VEA-bility Analysis of Network Diversification

VEA-bility Analysis of Network Diversification Melanie Tupper Supervised by Nur Zincir-Heywood Faculty of Computer Science, Dalhousie University tupper@cs.dal.ca zincir@cs.dal.ca August 31, 2007 Abstract: In nature, genetic diversity increases a species' ability to survive when faced with biological, environmental, or predatory threats. In this work, we investigate whether computer networks may benefit from diversification in much the same way. To this end, we compare different network configurations to explore the ability of a computer network to defend itself against threats. In order to compare configurations, we also present a novel security metric, VEA-bility, which measures the desirability of a specific network configuration. Our metric is a numeric value that is computed by considering network vulnerabilities, network topology, intruder connectivity, and an attack graph of the network. The higher the value, the greater the ability of the network to defend against attacks.

Acknowledgements I would like to thank The Computer Research Association s Committee on the Status of Women in Computing Research (CRA-W) and The Natural Sciences and Engineering Research Council of Canada (NSERC) for supporting this research. I would like to thank my mentor, Dr Nur Zincir-Heywood for her inspiration and guidance that made this project an enjoyable and rewarding experience. I would like to thank the entire Tech Support team at Dalhousie University for their cooperation and assistance for the duration of this project they made this project possible. I would also like to thank my husband, Stewart Hardie, for his encouragement, love, and support. 1

Contents Section 1 - Introduction...5 1.1 Motivation... 5 1.2 Overview... 8 Section 2 Related Work... 9 2.1 Diversity... 9 2.2 Security Metrics... 12 Section 3 Attack Graphs... 13 Section 4 - Methodology... 16 4.1 Data Collection... 16 4.2 Feature Selection... 17 4.3 Virtual Network Modeling and Simulation... 18 4.4 VEA-bility analysis... 23 Section 5 - Results... 26 5.1 Nessus Scans Results... 26 5.2 Selected Features... 27 5.3 Virtual Network Diversification... 28 Section 6 Network Example... 35 Section 7 Conclusions and Future Work... 40 References... 42 Appendix A: Sample Attack Graphs... 44 Appendix B: Sample XML File... 45 Appendix C: Network Configurations... 47 Appendix D: Sample Nessus Scan Result... 61 Appendix E: Configuration Data... 66 2

List of Figures Figure 2.1 Monoculture Computer Networks...10 Figure 2.2 Diversified Computer Networks...10 Figure 2.3 Diversified Network Example...11 Figure 3.1 Sample Attack Graph...12 Figure 4.1 Base and Diversified Base Configurations...16 Figure 4.2 Initial and Diversified Firewalled Configurations...20 Figure 4.3 Initial and Diversified DMZ Configurations...20 Figure 4.4 Screenshot of Temporal Generation Interface...23 Figure 5.1 A Scan Result Sample...26 Figure 5.2 Average VEA-bility for Diversified Network Configurations...33 Figure 5.3 VEA-bility Distribution...34 Figure 6.1 Network Configuration 1...35 Figure 6.2 Network Configuration 35...36 Figure 6.3 Network Configuration 106...36 Figure 6.4 Network Configuration 140...37 Figure 6.5 Network Configuration 211...38 Figure 6.6 Network Configuration 223...38 Figure 6.7 Network Configuration 279...39 3

List of Tables Table 4.1 Virtual Network Configurations...22 Table 5.1 Nessus Scan Results by Operating System...26 Table 5.2 Virtual Network Vulnerabilities...27 Table 5.3 Average V N Dimensin s...29 Table 5.4 Average E N Dimensin s...30 Table 5.5 Average A N Dimensin s...31 Table 5.6 Average VEA-bility s...32 4

Section 1 Introduction 1.1 Motivation A primary objective for a computer network administrator is to maintain a stable, secure environment for network users. This objective includes ensuring that malicious computer users, known as attackers or intruders, are kept from compromising the network. An intruder exploits vulnerabilities in the software running on hosts within the network to gain access to network resources. The effects of such an exploit can range from low-level information access on one host to remote root access on multiple hosts, often including critical servers. Preventing intruder exploits includes detecting and patching, if possible, any known software vulnerabilities. Software vulnerabilities, most often the result of careless programming, are weaknesses in software that attackers can use to gain or escalate network privileges. A network service running software that is vulnerable is considered to be a vulnerable service. A common vulnerability, known as a buffer overflow or overrun, is the failure to check the size of a data buffer. Other vulnerabilities include format string vulnerabilities and logic errors. An exploit, or attack, is a way for the attacker to take advantage of vulnerabilities and can take the form of a piece of software, sequence of commands, or a block of data. If successful, the intruder will have gained privileges equal to that of the vulnerable program, allowing the intruder to access information or escalate privileges on the target host. Buffer overflow vulnerabilities can be exploited by an attacker by sending input that is too large for the intended buffer. This results in the extra data being overwritten to unauthorized memory locations, allowing the intruder to execute arbitrary code. Another common attack, the denial-ofservice (DoS) attack, occurs when an intruder sends more requests than the system can handle, thus rendering the service unable to process any valid requests. Although a DoS attack does not generally permit intruder access to the target host, this attack may serve as a distraction for a more serious attack on another host within the network. 5

There are several online software vulnerability databases, including the Common Vulnerabilities and Exposures (CVE) list [1]. This list is an initiative to standardize vulnerability references and gives vulnerabilities a name in the form CVE-YYYY-XXXX, where YYYY is the year in which the vulnerability was first reported. The CVE list, which is searchable by the CVE name, also provides the status of the vulnerability, a description, and a list of references. Network hardening refers to the various methods that can be employed to secure a system, which includes patching software vulnerabilities. Other ways in which a network can be made more secure is through the addition of firewalls, Demilitarized zones (DMZs), intrusion detection systems (IDSs), or intrusion prevention systems (IPSs). Firewalls and DMZs protect a network by restricting the connectivity of hosts both inside and outside a network. An IDS monitors network traffic and can be configured to alert the administrator to suspicious activity. An IPS not only monitors network activity, but can also react in real time to block or prevent malicious activity. A firewall, which can be either a hardware or software device, restricts the connectivity of hosts outside a network to hosts inside a network, or vice versa, and aims to filter unwanted behavior to or from the network. A DMZ, also known as a perimeter network, is an area between an internal and external network. A DMZ configuration allows hosts on both the internal and external networks to connect to hosts in the DMZ. However, hosts inside the DMZ are generally restricted from connecting to hosts on the internal network. A DMZ configuration serves to protect the internal network from attack should an intruder be able to gain elevated privileges on a host inside the DMZ. Web servers are commonly placed in the DMZ to allow external users to request information while protecting hosts inside the network. An IDS or IPS is a software application that monitors malicious traffic between hosts, including worms and viruses, which would go undetected by a firewall. The presence of this type of unwanted traffic will alert the system administrator to suspicious activity. Even with the best security practices, it is inevitable that systems become vulnerable. A common practice for detecting vulnerabilities on a network employs a vulnerability scanner: a software program that can be used to pinpoint weaknesses in a network. A vulnerability scanner generally 6

searches for active IP addresses, then attempts to connect to open ports and determine the operating system and applications running on the host. If a vulnerable service is found, most scanners will identify the related CVE number and assign a risk factor to the vulnerability. A popular, open-source vulnerability scanner, which is available for numerous operating systems, is the Nessus Vulnerability Scanner [2]. Once vulnerabilities have been identified, the administrator can attempt to fix the hole by finding the corresponding vulnerability in one of the online databases, then downloading and applying a software patch, if one is available. While these tools are useful for increasing security, they can only be used on existing physical networks. Since new software vulnerabilities are being discovered at an alarming rate of approximately 18 vulnerabilities per day [3], exploit prevention has become an attractive research area. Network diversification is an emerging trend that is suggested to increase the security of a computer network [4]. In much the same way as biodiversity increases a species likelihood of survival and portfolio diversity increases an investor s chances of receiving a return on investment, network diversification aims to eliminate monoculture configurations by introducing additional hosts, operating systems, and applications into the network. However, to the authors knowledge, there has been no previous research investigating this hypothesis. Thus, the purpose of this research project is to explore the effects of network diversity on network security. Our approach involves 3 phases: (i) A data collection phase; (ii) An experimental phase; and (iii) An evaluation phase. In the data collection phase, we use a vulnerability scanner to gather host information from the test bed. An important aspect of this research is the ability to model realistic network topologies with known vulnerabilities. For this reason, we use a set of hosts on the Faculty of Computer Science network at Dalhousie University as our test bed. In the experimental phase, we use the data collected to model diversified networks from initial configurations, including known vulnerabilities. To this end, we use a software package to generate an attack graph, which is a tree based data structure describing all known courses of actions that an attacker can take to compromise the network. 7

In the final phase, evaluation, we propose a novel quantitative metric, VEA-bility, which can be used to compare different network configurations. The underlying idea behind our VEA-bility metric is that the security of a network is influenced by many factors, including the severity of existing vulnerabilities, distribution of services, connectivity of hosts, and possible attack paths. These factors are modeled into three network dimensions: vulnerability, exploitability, and attackability. The VEA-bility score, a numeric value in the range [0,10], is a function of these three dimensions. Based on our findings, we conclude that the diversification theory, when properly applied, can be employed to increase the security of a network. These conclusions are important to network administrators as they strive to provide secure, yet functional, network configurations. 1.2 Overview The following section, Section 2, provides an overview of network diversification and security metrics. Section 3 gives a detailed description of attack graphs. Section 4 reviews each phase of our research and Section 5 summarizes our results. Section 6 provides an example of our VEAbility metric applied to a diversified network. Section 7 presents our conclusions and suggests further directions for this research. 8

Section 2 Related Work 2.1 Diversity While network diversification is only an emerging strategy, the benefits of diverse populations have been extensively researched in many disciplines, including biology and economics. Biodiversity, the variation of life, has three levels: genetic diversity, species diversity, and ecosystem diversity [5]. While all three are essential to survival, we use the genetic level to demonstrate how diversity can impact the survival of a species or population. Genetic diversity refers to the differences in the raw material of a species; DNA and RNA are two forms of this raw material. It is this raw material that makes it possible for a species to evolve and adapt, thus surviving biological, environmental, or predatory threats. A decline in the number of individuals in a population lowers the genetic variability, which is common in endangered species. One such species that is vulnerable to extinction due to the lack of genetic variation is the cheetah. The few cheetah populations that do remain are being forced to compete for resources in limited and diminishing habitats. The cheetah s limited genetic diversity may affect their ability to adapt to these environmental changes, further limiting their chances of survival. Diversification in finance involves choosing different investments to comprise a portfolio. Two general categories of financial diversification are horizontal and vertical. Horizontal diversification refers to diversity among the same type of investment, stocks, for example. Vertical diversification refers to investing in different types of investments such as stocks, bonds and mutual funds. The main benefit of portfolio diversification is reduced risk. This strategy relies on historical evidence that it is unlikely that different investments will simultaneously move in the same direction, thus reducing the risk. 9

Figure 2.1 depicts a monoculture computer network consisting of hosts running the same operating system and offering like services. A firewall is used solely to illustrate the idea of an internal network. An example of a monoculture network would be a computer lab where the operating system and services on each machine are similar, if not identical. If an attacker were able to compromise one host, it is likely that the attacker would be able to compromise every host on the network. Unfortunately, the monoculture configuration approach is common in many organizations due to limited resources, including time, money, and network administration personnel. Figure 2.1 Monoculture Computer Network On the other hand, Figure 2.2 demonstrates how a network can be diversified by dispersing the network services among hosts running different operating systems. Industry professionals propose that monoculture networks can benefit from diversity strategies in much the same way as species benefit from genetic diversity and investors benefit from diverse portfolios [4,6,7]. As the saying goes, Don t put all your eggs in one basket! Figure 2.2 Diversified Computer Network 10

Jajodia et al. [8] attempt to demonstrate that network diversity is not always beneficial to a network. Although we do not disagree with their conclusion, that is, we agree that it is possible to obtain a less secure network through diversification, we believe that in order to make such a claim, a thorough investigation is needed providing examples to show whether the result they obtained is typical or exceptional. Their example case, shown in Figure 2.3, compares two firewalled configurations, which we refer to as configuration 1 and configuration 2. The internal network in configuration 1 consists of a vulnerable mail server running on a Linux host, an ftp directory running on a Linux host, and a Database Management System (DBMS) running on a FreeBSD host. Configuration 2 includes an additional vulnerable mail server running on a Windows 2000 host. Figure 2.3 Diversified Network Example Jajodia et al. [8] state that a network configuration that has less attack paths is considered to be more secure. They argue that since configuration 2 has twice as many attack paths as configuration 1, configuration 2 is less secure which shows that diversification may be detrimental to a network. However, the main purpose of their work is the proposition of a security metric that does not consider the number of attack paths. While we agree that the number of attack paths to the target host should be considered, we believe this exemplifies the need to include a myriad of factors when determining the level of security offered by a network configuration. 11

2.2 Security Metrics Comparing the desirability of different network configurations requires a security metric. In general, a metric is a quantifiable measurement that allows for comparison. A security metric can be either qualitative or quantitative, and measures the degree of security controls, policies and procedures. For a security administrator, a security metric allows for comparison of different network configurations. When reviewing related work, our primary interest lies with quantitative security metrics that generate a numeric score. Jajodia et al. [8] present such a metric based on the strength of the weakest adversary that can compromise the network. The algorithm they present starts with a goal state and works in reverse, decomposing the requirements for the previous network state until an initial state is found. This produces the minimum set of initial attributes that an attacker would need to compromise a specified host and can be compared to other configurations. Manadhata et al. [9] use an attack surface metric to compare the attack surfaces of two ftp servers. Although this research does not offer a metric to compare the security of networks, the contributions of this paper inspire us to consider a security metric for a network with multiple dimensions. Their metric computes a discrete score for each of their three proposed dimensions: methods, channels, and data. The metric proposed by Adedin et al. [10] to evaluate network security policies generates one unified score that is a weighted aggregation of different factors. These factors include existing network vulnerabilities, vulnerability history of exposed services, exposure of services, and traffic volumes handled by services. In calculating the Existing Vulnerability Measure (EVM), this metric uses an exponential average to ensure that the resulting score will be at least as high as the highest vulnerability score present in the system. Additional vulnerabilities serve only to increase this score. We also use an aggregated, unified security score as well as exponential averages in our own VEA-bility metric for network security. Another tool we use in calculating the VEA-bility of a network is an attack graph. The next section details how an attack graph is generated, the information it represents, and how we use this information in our VEA-bility metric. 12

