Determining the weight value of each alternative on each criterion to show the match rating of each alternative in each criterion is assessed by 1 to 5 namely: According to Muhammad Bakri's presentation to [2] Cloud Computing can be identified as a model that allows us to access everywhere, network access is done jointly in accordance with the demands of configurable computing resources and can be quickly established and released with minimal service provider management and interaction efforts.
The cloud computing architecture does not focus on "how" solution design and implementation, but rather focuses on the "what" services are provided. The reference International Journal Of Science, Technology & Management ISSN: 2722-4015 http://ijstm.inarah.co.id architecture aims to provide an understanding of the ins and outs of operations in cloud computing services. It does not represent the system of a particular cloud computing architecture, as a tool for discussing, describing, and developing a cloud computing architecture that uses a general frame of reference.
Here's the purpose of the cloud computing reference architecture design: 1. To understand and describe cloud services in the context of cloud computing as a whole.

2.
To provide technical references for institutions and other consumers to discuss, understand and compare cloud services. 3. To facilitate in the analysis of interprobability and probability, security, and reference implementation.

Fig1. Examples of cloud computing architecture
From what we see in the picture above, you can understand that cloud computing architecture consists of: This infrastructure includes all hardware/hardware and supporting software/software, e.g. network infrastructure devices, servers, and storage media.

Platform
A combination of hardware and software architecture, consisting of storage objects, user identity or information, running programs in use, data queues, and database systems used, is also inseparable from the hardware infrastructure that supports the platform.

Application
Consists of a variety of applications for monitoring, content management, communication, and many more that can be used according to their respective functions and needs. Each of the above components must support each other in order to function properly and run as needed [15]. Previous research related to Cloud Computing services can be referenced as follows: 1. Research entitled Comparison Analysis of Azure Cloud System And Google Cloud explains that the use of the operating system as a link between hardware media and ISSN: 2722-4015 http://ijstm.inarah.co.id software as a controller aimed at avoiding errors in performing the operating steps of a system. The purpose of this research is to compare several points of technology, system development, integrated products as well as the quality of each service. Cloud computing may still be vaguely audible to laypeople. Therefore, the comparison done by researchers between the two services is expected to help the public in understanding what cloud computing services are and the shortcomings and advantages of each service, however, these two azure and google cloud services have many similar features and have their own advantages and disadvantages that we cannot compare in general [1]. 2. Research entitled Comparison of Microsoft Onedrive Cloud Computing, Dropbox,and Google Drive explains that c loudcomputing is one of the latest webbased information technologies that provide services capable of providing resources and services. The purpose of this research is to better give an overview to the wider community about the detailed function of each service and is expected to provide convenience in supporting the industrial revolution in this new round. The era of industrial revolution 4.0 is expected all activities can be done by utilizing cloud computing technology and supporting ioT or internet of things. Cloud computing services currently face obstacles yatiu, the lack of maximum use of the services provided because the public does not understand the features that have been provided, such as shared media storage on google drives that can be used to share files in the cloud with other users who are given access, but the public still chooses to share files manually such as share via social media. Therefore, the results of this research are appropriately used to further improve the detailed understanding of the features provided by service providers are very important to support the industrial revolution 4.0 which is expected that the entire work process can be done through the cloud system [6].

Research Instruments
The types of instruments used in this research are: 1. Observation This method is done by visiting, and observing PT. Deptech Digital directly to obtain the required data.

Wawancra (Interview)
This method is done by conducting a question and answer session by asking some questions orally to the Head Of Software Developer section, to collect the required information.

Literature Review
This method is done by reading and analyzing some research related to cloud computing services to understand and strengthen the points of information that has been obtained before.

Data Collection Method
The method used to collect data on this research is done by observation, interview and research on internet sources to obtain information services to be analyzed.

Population and Sample Research
In conducting population data collection and sampling in this research, conducted through direct observation and interview with the company through conference media, the population taken is a list of cloud computing services commonly used by companies as many as 10 cloud computing services. Of these populations were taken 6 samples of the most frequently used services.

Data Analysis Methods
In conducting analytical research is an important part in the methodology of scientific research, because by analyzing a data we can get meaning and meaning that is useful in solving a form of problem. The method used in this research is Weighted Product (WP) is a method of decision making that is calculated quantitatively using mathematical calculations according to weighted product formula.

