| Summary: | 碩士 === 國立臺灣大學 === 資訊管理學研究所 === 100 === In the past, individuals and enterprises are usually the main subjects in the area of information security. Now the issue about information warfare between nation-sates is getting much attention. When discussing the resource allocation based on the subject of a nation-state, except for the allocation of defense resources, the resources allocated on attack should also be concerned. Historically, preventive strike and the corresponding retaliation from another nation-state are common in the war between two nation-states. In addition, there would be various information experts launching an attack together for a nation-state, which is called collaborative attacks that different from the situation of only one attacker in an ordinary cyber attack. Therefore, we consider two players that could attack and defend simultaneously and adopt the concept of collaborative attacks in our research model.
How to efficiently evaluate the network survivability is an important issue and worthy of discussion. In this thesis, the Average Degree of Disconnectivity (Average DOD) metric is adopted to measure the network survivability. The Average DOD combines the concept of probability with DOD metric to evaluate the damage degree of the network. The larger the Average DOD value, the higher the damage degree of the network. In our scenario, there are two players who have the dual-roles as an attacker and a defender; furthermore, both of them do not know the vulnerability information about their networks. However, the counterpart knows some. Therefore, after being attacked, they would update their vulnerabilities information and patch the vulnerabilities.
We develop a multi-round network attack-defense scenario, and establish a mathematical model to optimize resource allocation and then predict their own network survivability by the Average DOD. In each round, the players could allocate their attack resources on the nodes of their own network and on another player’s network after updating related information about another player’s. Furthermore, they could reallocate existing defense resources and repair compromised nodes. To solve the problem, the “gradient method” and “game theory” would be adopted to find the optimal resource allocation strategies for both players.
|
|---|