Section 3 Attack Graphs While identifying single vulnerabilities is useful, the security threat increases exponentially with multiple network vulnerabilities. One way to explore the effects of multiple vulnerabilities on a network is through an attack graph. An attack graph is a pictorial representation of the paths an attacker can take to exploit network vulnerabilities. The paths in the graph, known as attack paths, represent all the ways an intruder can penetrate the network. This can be used by a network administrator to identify how vulnerable the network is and what hardening measures should be taken to thwart attacks. Generation of an attack graph requires an attack model and a wealth of information regarding network topology and existing vulnerabilities. The network is usually modeled as state-machine where the nodes of the resulting attack graph represent a network state and the graph edges represent transitions, including actions taken by the attacker. Traditionally, attack graphs were produced manually by groups known as Red Teams who would generate hand drawn attack graphs on a large white board. Since attack graphs scale exponentially, this requires a substantial commitment of time and resources. Several software packages have been designed to automate the production of attack graphs, many of which are discussed in the following section. The main challenge of automating the generation of attack graphs is the exponential scaling of the graph with additional vulnerabilities. Most automated attack graph generators produce a pruned attack graph; the graph generated contains all the paths to a specific target host as opposed to the whole network. This type of graph can be aggregated for each host to comprise a network attack graph. A pruned attack graph is useful to an administrator wishing to protect a critical server. As the name suggests, a critical server offers essential network services; therefore, the security of a critical server is given a higher priority than other hosts on the network. When the critical server is identified as the target host, the pruned attack graph will identify attack paths to the critical server but not to other hosts on the network. 13

Network attack graphs are considered to be valuable tools for evaluating the security of a network. Therefore, much work has been done in the area of automated attack graph generation and analysis. Amman et al. [11] present an algorithm that scales well, and is implemented by Jajodia et al. [12] in their Topological Vulnerability Analysis (TVA) tool. The TVA tool automatically imports results from Nessus scans, but also requires that exploit and goal state information be entered by hand. TVA does not model firewall and router rules into the network model, but rather, relies on data collected from the Nessus scans to determine host connectivity. Since TVA requires Nessus scan info, it can only be used on an actual network and not on virtual networks or network simulations directly. Michael Artz s NetSPA tool [13] also requires information collected form Nessus scans, but must be entered into a database by hand. The NetSPA architecture relies on a software database and an action database to generate the network model. Software database entries can either be populated by hand or directly from an online vulnerability database. Action database entries must be entered manually and are used to model pre-conditions and post-conditions of intruder actions. The toolkit developed by Sheyner et al. [14] requires a user defined XML file describing the network for input. Therefore, this file can describe both virtual networks as well as existing physical networks, allowing analysts to consider alternate configurations before implementation. The input file specifies host information, connectivity among hosts, trust relations, a model of the intruder, intruder actions, and IDS information. This toolkit generates a pruned attack graph to a specified host. Although this toolkit does not produce a full attack graph, individual attack graphs to specific hosst can be combined if required. Our original research goal involved comparing the different automated attack graph generation software packages as described above on a test bed to determine which implementation had the most potential for further development. However, in spite of our efforts, we were only able to obtain a copy of one toolkit, which prompted the current research direction to use this toolkit to study the effects of diversity on security. 14

Our research uses a toolkit based on the work of Sheyner et al. This toolkit has been updated and is currently being maintained by David Swasey and is freely available for download [15]. We refer to this updated toolkit as the Sheyner/Swasey toolkit since the original toolkit designed by Sheyner et al. is still available, but not supported. Figure 3.1 is an example of a simple attack graph produced by the Sheyner/Swasey toolkit. The network configuration consists of one attacker and two internal hosts, A and B, each running one vulnerable service. For simplicity, we assume no restrictions on connectivity. Figure 3.1 Sample Attack Graph This is a pruned attack graph representing the attack paths to host A. Transitions from a parent node to a child are the result of a state change, represented as edges, and stored as Boolean values. Node 0 represents the initial state of the network with a simple transition to Node 1, most likely as the result of a network scan. Node 2 represents the attacker gaining access to host B, which is used to gain user access to host A (Node 3). The attacker escalates the privilege to root, resulting in Node 4. Alternatively, the attacker can directly gain user access to host A, Node 5, and escalate the user privilege to root, resulting in Node 6. Examples of other attack graphs can be found in Appendix A. 15

Section 4 Methodology 4.1 Data Collection Our goal to produce realistic virtual networks prompted us to seek permission to use the Dalhousie University network as our test bed. To accurately model networks mimicking the test bed, we use the Nessus Vulnerability Scanner to collect network topology information. The Nessus Scanner is an attractive research tool primarily because it allows safe checks. When the scanner is configured to scan a network using this option it will not attempt to exploit vulnerabilities, enabling us to compile vulnerability information without causing harmful Denial of Service (DoS) attacks. We run the scans from a Windows XP platform; however, the Nessus Scanner is also available for Mac OS X, Linux, FreeBSD, and Solaris. The Windows installation automatically installs and configures both server and client software, which may need to be configured separately for other operating systems. The Nessus Scanner gathers information by sending requests to all ports on hosts identified in the scan parameters. We use the scanner s default settings, but limit the number of hosts scanned for each scan to 20. The default range is 40 hosts per scan, as suggested in the Nessus 3.0 Client Guide [16]; however, we limit the number of hosts to 20 to further reduce the risk of overwhelming the system. The scan results are saved in XML format, but can also be viewed and resaved in plain text. In total, we scan 250 hosts, and generate results for 85 of these hosts. Failure to generate a report indicates that the Nessus Scanner was unable to connect to that host, and therefore, unable to extract any host information. The test bed is comprised of hosts located in diverse physical locations on the network, and includes network servers, faculty machines, and student machines. 16

4.2 Feature Selection The purpose of the Nessus scans is to model a realistic virtual network, which requires a wealth of host information. From the scan results we extract the following information: IP address Operating System Number of open ports Number of notes Number of warnings Number of holes Port number and corresponding services running on the open port CVE identification numbers and risk factors associated with vulnerabilities The magnitude of information requires that we construct a more compact representation of data from which to choose a set of operating systems to use in our experiments. We reorganize the host information into categories by operating system, further decomposing the operating systems by version or distribution. For each category we record: Number of hosts Number of warnings Number of holes Number of hosts with at least 1 vulnerability in each of the three highest risk categories After choosing three suitable operating systems we refer back to the original table to isolate similar vulnerabilities, also referring extensively to the National Vulnerability Database [17] to compare vulnerability characteristics including CVSS base score, access vector, access complexity, authentication, impact type, and age. We are concerned with these characteristics 17

because they express the severity of the vulnerability. By limiting the variation among these characteristics, we assure that our experiments measure the effects of different network configurations as opposed to different vulnerability characteristics. 4.3 Virtual Network Modeling and Simulation Before explaining how we diversify our virtual networks, it is important to understand what we mean by the term diversification. We define the diversification of a network as altering the topology of a network by redistributing services among additional hosts running either the same or another operating system as the original configuration. We model our diverse virtual networks into XML files to produce attack graphs using the Sheyner/Swasey toolkit. We chose this toolkit because of its availability, GUI interface, and ability to model virtual networks. Each XML file describes a network attack model comprised of the following components: Set of hosts connected to the network including operating system, IP address, vulnerabilities, open ports, and corresponding services; Connectivity information describing network topology and host reachability; Trust relations among hosts; Intruder information including connectivity information and level of privilege the intruder has on each of the other hosts; Set of exploits including CVE of exploited vulnerability, name, description, local and global exploit preconditions, and local and global exploit effects; A model of the intrusion detection system. Following the model provided by Sheyner et al. [14], we do not model trust relations among hosts within the network. We use the Sheyner/Swasey toolkit to generate attack graphs for each 18

configuration. The toolkit produces a pruned attack graph to a specified target host; therefore, a separate file must be generated for each host inside the network. Relevant data recorded includes number of nodes, number of edges, number of attack paths, and number of ways to traverse the network. A sample XML file can be found in Appendix B. We begin our modeling with the following three topologies: 1. One intruder and one network host running four services on Solaris Apache vulnerability 2. One intruder and one network host running four services on Windows rpc vulnerability 3. One intruder and one network host running four services on Linux ftp vulnerability These three topologies are considered to be the base configurations. We purposefully start with configurations that result in a low VEA-bility score, and then test the effects of diversification by adding additional hosts and operating systems; we refer to these diversified configurations as the diversified base configurations. Figure 4.1 depicts the Linux base configuration and a diversified configuration with 2 hosts running different operating systems. Figure 4.1 Base and Diversified Base Configurations We then add a firewall or DMZ to the base configurations to produce three initial firewalled configurations and three initial DMZ configurations, respectively. These new configurations act as a control and are also diversified by adding hosts and additional operating systems to produce diversified firewalled configurations and diversified DMZ configurations. The DMZ 19

configurations are constructed by isolating the host running a web server in the DMZ. This results in 12 fewer configurations since some configurations are duplicated. Figure 4.2 shows an initial firewall configuration and a diversified firewall configuration with 2 operating systems. Figure 7 depicts an initial DMZ configuration and a diversified DMZ configuration with 3 operating systems. Figure 4.2 Initial and Diversified Firewalled Configurations Figure 4.3 Initial and Diversified DMZ Configurations 20

In total, we test 303 configurations: 3 base configurations 24 diversified base configuration with 1 operating system 42 diversified base configurations with 2 operating systems 36 diversified base configurations with 3 operating systems 3 initial firewalled configurations 24 diversified firewalled configurations with 1 operating system 42 diversified firewalled configurations with 2 operating systems 36 diversified firewalled configurations with 3 operating systems 3 initial DMZ configurations 12 diversified DMZ configurations with 1 operating system 42 diversified DMZ configurations with 2 operating systems 36 diversified DMZ configurations with 3 operating systems Firewalled and DMZ configurations act initially as a control to show the level of security that can be achieved without diversification. Firewalls and DMZs are tools that administrators often use to increase the security of a network; therefore, we diversify the initial firewalled and DMZ configurations to show how the diversification approach in combination with existing security strategies can further affect security. A detailed verbal description of the test configurations can be found in Appendix C. 21

Table 4.1 matches the configurations in each category with the descriptions in Appendix C. Table 4.1 Virtual Network Configurations Configuration group Number of Configurations in group Corresponding Configurations in Appendix C Base configurations 3 1, 10, 19 Diversified base configurations with 1 operating system Diversified base configurations with 2 operating systems Diversified base configurations with 3 operating systems 24 2-9, 11-18, 20-27 42 28 69 36 70 105 Initial firewalled configurations 3 106, 115, 124 Diversified firewalled configurations with 1 operating system Diversified firewalled configurations with 2 operating systems Diversified firewalled configurations with 3 operating systems 24 107 114, 116 123, 125-132 42 133 174 36 175 210 Initial DMZ configurations 3 211, 216, 221 Diversified DMZ configurations with 1 operating system Diversified DMZ configurations with 2 operating systems Diversified DMZ configurations with 3 operating syatems 12 212 215, 217 220, 222-225 42 226 267 36 268-303 22

4.4 VEA-bility analysis Given that the objective of this work is to explore the effects of diversification on network security, a metric is required to compare the different aforementioned network configurations. Thus, a security metric is defined to capture the numerous factors that influence the security of a network. To this end, we propose the VEA-bility metric to be a function of the security scores along three dimensions: vulnerability, exploitability, and attackability. For the sake of simplicity, the vulnerability, exploitability, and attackability scores will be represented in equations as V, E, and A, respectively. Each of the three dimension scores is a numeric value in the range [0,10]. The VEA-bility metric uses data from three sources: network topology, attack graphs, and scores as assigned by the Common Vulnerability Scoring System (CVSS) [18]. VEA-bility uses the following CVSS values: impact score, temporal score, and exploitability score. The impact score measures the impact that a successful exploit will have on the availability, integrity, and accessibility of information resources. The temporal score assigns a value based on the age of the vulnerability, the remediation status of a patch, and the credibility of the patch source. The vulnerability of a network is the degree to which an exploit can impact a system; a measure that is influence by time. Therefore, our vulnerability dimension is a function of the impact and temporal scores. Figure 4.4 shows a screenshot of the online temporal score interface. Our exploitability dimension is a function of the exploitability score, which evaluates the likelihood of exploitation. Temporal Metrics These metrics describe elements about the vulnerability that change over time. If all of these values are left as 'Undefined', the environmental score will be based on the base score. Availability of exploit (Exploitability) Functional exploit exists Type of fix available (RemediationLevel) Official fix Level of verification that vulnerability exists (ReportConfidence) Confirmed Figure 4.4 Screenshot of Temporal Generation Interface 23

Since a network is only as secure as its host, we define the three dimensions for a network configuration as a function of the three dimensions for each host on the network. The vulnerability score of a network is the exponential average of the vulnerability scores of each host on the network, or a maximum of 10. This captures the requirement that the vulnerability score of the network is at least as large as the largest host vulnerability score; additional vulnerable hosts serve only to increase this value, which can be a maximum of 10. The exploitability and attackability scores of a network are the summation of the exploitability and attackability scores of each host, respectively. Therefore, the following equations represent the Vulnerability, Exploitability, and Attackability dimension scores for a network: For a network, N, let HV(N) be the set of hosts in N that contain vulnerabilities. A network void of vulnerabilities scores a 0 along each dimension; otherwise, we define the network dimensions: Vhost V N = min(10, ln e ) host HV ( N ) E N = host HV ( N ) E host A N = host HV ( N ) A host We propose that a host with multiple vulnerabilities is less secure than a host with a single vulnerability, which is modeled into the vulnerability and attackability dimensions by taking the exponential average of the values for all vulnerabilities. Again, this allows the value to be at least as large as the highest value, and additional scores serve to increase this value to a maximum value of 10. Let each vulnerability, v, have an impact score, temporal score, and exploitability score as defined by the CVSS [18]. An impact and exploitability sub-scores are automatically generated for each CVE name, the temporal score requires user input. 24