Weighted Product (WP) Calculations
The Weighted Product method is one solution to the decision support system. This method evaluates several alternatives against a set of attributes / criteria, where each attribute is independent of one another. According to Yoon, the Weighted method Product uses the multiplication technique for relates the attribute rating, where the rating of each attribute must first rank with the attribute weight concerned. This process is the same as a process normalization.
In determining the services to be used in the initialization process of pt Deptech Digital Indonesia project, usually using 6 types of cloud computing service providers,which in this research we select as a sample of weighted product method calculation, which will then be processed according to the stages in the Weighted Product method. The stages of the Weight Product method are:

a. Determination of Criteria Weight
In determining the weighting of criteria on the selection of pt cloud computing services. Deptech Digital Indonesia, there are 11 (eleven) criteria that become consideration material, namely: server location, service products, operating system, downtime, backup service, service price, customer support /, security, scalability, reputation, and network speed.
Weight assessment indicators on each criteria are taken using quality scales that can be seen in Table 2.

c. Conversion of Alternative Values (Services)
Determination of the match rating of each service with each criterion. , This criterion is used to consider the server location based on the needs of the project and the nature of the data to be stored C2: Service Products (Benefit), This criterion is used to assess the different types of services that can be used to support the performance and performance of the application. C3: Operating System (Benefit), This criterion is used to consider the required operating system applications, and the operating system used on the server. C4: Downtime (Benefit), This criterion is used for historical downtime indicators or times when the service cannot be used. C5: Backup Service (Benefit), This criterion is to assess whether the cloud computing service supports backup services or not, especially in favor of snapshots. C6: Service Price (Cost), This criterion is used to determine the price of the service, depending on needs and conditions. C7: Customer Support/Warranty (Benefit), This criterion is an assessment of support services from service providers in case of constraints along with the channels offered. C8: Security (Benefit), This criterion is an assessment of the security aspects used by cloud computing service providers to support corporate data security.
C9: Scalability (Benefit), These criteria assess how cloud computing service providers support in improving specifications, to improve the performance of the services we use. C10: Reputation (Benefit), Reputation assessment criteria or service rating from a cloud computing service provider. C11: Network Speed (Benefit), Network speed criteria offered by service providers.

d. Level of Interest
Weight determination for criteria is determined based on the value of the interest level scale where the value of the level of interest starts from 1 to 5. Interest level scale data can be seen in table 6. Table 6. Importance Scale Value Importance Scale 1 Highly Ins important 2 Not Important (TP) 3 Simply Important (CP) 4 Important (P) 5 Very Important (SP) Here is a list of initial weight values (W) of each criteria that have been determined by the PT. Deptech Digital Indonesia based on the value of interest level scale that can be seen in table 7.

e. Weight Improvement Criteria
After the weight value for each criterion has been determined. Furthermore, weight improvement will be carried out from the predetermined initial weight value. The calculation process to get the results can be seen in the calculation as follows: 1. The result of calculation on the weight of server location criteria with a weight value of 4 is 0.9.
2. The result of calculation on the weight of service product criteria with a weight value of 4 is 0.9.
3. The result of calculation on the weight of operating system criteria with a weight value of 5 is 0.11. 11. The result of calculation on the weight of network speed criteria with a weight value of 5 is 0.11.
After the process of calculating the improvement of the initial weight value on each criterion, here is a summary of the calculation results that can be seen in Table 8  Table 8 After the improvement of the initial weight value, the next process will be continued to the calculation of vector value (S) where each of the initial weight of the criteria will be raised with the weight of the improvement result for each alternative, where the calculation is as follows: After the calculation of vector value (S) is done the following is the result of the preference of vector value (S) of each alternative along with the total amount that can be seen in Table 9.

a. Vector Value Calculation (V)
At this stage, the calculation that has been done to get vector value (S) will be re-processed to perform data that has the highest result based on vector value calculation (V). The results of the calculation are obtained by the following calculation process: After calculating the vector value (V) for each alternative, here's a summary of the results that can be seen in Table 10. After passing a series of calculation processes that have been done at several stages above, it is found the results of the ranking that can be seen in table 11 below:

IV. CONCLUSION
With the results of this research is expected to help companies in conducting selection of cloud computing services. Weighting the criteria will accelerate the selection process by conducting a thorough comparison according to the criteria set out in each alternative, this will certainly cut the comparative time performed directly against each service. With this research, companies can develop it by adding additional criteria or alternatives. The calculation results using the WP method give preference to the top 3 services: Google Cloud (A5), Amazon Web Services (A1) and Microsoft Azure (A2). This proves that research with the WP method can be applied to various services that will be used in the future according to predetermined criteria. Based on these results, the system can recommend cloud computing services according to the needs and a good level of accuracy.