國立虎尾科技大學 |

Attack Graph Analysis Based on Markov Decision Process.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Attack Graph Analysis Based on Markov Decision Process./
作者:	Li, Ming.
面頁冊數:	1 online resource (93 pages)
附註:	Source: Masters Abstracts International, Volume: 56-04.
Contained By:	Masters Abstracts International56-04(E).
標題:	Electrical engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9781369724912

Attack Graph Analysis Based on Markov Decision Process.
Li, Ming.

Attack Graph Analysis Based on Markov Decision Process. - 1 online resource (93 pages)

Source: Masters Abstracts International, Volume: 56-04.

Thesis (M.Eng.)--The University of Tulsa, 2017.

Includes bibliographical references

An Attack Graph (AG) is an abstract representation of all the existing paths attackers can take to compromise a network. The quantitative analysis of an AG can be based on using probabilities, and the analytical results can indicate the paths that attackers are more likely to choose and the hosts in the network that are highly valuable goals for attackers. Markov Decision Processes (MDP) are one of the probability based analyses. By introducing rewards and alternatives, an AG can be fully described by a MDP. This thesis uses insertion of states to implement the conversion from an AG into a MDP, and then uses value iteration and policy iteration methods to conduct reward analysis on MDPs. The output of reward analysis is the Maximum Total Expected Reward (MTER) and the Optimal Policy Vector (OPV). For the analysis of an AG, the OPV can indicate specific transitions on an attacking path that may bring the MTER to attackers. Security engineers can use it to find potential security holes in the network and take preventive measures.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9781369724912Subjects--Topical Terms:

596380
Electrical engineering.
Index Terms--Genre/Form:

554714
Electronic books.

Attack Graph Analysis Based on Markov Decision Process.
LDR:02721ntm a2200361Ki 4500 001 919565
005 20181129115238.5
006 m o u
007 cr mn||||a|a||
008 190606s2017 xx obm 000 0 eng d
020 $a 9781369724912
035 $a (MiAaPQ)AAI10270833
035 $a (MiAaPQ)utulsa:10196
035 $a AAI10270833
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Li, Ming. $3 898736
245 1 0 $a Attack Graph Analysis Based on Markov Decision Process.
264 0 $c 2017
300 $a 1 online resource (93 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 56-04.
500 $a Adviser: Peter J. Hawrylak.
502 $a Thesis (M.Eng.)--The University of Tulsa, 2017.
504 $a Includes bibliographical references
520 $a An Attack Graph (AG) is an abstract representation of all the existing paths attackers can take to compromise a network. The quantitative analysis of an AG can be based on using probabilities, and the analytical results can indicate the paths that attackers are more likely to choose and the hosts in the network that are highly valuable goals for attackers. Markov Decision Processes (MDP) are one of the probability based analyses. By introducing rewards and alternatives, an AG can be fully described by a MDP. This thesis uses insertion of states to implement the conversion from an AG into a MDP, and then uses value iteration and policy iteration methods to conduct reward analysis on MDPs. The output of reward analysis is the Maximum Total Expected Reward (MTER) and the Optimal Policy Vector (OPV). For the analysis of an AG, the OPV can indicate specific transitions on an attacking path that may bring the MTER to attackers. Security engineers can use it to find potential security holes in the network and take preventive measures.
520 $a In the implementation of reward analysis, this thesis uses a hybrid programming model which combines Message Passing Interface (MPI) and Open Computing Language (OpenCL) on a heterogeneous computing cluster. The performance test conducted on the high performance computing cluster indicates that the parallel program based on the model can efficiently solve MDP with up to 1,000,000 states, which satisfies the demands to analyze today's large-scale AGs.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Computer engineering. $3 569006
650 4 $a Computer science. $3 573171
655 7 $a Electronic books. $2 local $3 554714
690 $a 0544
690 $a 0464
690 $a 0984
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The University of Tulsa. $b Electrical Engineering. $3 1194175
773 0 $t Masters Abstracts International $g 56-04(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10270833 $z click for full text (PQDT)