We define the severity, S, of a vulnerability to be the average of the impact and temporal scores: S v = (Impact v + Temporal v ) / 2 The host vulnerability score is an exponential average of the severity scores of the vulnerabilities on a host, or 10, whichever is lower. The exploitability is the exponential average of the exploitability score for all host vulnerabilities multiplied by the ratio of network services on the host. Attackability is the ratio of attack paths produced by attack graphs to total number of attack paths, and is multiplied by a factor of 10 to produce a number in the range [0,10], ensuring that all dimensions have the same range. For a host, host, let V(host) be the set of vulnerabilities on a host. We then define the three host dimensions as: V host = min(10, ln e Sv ) v V ( host ) Exploitability v E host = (min(10, ln e )) ( services on host) / ( network services) v V ( host ) A host = (10) (attack paths) / ( network paths) The final equation for network VEA-bility is: VEA-bility N = 10 ((V+E+A) N / 3) According to the NIST Security Metrics Guide for Information Technology Systems [19], a metric must yield quantifiable information, be useful for tracking system performance, measure a repeatable process, and supporting data must be readily obtainable. The metric we propose, VEA-bility, conforms to these standards in that it is quantifiable, that is, it is expressed as a numeric value in the range [0,10]. By using the Nessus scanner and the Sheyner/Swasey toolkit, both freely available online, the methods we employ can be easily duplicated. Also, this metric serves to track the performance of a network configuration by comparing the score to other possible configurations. Thus, a network administrator can use the VEA-bility metric to direct or reallocate resources. 25

Section 5 Results 5.1 Nessus Scans Results Scan results can be viewed in HTML, XML or plaintext format. Figure 5.1 is an excerpt from one of the resulting plaintext output files; the complete result can be found in Appendix D. An example is provided to show the structure of the report and type of information provided. All IP address references and other information that could potentially identify this host have been removed. surf (1010/tcp) 11219 NOTE Port is open; surf (1010/tcp) 11111 NOTE RPC program 100024 version 1 'status' is running on this port;; vnc (5900/tcp) 19288 NOTE The remote VNC server supports those security types:;+ 30;; netbios-ssn (139/tcp) 11011 NOTE An SMB server is running on this port; Figure 5.1. A Scan Result Sample Due to the nature of the information, we are unable to publish our original table of results. However, Table 5.1 presents a summary of the data we obtained to select three suitable operating systems. Table 5.1. Nessus Scan Results by Operating System Operating System Hosts Warnings Holes hosts with 1+ medium CVE hosts with 1+ high CVE hosts with 1+ critical CVE Mac OS X 16 8 10 0 0 7 Windows 10 11 12 3 0 3 Linux 23 26 20 4 3 3 Solaris 7 45 37 4 4 4 FreeBSD 2 0 0 0 0 0 26

5.2 Selected Features From Table 5.1, we choose to model our networks using Linux, Solaris, and Windows operating systems. These were chosen because we were able to identify similar vulnerabilities on each operating system on our test bed. Table 5.2 summarizes the identified vulnerabilities. Table 5.2 Virtual Network Vulnerabilities Operating System Vulnerable Service CVE Name Description Linux File Transfer Protocol (ftp) CVE-2004-0185 Buffer overflow Solaris Apache (web) CVE-2004-0492 Heap-based buffer overflow Windows Remote Procedure Call (rpc) CVE-2004-0212 Stack-based buffer overflow All three vulnerabilities have a CVSS base score of 10. They are network exploitable and low complexity. They do not require authentication, but do provide administrator access. Finally, all were first reported in 2004. We believe that by keeping the exploits similar, the resulting VEAbility scores will reflect the effects of changes in the network configurations and will not be influenced by different vulnerability factors. In all experiment scenarios, a tested virtual network offers the following services: rpc on port 135, web server on port 80, ftp server on port 21 and a database. Port numbers are assigned to network exposed services as required to model the XML files. The database service is for illustration only and represents any critical service running on the network; it is used to show how a critical service could be exploited without being directly exposed to the network. 27

5.3 Virtual Network Diversification Detailed results for each network configuration tested can be found in Appendix E. These results correspond to the configurations in Appendix C. For each configuration, we have recorded the impact score, temporal score, exploitability score, number of attack paths, number of ways to traverse the network, network dimension scores, and the VEA-bility score of the network. As discussed earlier, the vulnerabilities were intentionally chosen to be similar as not to skew the results. As a consequence, the severity, S, of the three vulnerabilities is constant. A temporal score of 8.7 was generated by using the online metric with availability of exploit field set to widespread, the type of fix available field set to Official Fix, and the Level of verification that vulnerability exists field set to Confirmed. The severity score for the three vulnerabilities is: S v = (Impact v + Temporal v ) / 2 = (10 + 8.7) / 2 = 9.35 We are only dealing with one vulnerability per host; therefore, V host = S v = 9.35 if a vulnerability exists, and 0 otherwise. Our network configurations may have either 0, 1, 2, or three hosts, the respective V N values will be 0, 9.35, 10, and 10. This demonstrates the value of using an exponential average. A network with one vulnerability has a vulnerability dimension value of 9.35 since this is the only vulnerability. However, when there are more vulnerabilities, the risk to the network is greater than 9.35 depending on the severity of the additional vulnerabilities. A simple average of severity scores would yield a 9.35 rating for a configuration with 2 vulnerabilities, which would not accurately represent the increased risk. 28

Due to the large dataset, comparing individual results is not practical. Therefore, in Table 5.3, we present the average vulnerability dimension scores for comparison. A higher score indicates a more vulnerable network configuration. Table 5.3 Average V N Dimension s Configuration Average V N Dimension Base configurations 9.35 Initial firewalled configurations 9.35 Initial DMZ configurations 9.35 Diversified base configurations with 1 O/S 9.35 Diversified base configurations with 2 O/S 6.86 Diversified base configurations with 3 O/S 6.40 Diversified firewalled configurations with 1 O/S 9.35 Diversified firewalled configurations with 2 O/S 6.86 Diversified firewalled configurations with 3 O/S 6.40 Diversified DMZ configurations with 1 O/S 9.35 Diversified DMZ configurations with 2 O/S 6.86 Diversified DMZ configurations with 3 O/S 6.40 These results indicate that with diversification by adding different operating systems it is possible to reduce the vulnerability of a network. This occurs since running a vulnerable service on a different operating system removes the vulnerability, provided the software is not vulnerable on the alternate operating system. It should be noted here that, based on our experiments, the addition of firewalls and DMZs seems to affect connectivity only and has no effect on the network vulnerability dimension. 29

Limiting vulnerabilities to one per host results in the following calculation: E host = (Exploitability v ) ( services on host) / ( network services) The E N score is then the summation of E host values. Table 5.4 shows the average exploitability dimension scores, E N, for each configuration category. A higher score indicates a more exploitable network configuration. Table 5.4 Average E N Dimension s Configuration Average E N Dimension Base configurations 10 Initial firewalled configurations 10 Initial DMZ configurations 5.8 Diversified base configurations with 1 O/S 5.3 Diversified base configurations with 2 O/S 5.7 Diversified base configurations with 3 O/S 3.8 Diversified firewalled configurations with 1 O/S 5.3 Diversified firewalled configurations with 2 O/S 5.7 Diversified firewalled configurations with 3 O/S 3.8 Diversified DMZ configurations with 1 O/S 3.3 Diversified DMZ configurations with 2 O/S 3.9 Diversified DMZ configurations with 3 O/S 3.1 These results show that exploitability is best controlled by the addition of a DMZ, but can be improved through diversification. Since the exploitability dimension is related to the number of services on hosts with vulnerabilities, it seems to reason that isolating the web server on a host will best affect this dimension. 30

Attackability is a function of the ratio of attack paths to total paths through the network. This ratio is multiplied by 10 to generate a number compatible with the other two dimensions. Average attackability, A N, dimension scores for each configuration category are presented in Table 5.5. A higher score indicates a more attackable network configuration. Table 5.5 Average A N Dimension s Configuration Average A N Dimension Base configurations 10 Initial firewalled configurations 6.7 Initial DMZ configurations 1.7 Diversified base configurations with 1 O/S 2.2 Diversified base configurations with 2 O/S 3.9 Diversified base configurations with 3 O/S 1.6 Diversified firewalled configurations with 1 O/S 1.5 Diversified firewalled configurations with 2 O/S 2.6 Diversified firewalled configurations with 3 O/S 1.0 Diversified DMZ configurations with 1 O/S.6 Diversified DMZ configurations with 2 O/S 1.3 Diversified DMZ configurations with 3 O/S 1.1 These figures highlight both the benefits of diversity and a DMZ on the network security as represented in the attack graphs. While many system administrators use attack graphs alone to evaluate the security of a network [14], we propose that they are more useful when aggregated with other network factors. 31

For example, consider a vulnerable host on a network that cannot be exploited due to connectivity restrictions. We consider this network less secure than a network with no software vulnerabilities, but more secure than a network with no connectivity restrictions. This is reflected in our VEA-bility metric score. The overall average VEA-bility scores are found below in Table 5.6. A higher score indicates a more secure configuration, which we call more VEA-ble. Table 5.6 Average VEA-bility s Configuration Average VEA-bility Base configurations.2 Initial firewalled configurations.6 Initial DMZ configurations 4.4 Diversified base configurations with 1 O/S 4.4 Diversified base configurations with 2 O/S 4.5 Diversified base configurations with 3 O/S 6.1 Diversified firewalled configurations with 1 O/S 4.6 Diversified firewalled configurations with 2 O/S 5.0 Diversified firewalled configurations with 3 O/S 6.3 Diversified DMZ configurations with 1 O/S 5.6 Diversified DMZ configurations with 2 O/S 5.8 Diversified DMZ configurations with 3 O/S 6.5 Although these scores are averages, it is evident that it is possible to increase the security rating of a network configuration through diversification. Figure 5.1 is a pictorial representation of this data which highlights the benefits of diversification alone or in combination with a firewall or DMZ. 32

The significant advantage of network diversity is apparent when comparing the average scores for the diversified base configurations with three operating systems to the diversified DMZ configurations with three operating systems. As indicated, the DMZ adds only.4 to the final VEA-bility score. Average VEA-bility of Diverse Network Configurations No Firewall Firewall DMZ 6.46 5.95 5.59 6.27 6.1 4.62 4.95 4.38 4.5 4.37 1.33 0.22 Initial Configurations Same O/S Two O/S Three O/S Figure 5.2 Average VEA-bility for Diversified Network Configurations 33

The following table shows the distribution of diversified network VEA-bility scores. Although it is possible to decrease security, the tables show the distribution of configurations which resulted in lower and higher VEA-bility scores than the average initial scores of.22,.6, and 4.4. To reduce influencing the results by including firewalled and DMZ configurations, we divide the results into three categories: base, firewalled, and DMZ. VEA-bility Distribution Less VEA-ble More VEA-ble Percentage of Hosts 100 80 60 40 20 0 85.3 96.1 81.1 14.7 3.9 18.9 Base Firewalled DMZ Figure 5.3 VEA-bility Distribution 34

Section 6 Network Example The following is an example of how a network configuration can be diversified to increase security. We start with a simple configuration of an attacker and one host running four services and explore the effects of adding a firewall, DMZ, additional hosts, and additional operating systems. All of the following configurations can be found in Appendix B; the first is configuration 1. Figure 6.1 Network Configuration 1 Since there is only one host, network dimensions are host dimensions: V N = (10+8.7) / 2 = 9.35 E N = 10 (4/4) = 10 A N = (1/1)10 = 10 VEA-bility = 10- ((9.35+10+10)/3) =.22 35

On the other hand, configuration 35 results from isolating the database on the Solaris host and dispersing the services onto an additional host running Windows operating system. Figure 6.2 Network Configuration 35 V N = (10+8.7) / 2 = 9.35 E N = 10 (3/4) = 7.5 A N = (1/4)10 = 2.5 VEA-bility = 10- ((9.35+7.5+2.5)/3) = 3.6 Alternatively, adding a firewall to configuration 1 creates configuration 106: Figure 6.3 Network Configuration 106 V N = (10+8.7) / 2 = 9.35 E N = 10 (4/4) = 10 A N = (1/1)10 = 10 VEA-bility = 10- ((9.35+10+10)/3) =.22 36

The VEA-bility scores for configurations 1 and 106 are both.22 as a result of the firewall rules restricting connectivity to the rpc service, but not to the web of ftp services. Configuration 140 is the result of isolating the database on the Solaris host and dispersing the remaining services onto an additional host running Windows operating system. Figure 6.4 Network Configuration 140 V N = (10+8.7) / 2 = 9.35 E N = 10 (3/4) = 7.5 A N = (0/4)10 = 0 VEA-bility = 10- ((9.35+7.5)/3) = 4.4 There are no attack paths to hosts inside the network because the firewall restricts connectivity to the vulnerable rpc service. However, the fact that rpc service is vulnerable on Windows and 75% of the network services are on this host gives a resulting VEA-bility score of 4.4. This vulnerability could be exploited if the attacker discovers another way to get inside the network. 37

We generate configuration 211 from configuration 1 by adding another Solaris host to a DMZ to act as the web server. Figure 6.5 Network Configuration 211 V N = (10+8.7) / 2 = 9.35 E N = 10 (1/5) = 2.5 A N = (1/2)10 = 5.0 VEA-bility = 10- ((9.35+2.5+5.0)/3) = 4.4 This allows the attacker to gain root access on the DMZ host, but lack vulnerabilities inside the network to prevent the attacker from advancing. The VEA-bility represents the potential for a vulnerable internal attack if the attacker were to discover an alternate vulnerability. Furthermore, by isolating the database on the Solaris host inside the network, and adding a Windows host for the remaining services, we have configuration 233. Figure 6.6 Network Configuration 233 38

