Aryono, Teddy and Galinium, Maulahikmah and Suntana, Asep S. (2017) Modelling Network Marketing Semi-Structured Data Using Graph Database Management System. Masters thesis, Swiss German University.

Preview	Text Teddy Aryono 21551005 TOC.pdf Download (245kB) | Preview
	Text Teddy Aryono 21551005 1.pdf Restricted to Registered users only Download (216kB)
	Text Teddy Aryono 21551005 2.pdf Restricted to Registered users only Download (217kB)
	Text Teddy Aryono 21551005 3.pdf Restricted to Registered users only Download (852kB)
	Text Teddy Aryono 21551005 4.pdf Restricted to Registered users only Download (1MB)
	Text Teddy Aryono 21551005 5.pdf Restricted to Registered users only Download (168kB)
Preview	Text Teddy Aryono 21551005 Ref.pdf Download (168kB) | Preview

Abstract

Network Marketing Organizations (NMOs) are retailing selling channels that use independent distributors, to buy and resell product at retail and to recruit new distributors. When they recruit other distributors, it becomes their downlines and they earn referral fees. Compensation calculation is done by a system to serve real time information. Due to the network marketing model that require a heavily related data and shape a giant tree, the system needs to provide an efficient method to store the data and do calculations among them. This research simplifies a storing method for semi-structured data in network marketing by modelling a data structure of network marketing compensation schema into graph to be implemented and stored in Graph Database Management System. Graph database is a technology to store data with graph structure where each data stored in it is formed a node. Graph database is excellent at managing highly connected data and complex query. Performance measurement is measured from speed to run a query to retrieve data, how much storage occupied for particular amount of data, and how CPU and memory are utilized. Speed measurement is accomplished by running a query that is similar to the RDBMS machine experiment and experiment GDBMS machine. Storage occupation measurement is done by importing data to both engine and measure the physical data using Operating System (OS) utility tools. CPU and memory utilization is done by utilizing top command of UNIX while running the query. To run 61 levels of graph traversal, Neo4j is 1425.93 times faster than MySQL. However, the storage occupied by MySQL is 47.82% more efficient than Neo4j at storing 100,000 records/nodes. In term of CPU and memory usage, to run 61 levels of traversal, Neo4j needs 56.72% less CPU and 55.86% less memory to be consumed than MySQL.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Network Marketing; Graph Database; Relational Database; Neo4j; Cypher Query Language,
Subjects:	Q Science > QA Mathematics > QA76 Computer software
Divisions:	Faculty of Engineering and Information Technology > Department of Information Technology
Depositing User:	Astuti Kusumaningrum
Date Deposited:	13 May 2020 14:50
Last Modified:	13 May 2020 14:50
URI:	http://repository.sgu.ac.id/id/eprint/338

Actions (login required)

View Item