Widianto, Ahmad Rinaldi and Lim, Charles and Eng, Kho I (2015) Improving Performance of Intrusion Detection System Using General-Purpose Computing on Graphic Processing Unit (GPGPU). Bachelor thesis, Swiss German University.

Preview	Text Ahmad Widianto 12111019 TOC.pdf Download (240kB) | Preview
	Text Ahmad Widianto 12111019 1.pdf Restricted to Registered users only Download (174kB)
	Text Ahmad Widianto 12111019 2.pdf Restricted to Registered users only Download (483kB)
	Text Ahmad Widianto 12111019 3.pdf Restricted to Registered users only Download (663kB)
	Text Ahmad Widianto 12111019 4.pdf Restricted to Registered users only Download (605kB)
	Text Ahmad Widianto 12111019 5.pdf Restricted to Registered users only Download (155kB)
Preview	Text Ahmad Widianto 12111019 Ref.pdf Download (191kB) | Preview

Abstract

The development of computer network and the Internet enables more machines to be- come connected. Together with mobile computing boom increases the number of data passed via network. More and more data stored in the cloud to meets mobile demands. This leads to even higher security threats. On the other hand, intrusion detection tech- nology are mostly single threaded application. This solution works in the past where most computers are single threaded machines. With current changes in computing paradigm and the increase in network traffic this method becomes obsolete. The fo- cus of this research is to develop a parallel intrusion detection by exploiting GPGPU. In parallel computing packets are processed concurrently, thus increasing system ca- pacity in high throughput environment. GPGPU further exploit parallelism inside mod- ern GPU, which is essentially SIMD engines. OpenCL will be used as programming library during this research. The result are in favour of GPU with original design im- proving performance by average of 3.51x against single-thread CPU implementation, and by average of 2.54x against multi-thread implementation. After some optimization the overall performance improved on the average by a factor of 11.21 compared to the older version. The performance advantage against CPU improve by a factor of 23.71 vs a factor of 12.28 faster than multi-thread implementation.

Item Type:	Thesis (Bachelor)
Uncontrolled Keywords:	Intrusion Detection; Graphic Processing Unit; GPGPU; Parallel Computing; OpenCL
Subjects:	T Technology > T Technology (General) > T58.5 Information technology T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105 Web Sites-Design > TK5105.59 Intrusion detection systems (Computer security, Firewalls
Divisions:	Faculty of Engineering and Information Technology > Department of Information Technology
Depositing User:	Atroridho Rizky
Date Deposited:	20 Jan 2021 14:42
Last Modified:	20 Jan 2021 14:42
URI:	http://repository.sgu.ac.id/id/eprint/1713

Actions (login required)

View Item