V N = (10+8.7) / 2 = 9.35 E N = 10 (2/4) = 5.0 A N = (0/5)10 = 0 VEA-bility = 10- ((9.35+5.0)/3) = 5.2 Again, there are no attack paths inside the network due to connectivity restrictions, however vulnerabilities still exist. This is the reason for the VEA-bility of 5.2 and reiterates our sentiment that the security of the network cannot be determined from attack graphs alone. Finally, by adding yet another host, Linux, we separate the services running on the Windows machine to get the following: configuration 279. Figure 6.7 Network Configuration 279 V N = 0 E N = 0 A N = 0 VEA-bility = 10- (0/3) = 10 This network is completely secure (very VEA-ble) as there are no vulnerabilities to exploit. 39

Section 7 Conclusions and Future Work In this work, the objective was to test and evaluate the effects of network diversity on network security. To this end, we gathered network topology information of our test bed using the Nessus Vulnerability Scanner. Using the collected data, we isolated the desired features and modeled diversified network configuration based on these features. Initial configurations were diversified by adding additional hosts and operating systems. We also added firewalls and DMZs as a control. Using these virtual models we generated attack graphs using the toolkit designed by Oleg Sheyner and updated by David Swasey. To compare the security measure of each resulting configuration, we proposed a quantitative security metric, VEA-bility. Our VEA-bility metric assigns a numeric value in the range [0,10] to each network configuration where 0 indicates a poorly configured network and 10 indicates the most secure network configuration possible. Using our VEA-bility metric applied to our diversified configurations, we find that diversification does indeed support the security of a network. Specifically, in our experiments, diversity assists in increasing the average score in each dimension, thus increasing the overall average VEA-bility of network configurations in each category. Therefore, observe that diversifying a computer network can result in a more secure configuration. We show that by diversifying services onto hosts where the resulting configuration is less vulnerable, the result is a more secure network. However, just as investors research and monitor their investments to maximize their profits, a network administrator should research the history of vulnerabilities for software intended for use on a system or network. Once configured, the system should be monitored by referring frequently to online vulnerability databases to patch or replace vulnerable software as required. There are a number of directions for extending this research in the future including improving the metric and using the VEA-bility metric to investigate specific aspects of network security. 40

Moreover, improving the network model is another future direction. In this research we did not have access to information provided by an intrusion detection system or intrusion prevention system on the test bed. Obtaining this information and including it in the network model would allow a network administrator to make more confident decisions concerning secure network topologies. Also, continuing from the work of Sheyner et al. [14], we did not model trust relations within the network, but rather modeled the resulting authentications as connectivity relations. Since the Sheyner/Swasey toolkit is designed to recognize host trust relations, representing these relations can provide more accurate representations of a network. One way to improve the VEA-bility metric includes adding more information provided by the CVSS [18]. One such example is the environmental score, which assigns a numeric value based on software implementation and network environment. The environmental score is calculated based on user defined input such as the potential for damage and the target distribution. Our VEA-bility metric could be used to investigate the security of a network to determine which of our three defined dimensions has a greater impact on the overall security. The results of this type of study would better allow network administrators to focus their efforts on events that would have the most impact on the security of their networks. 41

References [1] Common Vulnerabilities and Exposures (CVE). http://cve.mitre.org. [2] Nessus Vulnerability Scanner. http://www.nessus.org. [3] National Vulnerability Database Home. http://nvd.nist.gov. [4] Geer, D. E. (2007, April). The Evolution of Security [Electronic Version]. ACM Queue, 5(3). http://acmqueue.com/modules.php?name=content&pa=showpage&pid=478&page=4 [5] Campbell, N., Reece, J., Taylor, M., & Simon, E. (2006). Biology: Concepts and Connections. 5 th Edition. San Francisco: Benjamin Cummings. [6] Geer, D., Bace,R., Gutmann, P., Metzger, P., Pfleeger, C. P., Quarterman, J.S., and Schneier, B. (2003, Sept 27). CyberInsecurity: The Cost of Monopoly How the Dominance of Microsoft's Products Poses a Risk to Security. http://cryptome.org/cyberinsecurity.htm [7] Lemos, R. (2004, January). Agriculture epidemics may hold clues to Net viruses. CNET News.com. http://news.com.com/seeds+of+destruction/2009-7349_3-5140971.html [8] Jajodia, S., Pamula, J., Ammann, P., & Swarup, V. (2006). A Weakest Adversary Security Metric for Network Configuration Security Analysis. In QoP-2006: Quality of Protection workshop, October 2006. [9] Manadhata, P., Wing, J., Flynn, M., & McQueen, M. (2006). Measuring the Attack Surfaces of Two FTP Daemons. In QoP-2006: Quality of Protection workshop, October 2006. [10] Abedin, M., Nessa, S., Al-Shaer, E., & Khan, L. (2006). Vulnerability Analysis For Evaluating Quality of Protection Security Policies. In QoP-2006: Quality of Protection workshop, October 2006. 42

[11] Ammann, P., Wijesekera, D., & Kaushik, S. (2002). Scalable, Graph-Based Network Vulnerability Analysis. In Proceedings of the 9 th ACM Conference on Computer and Communications Security (CCS 02), November 2002. [12] Jajodia, S., Noel, S., & O Berry, B. (2005). Topological Analysis of Network Attack Vulnerability. In Managing Cyber Threats: Issues, Approaches and Challenges, pages 248-266. V. Kumar, J. Srivastava, and A. Lazarevic (Eds.), Springer-Verlag. [13] Artz, M. (2002). NetSPA, A Network Security Planning Architecture, M.S. Thesis. Cambridge: Massachusetts Institute of Technology, May 2002. [14] Sheyner, O. & Wing, J. M. (2004) Tools for Generating and Analyzing Attack Graphs. In Proceedings of Workshop on Formal Methods for Components and Objects, pp. 344-371. [15] Scenario and Attack Graphs. http://www.cs.cmu.edu/~scenariograph. [16] Nessus Client Guide. http://www.nessus.org/documentation/nessus_3.0_client_guide.pdf. [17] National Vulnerability Database Search Page. http://nvd.nist.gov/nvd.cfm. [18] A Complete Guide to the Common Vulnerability Scoring System (CVSS). http://www.first.org/cvss/v1/guide.html. [19] Security Metrics Guide for Information Technology Systems. http://csrc.nist.gov/publications/nistpubs/800-55/sp800-55.pdf. 43

Appendix A: Sample Attack Graphs Figure A.1 Single Path Attack Graph Figure A.2 Multiple Path Attack Graph Figure A.3 Multiple Path Attack Graph 44

Appendix B: Sample XML File <network> <host id="solaris" ip="192.168.1.1" network="inside"> <services> <web port="80" /> <ftp port="21" /> <rpc port="135" /> <database /> </services> <vulnerabilities> <CVE_2004_0492 /> </vulnerabilities> <connectivity> <remote id="solaris"> <web /> <ftp /> <rpc /> </remote> <remote id="intruder"> <web /> <ftp /> <rpc /> </remote> </connectivity> </host> <host id="intruder" ip="10.0.0.1" network="outside"> <connectivity> <remote id="solaris"> <web /> </remote> <remote id="intruder"> <web /> 45

<ftp /> <rpc /> </remote> </connectivity> </host> <adversary> <privileges> <privilege host="intruder" level="root" /> <privilege host="solaris" level="none" /> </privileges> <knowledge> <scan value="no" /> </knowledge> </adversary> <attack name="apache_buffer_overflow" description="gives a root shell on the target machine."> <local_preconditions> <privilege host="source" rel="gte" value="user" /> <privilege host="target" rel="lte" value="user" /> </local_preconditions> <global_preconditions> <service name="web" host="target" /> <vulnerability name="cve_2004_0492" host="target" /> <connectivity from="source" service="web" /> </global_preconditions> <local_effects> <privilege host="target" value="root" /> </local_effects> <global_effects> <service host="target" name="web" value="false" /> </global_effects> </attack> </network> 46

Appendix C: Network Configurations Base Configurations Intruder can connect to ports 80, 135, and 21. Configuration 1-1 intruder and 1 internal Solaris host running web, ftp, rpc, and database. Configuration 2 1 intruder, 1 internal Solaris host running rpc, and 1 internal Solaris host running web, ftp, and database. Configuration 3 1 intruder, 1 internal Solaris host running ftp, and 1 internal Solaris host running web, rpc, and database. Configuration 4 1 intruder, 1 internal Solaris host running web, and 1 internal Solaris host running rpc, ftp, and database. Configuration 5 1 intruder, 1 internal Solaris host running database, and 1 internal Solaris host running web, rpc, and ftp. Configuration 6 1 intruder, 1 internal Solaris host running database and web, and 1 internal Solaris host running ftp and rpc. Configuration 7 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Solaris host running ftp and web. Configuration 8 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Solaris host running rpc and web. Configuration 9 1 intruder, 1 internal Solaris host running database, 1 internal Solaris host running rpc, 1 internal Solaris host running ftp, and 1 internal Solaris host running web. Configuration 10 1 intruder and 1 internal Linux host running web, ftp, rpc, and database. Configuration 11 1 intruder, 1 internal Linux host running rpc, and 1 internal Linux host running web, database, and ftp. Configuration 12 1 intruder, 1 internal Linux host running ftp, and 1 internal Linux host running rpc, web, and database. Configuration 13 1 intruder, 1 internal Linux host running web, and 1 internal Linux host running rpc, ftp, and database. Configuration 14 1 intruder, 1 internal Linux host running database, and 1 internal Linux host running web, rpc, and ftp. Configuration 15 1 intruder, 1 internal Linux host running web and database, and 1 internal Linux host running ftp and rpc. Configuration 16 1 intruder, 1 internal Linux host running rpc and database, and 1 internal Linux host running ftp and web. Configuration 17 1 intruder, 1 internal Linux host running ftp and database, and 1 internal Linux host running web and rpc. Configuration 18 1 intruder, 1 internal Linux host running database, 1 internal Linux host running rpc, 1 internal Linux host running ftp, and 1 internal Linux host running web. Configuration 19 1 intruder and 1 internal Windows host running web, rpc, ftp, and database. Configuration 20 1 intruder, 1 internal Windows host running rpc, and 1 internal Windows host running ftp, web, and database. Configuration 21 1 intruder, 1 internal Windows host running ftp, and 1 internal Windows host running rpc, web, and database. Configuration 22 1 intruder, 1 internal Windows host running web, and 1 internal Windows host running ftp, rpc, and database. Configuration 23 1intruder, 1 internal Windows host running database, and 1 internal Windows host running rpc, ftp, and web. Configuration 24 1 intruder, 1 internal Windows host running web and database, and 1 internal Windows host running ftp and rpc. Configuration 25 1 intruder, 1 internal Windows host running rpc and database, and 1 internal Windows host running ftp and web. Configuration 26 1 intruder, 1 internal Windows host running ftp and database, and 1 internal Windows host running rpc and web. Configuration 27 1 intruder, 1 internal Windows host running database, 1 internal Windows host running rpc, 1 internal Windows host running ftp, and 1 internal Windows host running web. Configuration 28 1 intruder, 1 internal Solaris host running web, rpc, and database, and 1 internal Windows host running ftp. 47

Configuration 29 1 intruder, 1 internal Solaris host running database, ftp, and web, and 1 internal Windows host running rpc. Configuration 30 1 intruder, 1 internal Solaris host running database, ftp, and rpc, and 1 internal Windows host running web. Configuration 31 1 intruder, 1 internal Solaris host running web, ftp, and rpc, and 1 internal Windows host running database. Configuration 32 1 intruder, 1 internal Windows host running database, web, and rpc, and 1 internal Solaris host running ftp. Configuration 33 1 intruder, 1 internal Windows host running database, ftp, and web, and 1 internal Solaris host running rpc. Configuration 34 1 intruder, 1 internal Windows host running database, ftp, and rpc, and 1 internal Solaris host running web. Configuration 35 1 intruder, 1 internal Windows host running web, ftp, and rpc, and 1 internal Solaris host running database. Configuration 36 1 intruder, 1 internal Windows host running database and ftp, and 1 internal Solaris host running web and rpc. Configuration 37 1 intruder, 1 internal Windows host running database and web, and 1 internal Solaris host running ftp and rpc. Configuration 38 1 intruder, 1 internal Windows host running database and rpc, and 1 internal Solaris host running ftp and web. Configuration 39 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Windows host running ftp and web. Configuration 40 1 intruder, 1 internal Solaris host running database and web, and 1 internal Windows host running ftp and rpc. Configuration 41 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Windows host running web and rpc. Configuration 42 1 intruder, 1 internal Solaris host running database, rpc, and ftp, and 1 internal Linux host running web. Configuration 43 1 intruder, 1 internal Solaris host running database, ftp, and web, and 1 internal Linux host running rpc. Configuration 44 1 intruder, 1 internal Solaris host running database, rpc, and web, and 1interanal Linux host running ftp. Configuration 45 1 intruder, 1 internal Solaris host running ftp, web, and rpc, and 1 internal Linux host running database. Configuration 46 1 intruder, 1 internal Solaris host running ftp, and 1 internal Linux host running database, rpc, and web. Configuration 47 1 intruder, 1 internal Solaris host running rpc, and 1 internal Linux host running ftp, web, and database. Configuration 48 1 intruder, 1 internal Solaris host running web, and 1 internal Linux host running ftp, rpc, and database. Configuration 49 1 intruder, 1 internal Solaris host running database, and 1 internal Linux host running ftp, web, and rpc. Configuration 50 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Linux host running ftp and web. Configuration 51 1 intruder, 1 internal Solaris host running database and web, and 1 internal Linux host running ftp and rpc. Configuration 52 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Linux host running web and rpc. Configuration 53 1 intruder, 1 internal Solaris host running rpc and ftp, and 1 internal Linux host running database and web. Configuration 54 1 intruder, 1 internal Solaris host running web and rpc, and 1 internal Linux host running database and ftp. Configuration 55 1 intruder, 1 internal Solaris host running web and ftp, and 1 internal Linux host running rpc and database. Configuration 56 1 intruder, 1 internal Windows host running database, rpc, and web, and 1 internal Linux host running ftp. Configuration 57 1 intruder, 1 internal Windows host running web, ftp, and database, and 1 internal Linux host running rpc. Configuration 58 1 intruder, 1 internal Windows host running database, ftp, and rpc, and 1 internal Linux host running web. Configuration 59 1 intruder, 1 internal Windows host running ftp, rpc, and web, and 1 internal Linux host running database. Configuration 60 1 intruder, 1 internal Windows host running web, and 1 internal Linux host running database, ftp, and rpc. Configuration 61 1 intruder, 1 internal Windows host running rpc, and 1 internal Linux host running database, web, and ftp. 48

