Given the increasing cooperation between organizations, the flexible exchange of security information across the allied organizations is critical to effectively manage information systems (IS) security in a distributed environment. In this paper, we develop a cooperative model for IS security risk management in a distributed environment. In the proposed model, the exchange of security information among the interconnected IS under distributed environment is supported by Bayesian networks (BNs). In addition, for an organization’s IS, a BN is utilized to represent its security environment and dynamically predict its security risk level, by which the security manager can select an optimal action to safeguard the firm’s information resources. The actual case studied illustrates the cooperative model presented in this paper and how it can be exploited to manage the distributed IS security risk effectively.
With the increasing of collaboration between organizations, the management of information systems (IS) security risk is distributed across the allied organizations and the cooperative activities between organizations are imperative [
In this paper, a cooperative model for IS security risk management is proposed to estimate the risk level of each associated organization’s IS and support the decision making of security risk treatment in a distributed environment. In the model, the exchange of security information among the interconnected IS is achieved through Bayesian networks (BNs). Moreover, a BN is also exploited to model the security environment of an organization’s IS and predict its security risk level. However, it is difficult and critical task for a security manager to establish an appropriate BN, which is suitable for the environment of organization’s information systems [
The remaining sections of this paper are organized as follows. We first review the relevant literature in Section
There has been increased academic interest in the IS security risk management. From the technical literature, the security protocols [
In recent years, a new managerial perspective on IS security has emerged from the literature. This perspective focuses on the managerial processes that control the effective deployment of technical approaches and security resources to create a secure IS environment in an organization. From this perspective, Feng and Li [
Bayesian networks (BNs), also known as probabilistic belief networks, is a knowledge representation tool capable of representing dependence and independence relationships among random variables [
In this paper, the developed BN is not only used to facilitate the dynamical prediction of the security risk level of an organization’s IS, but also exploited to model the IS security environment.
In a distributed environment, the proposed model consists of many interconnected network information systems. We call these network information systems as “associated members.” Each associated member is installed with three kinds of components: monitor component, estimation component, and treatment component. Besides, the above three kinds of components, the registration component contains the information about each estimation component. It is required that all estimation components in the distributed environment must register with the registration component. The cooperative model architecture is demonstrated in Figure
Model architecture.
The interactions among the estimation component and the registration component are shown in Figure
Information exchange in the interactive process.
Exchange information | Description |
---|---|
Search request | It consists of the requester’s id, IP address, and the required input variables. The estimation component has a set of sharing variables. To find components capable of providing required input data, the estimation component sends a search request to the registration component. |
Search reply | It consists of the requested variable name, the IP address, and status of the component publishing the variable. Based on receiving a search request, the registration component searches its database to determine which components can provide the requested variables. |
Registration message | It consists of component id, IP address, list of published variables, and their possible states. Each estimation component registers with the registration component, which issues an acknowledgment message for entering the new component in its database. |
Communication between estimation components | It consists of the request id, the sender’s id, and the probability distribution of the requested variable. Upon receiving the list of components capable of providing the required input from the registration component, the request component sends requests directly to these components. Then, the sender sends the probability distribution of the requested variable. |
Interactions among the components.
In this section, we present an algorithm based on ant colony optimization (shown in Algorithm
(1) (2) (3) (4) (5) (6) (7) (8) Select two indexes (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) Update pheromone according to ( (21) (22) (23)
The equations appearing in the algorithm are as follows.
(1) Heuristic information:
(2) Updating rule:
(3) Probabilistic transition:
In this section, the proposed model is applied to a distributed environment, which is composed of four associated members with interconnected IS: two suppliers (S1 and S2), a manufacturer (M1), and a downstream transporter (DT1).
Based on the algorithm presented in Section
BN information of M1.
Node ID | Node name | State space | Parent nodes | Children nodes |
---|---|---|---|---|
M1_1 | Network access control |
|
|
|
M1_2 | Network security audit |
|
|
|
M1_3 | Change management |
|
|
|
M1_4 | Supplier threat level |
|
|
|
M1_5 | Transporter threat level |
|
|
|
M1_6 | Operational procedures and responsibilities |
|
|
|
M1_7 | Network security |
|
|
|
M1_8 | External systems security |
|
|
|
M1_9 | Operation security |
|
|
|
M1_10 | M1 threat level |
|
|
|
BN information of S1.
Node ID | Node name | State space | Parent nodes | Children nodes |
---|---|---|---|---|
S1_1 | Communication secrecy |
|
|
|
S1_2 | Audit logging |
|
|
|
S1_3 | Network access control |
|
|
|
S1_4 | Network security audit |
|
|
|
S1_5 | Network security |
|
|
|
S1_6 | Communication security |
|
|
|
S1_7 | S1 threat level |
|
|
|
BN structures of M1 and S1.
The manager interface of our proposed model is shown in Figure
Security manager interface.
Once the new evidence is obtained through the monitor components, the estimation component is able to make the BN modify its own belief (probability distribution on variable of risk level) in real time and exchange the update of beliefs of the security state with other associated members.
In a distributed environment, in order to effectively manage information systems (IS) security, a cooperative model based on Bayesian networks is presented and illustrated in this paper. We contribute to the IS security literature by supporting the exchange of security information among interconnected IS. Furthermore, for the modelling of IS security environment, an algorithm based on ant colony optimization facilitates to predict IS threat level more objectively. The model proposed in this paper has great potential for future extensions and refinements to provide more utility for the management of IS security.
The authors declare that there is no conflict of interests regarding the publication of this paper.
The research was supported by the National Natural Science Foundation of China (nos. 70901054 and 71271149) and the Program for New Century Excellent Talents in University (NCET). It was also supported by the China Postdoctoral Science Foundation funded Project (no. 2012M520025). The authors are very grateful to all anonymous reviewers whose invaluable comments and suggestions substantially helped improve the quality of this paper.