Configuration 62 1 intruder, 1 internal Windows host running ftp, and 1 internal Linux host running rpc, web, and database. Configuration 63 1 intruder, 1 internal Windows host running database, and 1 internal Linux host running web, ftp, and rpc. Configuration 64 1 intruder, 1 internal Windows host running database and web, and 1 internal Linux host running ftp and rpc. Configuration 65 1 intruder, 1 internal Windows host running database and rpc, and 1 internal Linux host running ftp and web. Configuration 66 1 intruder, 1 internal Windows host running database and ftp, and 1 internal Linux host running web and rpc. Configuration 67 1 intruder, 1 internal Windows host running ftp and rpc, and 1 internal Linux host running database and web. Configuration 68 1 intruder, 1 internal Windows host running ftp and web, and 1 internal Linux host running database and rpc. Configuration 69 1 intruder, 1 internal Windows host running web and rpc, and 1 internal Linux host running database and ftp. Configuration 70 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running ftp, and 1 internal Windows host running database and rpc. Configuration 71 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running database, and 1 internal Windows host running rpc and ftp. Configuration 72 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running rpc, and 1 internal Windows host running database and ftp. Configuration 73 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running database, and 1 internal Windows host running web and rpc. Configuration 74 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running rpc, and 1 internal Windows host running web and database. Configuration 75 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running web, and 1 internal Windows host running database and rpc. Configuration 76 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running web, and 1 internal Windows host running database and ftp. Configuration 77 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running ftp, and 1internal Windows host running web and database. Configuration 78 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running database, and 1 internal Windows host running ftp and web. Configuration 79 1 intruder, 1 internal Solaris host running database, 1 internal Linux host running ftp, and 1internal Windows host running rpc and web. Configuration 80 1 intruder, 1 internal Solaris host running database, 1 intern al Linux host running web, and 1internal Windows host running ftp and rpc. Configuration 81 1 intruder, 1 internal Solaris host running database, 1internal Linux host running rpc, and 1 internal Windows host running web and ftp. Configuration 82 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running ftp, and 1 internal Linux host running database and rpc. Configuration 83 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running rpc, and 1 internal Linux host running database and ftp. Configuration 84 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running database, and 1 internal Linux host running rpc and ftp. Configuration 85 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running rpc, and 1 internal Linux host running web and database. Configuration 86 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running web, and 1 internal Linux host running database and rpc. 49

Configuration 87 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running database, and 1 internal Linux host running web and rpc. Configuration 88 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running web, and 1 internal Solaris host running rpc and database. Configuration 89 1 intruder, 1 internal Linux host running database, 1 internal Windows host running ftp, and 1 internal Solaris host running web and rpc. Configuration 90 1 intruder, 1 internal Linux host running database, 1 internal Windows host running web, and 1internal Solaris host running rpc and ftp. Configuration 91 1 intruder, 1 internal Linux host running database, 1 internal Windows host running rpc, and 1 internal Solaris host running web and ftp. Configuration 92 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running web, and 1 internal Solaris host running database and ftp. Configuration 93 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running ftp, and 1 internal Solaris host running database and web. Configuration 94 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running database, and 1 internal Solaris host running web and ftp. Configuration 95 1 intruder, 1 internal Linux host running web, 1 internal Windows host running ftp, and 1 internal Solaris host running rpc and database. Configuration 96 1 intruder, 1 internal Linux host running web, 1 internal Windows host running database, and 1internal Solaris host running rpc and ftp. Configuration 97 1 intruder, 1 internal Linux host running web, 1 internal Windows host running rpc, and 1 internal Solaris host running ftp and database. Configuration 98 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running ftp, and 1internal Linux host running rpc and web. Configuration 99 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running rpc, and 1 internal Linux host running ftp and web. Configuration 100 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running web, and 1 internal Linux host running rpc and ftp. Configuration 101 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running web, and 1 internal Linux host running ftp and database. Configuration 102 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running ftp, and 1internal Linux host running web and database. Configuration 103 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running database, and 1 internal Linux host running ftp and web. Configuration 104 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running database, and 1 internal Solaris host running web and rpc. Configuration 105 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running rpc, and 1 internal Solaris host running database and web. Firewalled Configurations Intruder can connect to ports 80 and 21. Configuration 106-1 intruder and 1 internal Solaris host running web, ftp, rpc, and database. Firewall between intruder and internal network. Configuration 107 1 intruder, 1 internal Solaris host running rpc, and 1 internal Solaris host running web, ftp, and database. 50

Configuration 108 1 intruder, 1 internal Solaris host running ftp, and 1 internal Solaris host running web, rpc, and database. Configuration 109 1 intruder, 1 internal Solaris host running web, and 1 internal Solaris host running rpc, ftp, and database. Configuration 110 1 intruder, 1 internal Solaris host running database, and 1 internal Solaris host running web, rpc, and ftp. Configuration 111 1 intruder, 1 internal Solaris host running database and web, and 1 internal Solaris host running ftp and rpc. Configuration 112 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Solaris host running ftp and web. Configuration 113 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Solaris host running rpc and web. Configuration 114 1 intruder, 1 internal Solaris host running database, 1 internal Solaris host running rpc, 1 internal Solaris host running ftp, and 1 internal Solaris host running web. Configuration 115 1 intruder and 1 internal Linux host running web, ftp, rpc, and database. Firewall between intruder and internal network. Configuration 116 1 intruder, 1 internal Linux host running rpc, and 1 internal Linux host running web, database, and ftp. Firewall between intruder and internal network. Configuration 117 1 intruder, 1 internal Linux host running ftp, and 1 internal Linux host running rpc, web, and database. Firewall between intruder and internal network. Configuration 118 1 intruder, 1 internal Linux host running web, and 1 internal Linux host running rpc, ftp, and database. Firewall between intruder and internal network. Configuration 119 1 intruder, 1 internal Linux host running database, and 1 internal Linux host running web, rpc, and ftp. Firewall between intruder and internal network. Configuration 120 1 intruder, 1 internal Linux host running web and database, and 1 internal Linux host running ftp and rpc. Configuration 121 1 intruder, 1 internal Linux host running rpc and database, and 1 internal Linux host running ftp and web. Configuration 122 1 intruder, 1 internal Linux host running ftp and database, and 1 internal Linux host running web and rpc. Configuration 123 1 intruder, 1 internal Linux host running database, 1 internal Linux host running rpc, 1 internal Linux host running ftp, and 1 internal Linux host running web. Configuration 124 1 intruder and 1 internal Windows host running web, rpc, ftp, and database. Firewall between intruder and internal network. Configuration 125 1 intruder, 1 internal Windows host running rpc, and 1 internal Windows host running ftp, web, and database. Configuration 126 1 intruder, 1 internal Windows host running ftp, and 1 internal Windows host running rpc, web, and database. Configuration 127 1 intruder, 1 internal Windows host running web, and 1 internal Windows host running ftp, rpc, and database. Configuration 128 1intruder, 1 internal Windows host running database, and 1 internal Windows host running rpc, ftp, and web. Configuration 129 1 intruder, 1 internal Windows host running web and database, and 1 internal Windows host running ftp and rpc. 51

Configuration 130 1 intruder, 1 internal Windows host running rpc and database, and 1 internal Windows host running ftp and web. Configuration 131 1 intruder, 1 internal Windows host running ftp and database, and 1 internal Windows host running rpc and web. Configuration 132 1 intruder, 1 internal Windows host running database, 1 internal Windows host running rpc, 1 internal Windows host running ftp, and 1 internal Windows host running web. Configuration 133 1 intruder, 1 internal Solaris host running web, rpc, and database, and 1 internal Windows host running ftp. Configuration 134 1 intruder, 1 internal Solaris host running database, ftp, and web, and 1 internal Windows host running rpc. Configuration 135 1 intruder, 1 internal Solaris host running database, ftp, and rpc, and 1 internal Windows host running web. Configuration 136 1 intruder, 1 internal Solaris host running web, ftp, and rpc, and 1 internal Windows host running database. Configuration 137 1 intruder, 1 internal Windows host running database, web, and rpc, and 1 internal Solaris host running ftp. Configuration 138 1 intruder, 1 internal Windows host running database, ftp, and web, and 1 internal Solaris host running rpc. Configuration 139 1 intruder, 1 internal Windows host running database, ftp, and rpc, and 1 internal Solaris host running web. Configuration 140 1 intruder, 1 internal Windows host running web, ftp, and rpc, and 1 internal Solaris host running database. Configuration 141 1 intruder, 1 internal Windows host running database and ftp, and 1 internal Solaris host running web and rpc. Configuration 142 1 intruder, 1 internal Windows host running database and web, and 1 internal Solaris host running ftp and rpc. Configuration 143 1 intruder, 1 internal Windows host running database and rpc, and 1 internal Solaris host running ftp and web. Configuration 144 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Windows host running ftp and web. Configuration 145 1 intruder, 1 internal Solaris host running database and web, and 1 internal Windows host running ftp and rpc. Configuration 146 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Windows host running web and rpc. Configuration 147 1 intruder, 1 internal Solaris host running database, rpc, and ftp, and 1 internal Linux host running web. Firewall between intruder and internal network. Configuration 148 1 intruder, 1 internal Solaris host running database, ftp, and web, and 1 internal Linux host running rpc. Firewall between intruder and internal network. Configuration 149 1 intruder, 1 internal Solaris host running database, rpc, and web, and 1interanal Linux host running ftp. Firewall between intruder and internal network. Configuration 150 1 intruder, 1 internal Solaris host running ftp, web, and rpc, and 1 internal Linux host running database. Firewall between intruder and internal network. Configuration 151 1 intruder, 1 internal Solaris host running ftp, and 1 internal Linux host running database, rpc, and web. Firewall between intruder and internal network. 52

Configuration 152 1 intruder, 1 internal Solaris host running rpc, and 1 internal Linux host running ftp, web, and database. Firewall between intruder and internal network. Configuration 153 1 intruder, 1 internal Solaris host running web, and 1 internal Linux host running ftp, rpc, and database. Firewall between intruder and internal network. Configuration 154 1 intruder, 1 internal Solaris host running database, and 1 internal Linux host running ftp, web, and rpc. Firewall between intruder and internal network. Configuration 155 1 intruder, 1 internal Solaris host running database and rpc, and 1 internal Linux host running ftp and web. Configuration 156 1 intruder, 1 internal Solaris host running database and web, and 1 internal Linux host running ftp and rpc. Configuration 157 1 intruder, 1 internal Solaris host running database and ftp, and 1 internal Linux host running web and rpc. Configuration 158 1 intruder, 1 internal Solaris host running rpc and ftp, and 1 internal Linux host running database and web. Configuration 159 1 intruder, 1 internal Solaris host running web and rpc, and 1 internal Linux host running database and ftp. Configuration 160 1 intruder, 1 internal Solaris host running web and ftp, and 1 internal Linux host running rpc and database. Configuration 161 1 intruder, 1 internal Windows host running database, rpc, and web, and 1 internal Linux host running ftp. Configuration 162 1 intruder, 1 internal Windows host running web, ftp, and database, and 1 internal Linux host running rpc. Configuration 163 1 intruder, 1 internal Windows host running database, ftp, and rpc, and 1 internal Linux host running web. Configuration 164 1 intruder, 1 internal Windows host running ftp, rpc, and web, and 1 internal Linux host running database. Configuration 165 1 intruder, 1 internal Windows host running web, and 1 internal Linux host running database, ftp, and rpc. Configuration 166 1 intruder, 1 internal Windows host running rpc, and 1 internal Linux host running database, web, and ftp. Configuration 167 1 intruder, 1 internal Windows host running ftp, and 1 internal Linux host running rpc, web, and database. Configuration 168 1 intruder, 1 internal Windows host running database, and 1 internal Linux host running web, ftp, and rpc. Configuration 169 1 intruder, 1 internal Windows host running database and web, and 1 internal Linux host running ftp and rpc. Configuration 170 1 intruder, 1 internal Windows host running database and rpc, and 1 internal Linux host running ftp and web. Configuration 171 1 intruder, 1 internal Windows host running database and ftp, and 1 internal Linux host running web and rpc. Configuration 172 1 intruder, 1 internal Windows host running ftp and rpc, and 1 internal Linux host running database and web. Configuration 173 1 intruder, 1 internal Windows host running ftp and web, and 1 internal Linux host running database and rpc. 53

Configuration 174 1 intruder, 1 internal Windows host running web and rpc, and 1 internal Linux host running database and ftp. Configuration 175 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running ftp, and 1 internal Windows host running database and rpc. Configuration 176 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running database, and 1 internal Windows host running rpc and ftp. Configuration 177 1 intruder, 1 internal Solaris host running web, 1 internal Linux host running rpc, and 1 internal Windows host running database and ftp. Configuration 178 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running database, and 1 internal Windows host running web and rpc. Configuration 179 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running rpc, and 1 internal Windows host running web and database. Configuration 180 1 intruder, 1 internal Solaris host running ftp, 1 internal Linux host running web, and 1 internal Windows host running database and rpc. Configuration 181 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running web, and 1 internal Windows host running database and ftp. Configuration 182 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running ftp, and 1internal Windows host running web and database. Configuration 183 1 intruder, 1 internal Solaris host running rpc, 1 internal Linux host running database, and 1 internal Windows host running ftp and web. Configuration 184 1 intruder, 1 internal Solaris host running database, 1 internal Linux host running ftp, and 1internal Windows host running rpc and web. Configuration 185 1 intruder, 1 internal Solaris host running database, 1 intern al Linux host running web, and 1internal Windows host running ftp and rpc. Configuration 186 1 intruder, 1 internal Solaris host running database, 1internal Linux host running rpc, and 1 internal Windows host running web and ftp. Configuration 187 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running ftp, and 1 internal Linux host running database and rpc. Configuration 188 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running rpc, and 1 internal Linux host running database and ftp. Configuration 189 1 intruder, 1 internal Solaris host running web, 1 internal Windows host running database, and 1 internal Linux host running rpc and ftp. Configuration 190 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running rpc, and 1 internal Linux host running web and database. Configuration 191 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running web, and 1 internal Linux host running database and rpc. Configuration 192 1 intruder, 1 internal Solaris host running ftp, 1 internal Windows host running database, and 1 internal Linux host running web and rpc. Configuration 193 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running web, and 1 internal Solaris host running rpc and database. Configuration 194 1 intruder, 1 internal Linux host running database, 1 internal Windows host running ftp, and 1 internal Solaris host running web and rpc. Configuration 195 1 intruder, 1 internal Linux host running database, 1 internal Windows host running web, and 1internal Solaris host running rpc and ftp. 54

Configuration 196 1 intruder, 1 internal Linux host running database, 1 internal Windows host running rpc, and 1 internal Solaris host running web and ftp. Configuration 197 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running web, and 1 internal Solaris host running database and ftp. Configuration 198 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running ftp, and 1 internal Solaris host running database and web. Configuration 199 1 intruder, 1 internal Linux host running rpc, 1 internal Windows host running database, and 1 internal Solaris host running web and ftp. Configuration 200 1 intruder, 1 internal Linux host running web, 1 internal Windows host running ftp, and 1 internal Solaris host running rpc and database. Configuration 201 1 intruder, 1 internal Linux host running web, 1 internal Windows host running database, and 1internal Solaris host running rpc and ftp. Configuration 202 1 intruder, 1 internal Linux host running web, 1 internal Windows host running rpc, and 1 internal Solaris host running ftp and database. Configuration 203 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running ftp, and 1internal Linux host running rpc and web. Configuration 204 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running rpc, and 1 internal Linux host running ftp and web. Configuration 205 1 intruder, 1 internal Solaris host running database, 1 internal Windows host running web, and 1 internal Linux host running rpc and ftp. Configuration 206 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running web, and 1 internal Linux host running ftp and database. Configuration 207 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running ftp, and 1internal Linux host running web and database. Configuration 208 1 intruder, 1 internal Solaris host running rpc, 1 internal Windows host running database, and 1 internal Linux host running ftp and web. Configuration 209 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running database, and 1 internal Solaris host running web and rpc. Configuration 210 1 intruder, 1 internal Linux host running ftp, 1 internal Windows host running rpc, and 1 internal Solaris host running database and web. DMZ Configurations Intruder can only connect to ports 80 on host in DMZ. Configuration 211 1 intruder, 1 DMZ Solaris host running web, and 1 internal Solaris host running database, ftp, and rpc. Configuration 212 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database and ftp, and 1 internal Solaris host running rpc. Configuration 213 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database and rpc, and 1 internal Solaris host running ftp. Configuration 214 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database, and 1 internal Solaris host running rpc and ftp. Configuration 215 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database, 1 internal Solaris host running ftp, and 1 internal Solaris host running rpc. Configuration 216 1 intruder, 1 DMZ Linux host running web, and 1 internal Linux host running database, ftp, and rpc. Configuration 217 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running database and ftp, and 1 internal Linux host running rpc. 55

Configuration 218 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running database and rpc, and 1 internal Linux host running ftp. Configuration 219 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running database, and 1 internal Linux host running rpc and ftp. Configuration 220 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running database, 1 internal Linux host running ftp, and 1 internal Linux host running rpc. Configuration 221 1 intruder, 1 DMZ Windows host running web, and 1 internal Windows host running database, ftp, and rpc. Configuration 222 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database and ftp, and 1 internal Windows host running rpc. Configuration 223 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database and rpc, and 1 internal Windows host running ftp. Configuration 224 1 intruder, 1DMZ Windows host running web, 1 internal Windows host running database, and 1 internal Windows host running rpc and ftp. Configuration 225 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database, 1internal Windows host running ftp, and 1 internal Windows host running rpc. Configuration 226 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database and rpc, and 1 internal Windows host running ftp. Configuration 227 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database and ftp, and 1internal Windows host running rpc. Configuration 228 1 intruder, 1 DMZ Windows host running web, and 1 internal Solaris host running database, ftp, and rpc. Configuration 229 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running ftp and rpc, and 1 internal Windows host running database. Configuration 230 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database and rpc, and 1 internal Solaris host running ftp. Configuration 231 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database and ftp, and 1 internal Solaris host running rpc. Configuration 232 1 intruder, 1 DMZ Solaris host running web, and 1 internal Windows host running ftp, rpc, and database. Configuration 233 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running ftp and rpc, and 1 internal Solaris host running database. Configuration 234 1 intruder, 1 DMZ Solaris host running web, 1 internal Windows host running database and ftp, and 1internal Solaris host running rpc. Configuration 235 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database, and 1 internal Solaris host running ftp and rpc. Configuration 236 1 intruder, 1 DMZ Solaris host running web, 1 internal Windows host running database and rpc, and 1 internal Solaris host running ftp. Configuration 237 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running database and rpc, and 1 internal Windows host running ftp. Configuration 238 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database, and 1 internal Windows host running ftp and rpc. Configuration 239 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running rpc, and 1 internal Solaris host running database and ftp. Configuration 240 1 intruder, 1 DMZ Linux host running web and 1 internal Solaris host running ftp, rpc, and database. 56

Configuration 241 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running ftp and database, and 1 internal Linux host running rpc. Configuration 242 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database and rpc, and 1 internal Linux host running ftp. Configuration 243 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running ftp and rpc, and 1internal Linux host running database. Configuration 244 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running ftp, and 1 internal Linux host running database and rpc. Configuration 245 1 intruder 1 DMZ Linux host running web, 1 internal Solaris host running rpc, and 1 internal Linux host running ftp and database. Configuration 246 1 intruder, 1 DMZ Solaris host running web, and 1 internal Linux host running ftp, rpc, and database. Configuration 247 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running database, and 1 internal Linux host running ftp and rpc. Configuration 248 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running database and rpc, and 1 internal Linux host running ftp. Configuration 249 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running database, and 1internal Linux host running ftp and rpc. Configuration 250 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running rpc, and 1 internal Solaris host running database and ftp. Configuration 251 I intruder, 1 DMZ Linux host running web, 1 internal Linux host running database, and 1 internal Solaris host running rpc and ftp. Configuration 252 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running rpc, and 1 internal Linux host running ftp and database. Configuration 253 1 intruder, 1 DMZ Solaris host running web, 1 internal Solaris host running ftp, and 1 internal Linux host running database and rpc. Configuration 254 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database and rpc, and 1internal Linux host running ftp. Configuration 255 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running ftp and database, and 1 internal Linux host running rpc. Configuration 256 1 intruder, 1 DMZ Linux host running web, and 1 internal Windows host running ftp, rpc, and database. Configuration 257 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running ftp and rpc, and 1 internal Linux host running database. Configuration 258 1 intruder, 1 DMZ Windows host running web, and 1 internal Linux host running ftp, rpc, and database. Configuration 259 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running rpc, and 1 internal Linux host running database and ftp. Configuration 260 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running ftp, and 1 internal Linux host running rpc and database. Configuration 261 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running database, and 1 internal Linux host running ftp and rpc. Configuration 262 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running database, and 1 internal Linux host running ftp and rpc. Configuration 263 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running ftp, and1 internal Windows host running database and rpc. 57

Configuration 264 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running rpc, and 1 internal Windows host running database and ftp. Configuration 265 1 intruder, 1 DMZ Linux host running web, 1 internal Linux host running database, and 1 internal Windows host running ftp and rpc. Configuration 266 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running ftp, and 1 internal Linux host running rpc and database. Configuration 267 1 intruder, 1 DMZ Windows host running web, 1 internal Windows host running rpc, and 1 internal Linux host running ftp and database. Configuration 268 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running ftp, and 1 internal Windows host running rpc and database. Configuration 269 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running database, and 1 internal Windows host running rpc and ftp. Configuration 270 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running rpc, and 1 internal Windows host running ftp and database. Configuration 271 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running ftp, 1 internal Linux host running database, and 1 internal Windows host running rpc. Configuration 272 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running ftp, 1 internal Linux host running rpc, and 1 internal Windows host running database. Configuration 273 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running ftp, and 1 internal Windows host running rpc and database. Configuration 274 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running rpc, and 1 internal Windows host running ftp and database. Configuration 275 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running rpc, and 1 internal Windows host running ftp and database. Configuration 275 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running rpc, 1 internal Linux host running ftp, and 1 internal Windows host running database. Configuration 276 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running rpc, 1 internal Linux host running database, and 1 internal Windows host running ftp. Configuration 277 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running database, 1 internal Linux host running ftp, and 1 internal Windows host running rpc. Configuration 278 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running database, and 1 internal Windows host running ftp and rpc. Configuration 279 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running database, 1 internal Linux host running rpc, and 1 internal Windows host running ftp. Configuration 280 1 intruder, 1 DMZ Solaris host running web, 1 internal Windows host running ftp, and 1 internal Linux host running rpc and database. Configuration 281 1 intruder, 1 DMZ Solaris host running web, 1 internal Windows host running rpc, and 1 internal Linux host running ftp and database. Configuration 282 1 intruder, 1 DMZ Solaris host running web, 1 internal Windows host running database, and 1 internal Linux host running rpc and ftp. Configuration 283 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running ftp, 1 internal Windows host running rpc, and 1 internal Linux host running database. Configuration 284 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running ftp, and 1 internal Linux host running rpc and database. 58

Configuration 285 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running ftp, 1 internal Windows host running database, and 1 internal Linux host running rpc. Configuration 286 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running database, 1 internal Windows host running ftp, and 1 internal Linux host running rpc. Configuration 287 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running database, 1 internal Windows host running rpc, and 1 internal Linux host running ftp. Configuration 288 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running database, and 1 internal Linux host running ftp and database. Configuration 289 1 intruder, 1 DMZ Windows host running web, 1 internal Solaris host running rpc, and 1 internal Linux host running ftp and database. Configuration 290 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running rpc, 1 internal Windows host running ftp, and 1 internal Linux host running database. Configuration 291 1 intruder, 1 DMZ Linux host running web, 1 internal Solaris host running rpc, 1 internal Windows host running database, and 1 internal Linux host running ftp. Configuration 292 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running ftp, 1 internal Windows host running database, and 1 internal Solaris host running rpc. Configuration 293 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running ftp, 1 internal Windows host running rpc, and 1 internal Solaris host running database. Configuration 294 1 intruder, 1 DMZ Windows host running web, 1 internal Linux host running ftp, and 1 internal Solaris host running rpc and database. Configuration 295 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running database, 1 internal Windows host running ftp, and 1 internal Solaris host running rpc. Configuration 296 1 intruder, 1 DMZ Windows host running web, 1 internal Linux host running database, and 1 internal Solaris host running ftp and rpc. Configuration 297 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running database, 1 internal Windows host running rpc, and 1 internal Solaris host running ftp. Configuration 298 1 intruder, 1 DMZ Windows host running web, 1 internal Linux host running rpc, and 1 internal Solaris host running ftp and database. Configuration 299 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running rpc, 1 internal Windows host running ftp, and 1 internal Solaris host running database. Configuration 300 1 intruder, 1 DMZ Solaris host running web, 1 internal Linux host running rpc, 1 internal Windows host running database, and 1internal Solaris host running ftp. Configuration 301 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running ftp, and 1 internal Solaris host running rpc and database. Configuration 302 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running database, and 1 internal Solaris host running ftp and rpc. Configuration 303 1 intruder, 1 DMZ Linux host running web, 1 internal Windows host running rpc, and 1 internal Solaris host running ftp and database. 59

Appendix D: Sample Nessus Scan Result IP addresses and any other information that could potentially identify this host have been removed. dec_dlm (625/tcp) 11219 NOTE Port is open; afpovertcp (548/tcp) 11219 NOTE Port is open; afpovertcp (548/tcp) 10666 NOTE Synopsis :;;File sharing service is available.;;description :;;The remote host is running an AppleShare IP file service.;by sending DSIGetStatus request, it was;possible to disclose information about the remote host.;;risk Factor :;;None / CVSS Base : 0 ;(AV:R/AC:L/Au:NR/C:N/A:N/I:N/B:N);;Plugin output :;;This host is running an AppleShare File Services over IP.; Machine type: Macintosh; surf (1010/tcp) 11219 NOTE Port is open; surf (1010/tcp) 11111 NOTE RPC program 100024 version 1 'status' is running on this port;; vnc (5900/tcp) 11219 NOTE Port is open; vnc (5900/tcp) 19288 NOTE The remote VNC server supports those security types:;+ 30;; netbios-ssn (139/tcp) 11219 NOTE Port is open; netbios-ssn (139/tcp) 11011 NOTE An SMB server is running on this port; sunrpc (111/tcp) 11219 NOTE Port is open; sunrpc (111/tcp) 10223 NOTE The RPC portmapper is running on this port.;;an attacker may use it to enumerate your list;of RPC services. We recommend you filter traffic;going to this port.;;risk Factor : Low;CVE : CVE-1999-0632, CVE-1999-0189, CVE-1999-0632, CVE-1999-0189;BID : 205, 205; sunrpc (111/tcp) 11111 NOTE RPC program 100000 version 2 'portmapper' (portmap sunrpc rpcbind) is running on this port;; nfs (2049/tcp) 11219 NOTE Port is open; nfs (2049/tcp) 11111 NOTE RPC program 100003 version 2 'nfs' (nfsprog) is running on this port;rpc program 100003 version 3 'nfs' (nfsprog) is running on this port;; ssh (22/tcp) 11219 NOTE Port is open; ssh (22/tcp) 10330 NOTE An ssh server is running on this port; 60

ssh (22/tcp) 10267 NOTE Remote SSH version : SSH-2.0-OpenSSH_3.8.1p1;;Remote SSH supported authentication ssh (22/tcp) 10881 NOTE The remote SSH daemon supports the following versions of the;ssh protocol :;;. 1.99;. 2.0;;;SSHv2 svrloc (427/tcp) 11219 NOTE Port is open; svrloc (427/tcp) 23777 NOTE Synopsis :;;The remote server supports the Service Location Protocol. ;;Description :;;The remote server understands Service Location Protocol (SLP), a;protocol that allows network applications to discover the existence,;location, and configuration of various services in an enterprise;network environment. A server that understands SLP can either be a;service agent (SA), which knows the location of various services, or a;directory agent (DA), which acts as a central repository for service;location information. ;;See Also :;;http://www.ietf.org/rfc/rfc2608.txt;;solution:;;limit incoming traffic to this port if desired. ;;Risk Factor :;;None;;Plugin output :;;An SLP Service Agent is listening on this port.; microsoft-ds (445/tcp) 11219 NOTE Port is open; microsoft-ds (445/tcp) 11011 NOTE A CIFS server is running on this port; microsoft-ds (445/tcp) 10785 NOTE Synopsis :;;It is possible to obtain information about the remote operating;system.;;description :;;It is possible to get the remote operating system name and;version (Windows and/or Samba) by sending an authentication;request.;;risk Factor :;;None;;Plugin output :;;The remote Operating System is : Unix;The remote native lan manager is : Samba 3.0.10; microsoft-ds (445/tcp) 25240 NOTE Synopsis :;;An SMB server is running on the remote host.;;description :;;The remote host is running a SAMBA server, a CIFS/SMB;server for Unix.;;See Also :;;http://www.samba.org;;risk Factor :;;None; microsoft-ds (445/tcp) 25216 HOLE Synopsis :;;It is possible to execute code on the remote host through samba.;;description :;;The version of the Samba server installed on the remote host is affected;by multiple heap overflow vulnerabilities, which can be exploited remotely;to execute code with the privileges of the samba daemon.;;see Also :;;http://www.samba.org/samba/security/cve-2007-2446.html;;solution:;;upgrade to Samba version 3.0.25 or later. ;;Risk Factor :;;Critical / CVSS Base : 10.0;(AV:R/AC:L/Au:NR/C:C/I:C/A:C/B:N);CVE : CVE-2007-2446;BID : 23973, 24195, 24196, 24197, 24198;Other references : OSVDB:34732; microsoft-ds (445/tcp) 10394 NOTE Synopsis :;;It is possible to logon on the remote host.;;description :;;The remote host is running one of the Microsoft Windows operating;system. It was possible to logon 61

using one of the following;account :;;- NULL session;- Guest account;- Given Credentials;;See Also :;;http://support.microsoft.com/support/kb/articles/q143/4/74.asp;http://support.microsoft.com/support/kb/ articles/q246/2/61.asp;;risk Factor :;;none;;plugin output :;;- NULL sessions are enabled on the remote host;;cve : CVE-1999-0504, CVE-1999-0506, CVE-2000-0222, CVE-1999-0505, CVE-2002-1117, CVE- 1999-0504, CVE-1999-0506, CVE-2000-0222, CVE-1999-0505, CVE-2002-1117;BID : 494, 990, 11199, 494, 990, 11199; microsoft-ds (445/tcp) 10397 NOTE Synopsis :;;It is possible to obtain network information.;;description :;;It was possible to obtain the browse list of the remote;windows system by send a request to the LANMAN pipe.;the browse list is the list of the nearest Windows systems;of the remote host. ;;Risk Factor :;;None microsoft-ds (445/tcp) 17651 NOTE Synopsis :;;It is possible to retrieve password policy using the supplied credentials.;;description :;;Using the supplied credentials it was possible to extract the password;policy.;password policy must be conform to the Informationnal System Policy.;;Risk Factor :;;None / CVSS Base : 0 ;(AV:L/AC:H/Au:R/C:N/A:N/I:N/B:N);;Plugin output :;;The following password policy is defined on the remote host:;maximum password age (d): 0;Password must meet complexity requirements: Enabled;Minimum password age (d): 0;Forced logoff time (s): Not set;locked account time (s): 1800;Time between failed logon (s): 1800;Number of invalid logon before locked out (s): 0;; microsoft-ds (445/tcp) 10395 INFO Synopsis :;;It is possible to enumerate remote network shares.;;description :;;By connecting to the remote host using a NULL (or guest) session;nessus was able to enumerates the network share names.;;risk Factor :;;None;;Plugin output :;; microsoft-ds (445/tcp) 10859 NOTE Synopsis :;;It is possible to obtain remote host SID.;;Description :;;By emulating the call to LsaQueryInformationPolicy() it was;possible to obtain the host SID (Security Identifier).;;The host SID can then be used to get the list of local users.;;risk Factor : ;;None;;Plugin output ;CVE : CVE-2000-1200;BID : 959; asip-webadmin (311/tcp) 11219 NOTE Port is open; asip-webadmin (311/tcp) 10330 NOTE A SSLv2 server answered on this port;; asip-webadmin (311/tcp) 10330 NOTE A web server is running on this port through SSL; asip-webadmin (311/tcp) 21643 NOTE Synopsis :;;The remote service supports the use of weak SSL ciphers.;;description :;;The remote host supports the use of SSL ciphers that;offer either weak encryption or no encryption at all.;;see Also :;;http://www.openssl.org/docs/apps/ciphers.html;;solution:;;reconfigure the affected application if possible to avoid use of ;weak ciphers.;;risk Factor : ;;Low / CVSS Base 62

: 2 ;(AV:R/AC:L/Au:NR/C:P/A:N/I:N/B:N);;Plugin output :;;Here is a list of the SSL ciphers supported by the remote server :;; Export Ciphers; SSLv2; EXP-RC2-CBC-MD5 Kx=RSA(512) Au=RSA Enc=RC2(40) Mac=MD5 export ; EXP-RC4-MD5 Kx=RSA(512) Au=RSA Enc=RC4(40) Mac=MD5 export ; TLSv1; EXP1024-DES-CBC-SHA Kx=RSA(1024) Au=RSA Enc=DES(56) Mac=SHA1 export ; EXP1024-RC2-CBC-MD5 Kx=RSA(1024) Au=RSA Enc=RC2(56) Mac=MD5 export ; EXP1024-RC4-MD5 Kx=RSA(1024) Au=RSA Enc=RC4(56) Mac=MD5 export ; EXP1024-RC4-SHA Kx=RSA(1024) Au=RSA Enc=RC4(56) Mac=SHA1 export ;; Low Strength Ciphers (excluding export, < 128-bit key); SSLv2; DES-CBC-MD5 Kx=RSA Au=RSA Enc=DES(56) Mac=MD5 ; RC4-64-MD5 Kx=RSA Au=RSA Enc=RC4(64) Mac=MD5 ; SSLv3; DES-CBC- SHA Kx=RSA Au=RSA Enc=DES(56) Mac=SHA1 ; TLSv1; DES-CBC-SHA Kx=RSA Au=RSA Enc=DES(56) Mac=SHA1 ;; Medium Strength Ciphers (128-bit key); SSLv2; RC2-CBC-MD5 Kx=RSA Au=RSA Enc=RC2(128) Mac=MD5 ; RC4- MD5 Kx=RSA Au=RSA Enc=RC4(128) Mac=MD5 ; SSLv3; RC4-MD5 Kx=RSA Au=RSA Enc=RC4(128) Mac=MD5 ; RC4-SHA Kx=RSA Au=RSA Enc=RC4(128) Mac=SHA1 ; TLSv1; RC4-MD5 Kx=RSA Au=RSA Enc=RC4(128) Mac=MD5 ; RC4-SHA Kx=RSA Au=RSA Enc=RC4(128) Mac=SHA1 ;; High Strength Ciphers (> 128-bit key); SSLv2; DES-CBC3-MD5 Kx=RSA Au=RSA Enc=3DES(168) Mac=MD5 ; SSLv3; DES-CBC3-SHA Kx=RSA Au=RSA Enc=3DES(168) Mac=SHA1 ; TLSv1; DES-CBC3-SHA Kx=RSA Au=RSA Enc=3DES(168) Mac=SHA1 ; AES128-SHA Kx=RSA Au=RSA Enc=AES(128) Mac=SHA1 ; AES256-SHA Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 ;;The fields above are :;; {OpenSSL ciphername}; Kx={key exchange}; Au={authentication}; Enc={symmetric encryption method}; Mac={message authentication code}; {export flag};; asip-webadmin (311/tcp) 15901 INFO The SSL certificate of the remote service expired asip-webadmin (311/tcp) 20007 NOTE Synopsis :;;The remote service encrypts traffic using a protocol with known;weaknesses. ;;Description :;;The remote service accepts connections encrypted using SSL 2.0, which;reportedly suffers from several cryptographic flaws and has been;deprecated for several years. An attacker may be able to exploit these;issues to conduct man-in-the-middle attacks or decrypt communications;between the affected service and clients. ;;See Also :;;http://www.schneier.com/paperssl.pdf;;solution: ;;Consult the application's documentation to disable SSL 2.0 and use SSL;3.0 or TLS 1.0 instead.;;risk Factor : ;;Low / CVSS Base : 2 ;(AV:R/AC:L/Au:NR/C:P/A:N/I:N/B:N); tftp (69/udp) 11819 NOTE Synopsis :;;A TFTPD server is listening on the remote port.;;description :;;The remote host is running a TFTPD (Trivial File Transfer Protocol).;TFTPD is often used by routers and 63

diskless hosts to retrieve their;configuration. It is also used by worms to propagage.;;solution: ;;If you do not use this service, you should disable it.;;risk Factor :;;None;CVE : CVE-1999-0616, CVE-1999-0616; general/tcp 25220 NOTE Synopsis :;;The remote service implements TCP timestamps.;;description :;;The remote host implements TCP timestamps, as defined by RFC1323.;A side effect of this feature is that the uptime of the remote ;host can be sometimes be computed.;;see Also :;;http://www.ietf.org/rfc/rfc1323.txt;;risk Factor : ;;None; general/tcp 23971 NOTE Synopsis :;;The physical network is set up in a potentially insecure way.;;description :;;The remote host is on a different logical network. However, it is;on the same physical subnet.;;an attacker connecting from the same network as your Nessus;scanner is on could reconfigure his system to force it;to belong to the subnet of the remote host.;;this makes any filtering between the two subnets useless.;;solution: ;;Use VLANs to separate different logical networks.;;risk Factor : ;;Low;Plugin output :; general/tcp 11936 NOTE The remote host is running Mac OS X 10.4.6; general/tcp 19506 NOTE Information about this scan : ;;Nessus version : 3.0.5;Plugin feed version : 200609281415;Type of plugin feed : Registered (7 days delay);scanner IP :;Port scanner(s) : synscan ;Port range : default;thorough tests : no;experimental tests : no;paranoia level : 1;Report Verbosity : 1;Safe checks : yes;max hosts : 20;Max checks : 4;Scan Start Date : 2007/6/28 14:21;Scan duration : 448 sec;; unknown (7938/tcp) 19557 NOTE Synopsis :;;A backup software is running on this port.;;description :;;The remote host is running EMC Legato Networker backup software or one;of these derivates :;; - Sun StorEdge Enterprise Backup Software; - Sun Solstice Backup Software;;;Solution:;;If this service is not needed, disable it or filter incoming traffic;to this port.;;risk Factor : ;;None; unknown (7938/tcp) 19558 HOLE Synopsis :;;Arbitrary code can be executed on the remote host.;;description :;;The remote host is running one of the following product :;; - Legato Networker; - Sun StorEdge Enterprise Backup Software; - Sun Solstice Backup Software;;The remote version of this software is vulnerable to denial of service,;unauthorized access and remote command execution vulnerabilities.;;solution:;;http://www.legato.com/support/websupport/product_alerts/081605_nw- 7x.htm;http://sunsolve.sun.com/search/document.do?assetkey=1-26-101886-1;;Risk Factor :;;Critical / CVSS Base : 10 ;(AV:R/AC:L/Au:NR/C:C/A:C/I:C/B:N);CVE : CVE-2005-0357, CVE-2005-0358, CVE-2005-0359, CVE-2005-0357, CVE-2005-0358, CVE-2005-0359;BID : 14582, 14582; 64

unknown (1007/tcp) 11111 NOTE RPC program 100005 version 1 'mountd' (mount showmount) is running on this port;rpc program 100005 version 3 'mountd' (mount showmount) is running on this port;; unknown (1009/tcp) 11111 NOTE RPC program 100021 version 0 'nlockmgr' is running on this port;rpc program 100021 version 1 'nlockmgr' is running on this port;rpc program 100021 version 3 'nlockmgr' is running on this port;rpc program 100021 version 4 'nlockmgr' is running on this port;; unknown (7937/tcp) 11111 NOTE RPC program 390113 version 1 'nsrexec' (nsrexecd NetWorker client execution service) is running on this port;; sunrpc (111/udp) 11111 NOTE RPC program 100000 version 2 'portmapper' (portmap sunrpc rpcbind) is running on this port;; ftps-data (989/udp) 11111 NOTE RPC program 100005 version 1 'mountd' (mount showmount) is running on this port;rpc program 100005 version 3 'mountd' (mount showmount) is running on this port;; ufsd (1008/udp) 11111 NOTE RPC program 100021 version 0 'nlockmgr' is running on this port;rpc program 100021 version 1 'nlockmgr' is running on this port;rpc program 100021 version 3 'nlockmgr' is running on this port;rpc program 100021 version 4 'nlockmgr' is running on this port;; unknown (1019/udp) 11111 NOTE RPC program 100024 version 1 'status' is running on this port;; nfs (2049/udp) 11111 NOTE RPC program 100003 version 2 'nfs' (nfsprog) is running on this port;rpc program 100003 version 3 'nfs' (nfsprog) is running on this port;; ntp (123/udp) 10884 NOTE It is possible to determine a lot of information about the remote host ;by querying the NTP (Network Time Protocol) variables - these include ;OS descriptor, and time settings.;;it was possible to gather the following information from the remote NTP: processor='power Macintosh', system='darwin8.6.0', leap=3, stratum=16,;precision=-18, rootdelay=0.000, peer=0,;refid=0.0.0.0, reftime=0x00000000.00000000, poll=4, state=0, offset=0.000, frequency=0.000,;jitter=0.004, stability=0.000;;;;quickfix: Set NTP to restrict default access to ignore all info packets:; restrict default ignore;;risk Factor : Low; 65

Appendix E: Configuration Data Table E.1. Network Configuration Results Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility BASE CONFIGURATIONS 1 1 1 1 1 1 9.35 10 10.2 2 2 1 1 1 4 9.35 7.5 2.5 3.6 3 2 1 1 1 4 9.35 7.5 2.5 3.6 4 2 1 1 1 4 9.35 2.5 2.5 5.2 5 2 1 1 1 4 9.35 7.5 2.5 3.6 6 2 1 1 1 4 9.35 5 2.5 4.4 7 2 1 1 1 4 9.35 5 2.5 4.4 8 2 1 1 1 4 9.35 5 2.5 4.4 9 4 1 1 1 64 9.35 2.5.2 6.0 10 1 1 1 1 1 9.35 10 10.2 11 2 1 1 1 4 9.35 7.5 2.5 3.6 12 2 1 1 1 4 9.35 2.5 2.5 5.2 13 2 1 1 1 4 9.35 7.5 2.5 3.6 14 2 1 1 1 4 9.35 7.5 2.5 3.6 15 2 1 1 1 4 9.35 5 2.5 4.4 16 2 1 1 1 4 9.35 5 2.5 4.4 17 2 1 1 1 4 9.35 5 2.5 4.4 18 4 1 1 1 64 9.35 2.5.2 6.0 19 1 1 1 1 1 9.35 10 10.2 20 2 1 1 1 4 9.35 2.5 2.5 5.2 21 2 1 1 1 4 9.35 7.5 2.5 3.6 22 2 1 1 1 4 9.35 7.5 2.5 3.6 23 2 1 1 1 4 9.35 7.5 2.5 3.6 66

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 24 2 1 1 1 4 9.35 5 2.5 4.4 25 2 1 1 1 4 9.35 5 2.5 4.4 26 2 1 1 1 4 9.35 5 2.5 4.4 27 4 1 1 1 64 9.35 2.5.2 6.0 28 2 1 2 1 4 9.35 7.5 2.5 3.6 29 2 2 2 4 4 10 10 10 0 30 2 0 2 0 4 0 0 0 10 31 2 1 2 1 4 9.35 7.5 2.5 3.6 32 2 1 2 1 4 9.35 7.5 2.5 3.6 33 2 0 2 0 4 0 0 0 10 34 2 2 2 4 4 10 10 10 0 35 2 1 2 1 4 9.35 7.5 2.5 3.6 36 2 1 2 1 4 9.35 5 2.5 4.4 37 2 0 2 0 4 0 0 0 10 38 2 2 2 4 4 10 10 10 0 39 2 0 2 0 4 0 0 0 10 40 2 2 2 4 4 10 10 10 0 41 2 1 2 1 4 9.35 5 2.5 4.4 42 2 0 2 0 4 0 0 0 10 43 2 1 2 1 4 9.35 7.5 2.5 3.6 44 2 2 2 4 4 10 10 10 0 45 2 1 2 1 4 9.35 7.5 2.5 3.6 46 2 0 2 0 4 0 0 0 10 47 2 1 2 1 4 9.35 7.5 2.5 3.6 67

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 48 2 2 2 4 4 10 10 10 0 49 2 1 2 1 4 9.35 7.5 2.5 3.6 50 2 1 2 1 4 9.35 5 2.5 4.4 51 2 2 2 4 4 10 10 10 0 52 2 0 2 0 4 0 0 0 10 53 2 0 2 0 4 0 0 0 10 54 2 2 2 4 4 10 10 10 0 55 2 1 2 1 4 9.35 5 2.5 4.4 56 2 2 2 4 4 10 10 10 0 57 2 0 2 0 4 0 0 0 10 58 2 1 2 1 4 9.35 7.5 2.5 3.6 59 2 1 2 1 4 9.35 7.5 2.5 3.6 60 2 1 2 1 4 9.35 7.5 2.5 3.6 61 2 2 2 4 4 10 10 10 0 62 2 0 2 0 4 0 0 0 10 63 2 1 2 1 4 9.35 7.5 2.5 3.6 64 2 1 2 1 4 9.35 5 2.5 4.4 65 2 2 2 4 4 10 10 10 0 66 2 0 2 0 4 0 0 0 10 67 2 1 2 1 4 9.35 5 2.5 4.4 68 2 0 2 0 4 0 0 0 10 69 2 2 2 4 4 10 10 10 0 70 3 3 3 15 15 10 10 10 0 71 3 2 3 4 15 10 7.5 2.67 3.3 68

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 72 3 1 3 1 15 9.35 2.5.7 5.8 73 3 1 3 1 15 9.35 5.7 5.0 74 3 0 3 0 15 0 0 0 10 75 3 1 3 1 15 9.35 5.7 5.0 76 3 0 3 0 15 0 0 0 10 77 3 1 3 1 15 9.35 2.5.7 5.8 78 3 0 3 0 15 0 0 0 10 79 3 2 3 4 15 10 7.5 2.67 3.3 80 3 1 3 1 15 9.35 5.7 5.0 81 3 0 3 0 15 0 0 0 10 82 3 1 3 1 15 9.35 2.5.7 5.8 83 3 3 3 15 15 10 10 10 0 84 3 2 3 4 15 10 7.5 2.67 3.3 85 3 1 3 1 15 9.35 2.5.7 5.8 86 3 0 3 0 15 0 0 0 10 87 3 0 3 0 15 0 0 0 10 88 3 0 3 0 15 0 0 0 10 89 3 2 3 4 15 10 7.5 2.67 3.3 90 3 1 3 1 15 9.35 5.7 5.0 91 3 1 3 1 15 9.35 5.7 5.0 92 3 0 3 0 15 0 0 0 10 93 3 1 3 1 15 9.35 5.7 5.0 94 3 2 3 4 15 10 7.5 2.67 3.3 95 3 3 3 15 15 10 10 10 0 69

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 96 3 1 3 1 15 9.35 2.5.7 5.8 97 3 1 3 1 15 9.35 5.7 5.0 98 3 0 3 0 15 0 0 0 10 99 3 2 3 4 15 10 7.5 2.67 3.3 100 3 0 3 0 15 0 0 0 10 101 3 1 3 1 15 9.35 5.7 5.0 102 3 1 3 1 15 9.35 5.7 5.0 103 3 0 3 0 15 0 0 0 10 104 3 0 3 0 15 0 0 0 10 105 3 1 3 1 15 9.35 2.5.7 5.8 FIREWALLED CONFIGURATIONS 106 1 1 1 1 1 9.35 10 10.2 107 2 1 1 1 4 9.35 7.5 2.5 3.6 108 2 1 1 1 4 9.35 7.5 2.5 3.6 109 2 1 1 1 4 9.35 2.5 2.5 5.2 110 2 1 1 1 4 9.35 7.5 2.5 3.6 111 2 1 1 1 4 9.35 5 2.5 4.4 112 2 1 1 1 4 9.35 5 2.5 4.4 113 2 1 1 1 4 9.35 5 2.5 4.4 114 4 1 1 1 64 9.35 2.5.2 6.0 115 1 1 1 1 1 9.35 10 10.2 116 2 1 1 1 4 9.35 7.5 2.5 3.6 117 2 1 1 1 4 9.35 2.5 2.5 5.2 118 2 1 1 1 4 9.35 7.5 2.5 3.6 70

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 119 2 1 1 1 4 9.35 7.5 2.5 3.6 120 2 1 1 1 4 9.35 5 2.5 4.4 121 2 1 1 1 4 9.35 5 2.5 4.4 122 2 1 1 1 4 9.35 5 2.5 4.4 123 4 1 1 1 64 9.35 2.5.2 6.0 124 1 1 1 0 1 9.35 10 0 1.3 125 2 1 1 0 4 9.35 2.5 0 6.1 126 2 1 1 0 4 9.35 7.5 0 4.4 127 2 1 1 0 4 9.35 7.5 0 4.4 128 2 1 1 0 4 9.35 7.5 0 4.4 129 2 1 1 0 4 9.35 5 0 5.2 130 2 1 1 0 4 9.35 5 0 5.2 131 2 1 1 0 4 9.35 5 0 5.2 132 4 1 1 0 64 9.35 2.5 0 6.1 133 2 1 2 1 4 9.35 7.5 2.5 3.6 134 2 2 2 2 4 10 10 5 1.7 135 2 0 2 0 4 0 0 0 10 136 2 1 2 1 4 9.35 7.5 2.5 3.6 137 2 1 2 0 4 9.35 7.5 0 4.4 138 2 0 2 0 4 0 0 0 10 139 2 2 2 2 4 10 10 5 1.7 140 2 1 2 0 4 9.35 7.5 0 4.4 141 2 1 2 1 4 9.35 5 2.5 4.4 142 2 0 2 0 4 0 0 0 10 71

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 143 2 2 2 2 4 10 10 5 1.7 144 2 0 2 0 4 0 0 0 10 145 2 2 2 2 4 10 10 5 1.7 146 2 1 2 0 4 9.35 5 0 5.2 147 2 0 2 0 4 0 0 0 10 148 2 1 2 1 4 9.35 7.5 2.5 3.6 149 2 2 2 4 4 10 10 10 0 150 2 1 2 1 4 9.35 7.5 2.5 3.6 151 2 0 2 0 4 0 0 0 10 152 2 1 2 1 4 9.35 7.5 2.5 3.6 153 2 2 2 4 4 10 10 10 0 154 2 1 2 1 4 9.35 7.5 2.5 3.6 155 2 1 2 1 4 9.35 5 2.5 4.4 156 2 2 2 4 4 10 10 10 0 157 2 0 2 0 4 0 0 0 10 158 2 0 2 0 4 0 0 0 10 159 2 2 2 4 4 10 10 10 0 160 2 1 2 1 4 9.35 5 2.5 4.4 161 2 2 2 2 4 10 10 5 1.7 162 2 0 2 0 4 0 0 0 10 163 2 1 2 0 4 9.35 7.5 0 4.4 164 2 1 2 0 4 9.35 7.5 0 4.4 165 2 1 2 1 4 9.35 7.5 2.5 3.6 166 2 2 2 2 4 10 10 5 1.7 72

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 167 2 0 2 0 4 0 0 0 10 168 2 1 2 1 4 9.35 7.5 2.5 3.6 169 2 1 2 1 4 9.35 5 2.5 4.4 170 2 2 2 2 4 10 10 5 1.7 171 2 0 2 0 4 0 0 0 10 172 2 1 2 0 4 9.35 5 0 5.2 173 2 0 2 0 4 0 0 0 10 174 2 2 2 2 4 10 10 5 1.7 175 3 3 3 10 15 10 10 6.67 1.1 176 3 2 3 2 15 10 7.5 1.33 3.7 177 3 1 3 1 15 9.35 2.5.67 5.8 178 3 1 3 0 15 9.35 5 0 5.2 179 3 0 3 0 15 0 0 0 10 180 3 1 3 0 15 9.35 5 0 5.2 181 3 0 3 0 15 0 0 0 10 182 3 1 3 1 15 9.35 2.5.67 5.8 183 3 0 3 0 15 0 0 0 10 184 3 2 3 2 15 10 7.5 1.33 3.7 185 3 1 3 0 15 9.35 5 0 5.2 186 3 0 3 0 15 0 0 0 10 187 3 1 3 1 15 9.35 2.5.67 5.8 188 3 3 3 10 15 10 10 6.67 1.1 189 3 2 3 4 15 10 7.5 2.67 3.3 190 3 1 3 0 15 9.35 2.5 0 6.1 73

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 191 3 0 3 0 15 0 0 0 10 192 3 0 3 0 15 0 0 0 10 193 3 0 3 0 15 0 0 0 10 194 3 2 3 2 15 10 7.5 1.33 3.7 195 3 1 3 1 15 9.35 5.67 5.0 196 3 1 3 1 15 9.35 5.67 5.0 197 3 0 3 0 15 0 0 0 10 198 3 1 3 1 15 9.35 5.67 5.0 199 3 2 3 4 15 10 7.5 2.67 3.3 200 3 3 3 10 15 10 10 6.67 1.1 201 3 1 3 1 15 9.35 2.5.67 5.8 202 3 1 3 1 15 9.35 5.67 5.0 203 3 0 3 0 15 0 0 0 10 204 3 2 3 2 15 10 7.5 1.33 3.7 205 3 0 3 0 15 0 0 0 10 206 3 1 3 1 15 9.35 5.67 5.0 207 3 1 3 1 15 9.35 5.67 5.0 208 3 0 3 0 15 0 0 0 10 209 3 0 3 0 15 0 0 0 10 210 3 1 3 0 15 9.35 2.5 0 6.1 DMZ CONFIGURATIONS 211 2 1 1 1 2 9.35 2.5 5 4.4 212 3 1 1 1 5 9.35 2.5 2 5.4 213 3 1 1 1 5 9.35 2.5 2 5.4 74

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 214 3 1 1 1 5 9.35 2.5 2 5.4 215 4 1 1 1 16 9.35 2.5.625 5.8 216 2 1 1 0 2 9.35 7.5 0 4.4 217 3 1 1 0 5 9.35 5 0 5.2 218 3 1 1 0 5 9.35 2.5 0 6.1 219 3 1 1 0 5 9.35 5 0 5.2 220 4 1 1 0 16 9.35 2.5 0 6.1 221 2 1 1 0 2 9.35 7.5 0 4.4 222 3 1 1 0 5 9.35 2.5 0 6.1 223 3 1 1 0 5 9.35 5 0 5.2 224 3 1 1 0 5 9.35 5 0 5.2 225 4 1 1 0 16 9.35 2.5 0 6.1 226 3 1 2 1 5 9.35 2.5 2 5.4 227 3 2 2 2 5 10 5 4 3.7 228 2 0 2 0 2 0 0 0 10 229 3 1 2 1 5 9.35 2.5 2 5.4 230 3 1 2 0 5 9.35 5 0 5.2 231 3 0 2 0 5 0 0 0 10 232 2 2 2 2 2 10 10 10 0 233 3 1 2 0 5 9.35 5 0 5.2 234 3 1 2 1 5 9.35 2.5 2 5.4 235 3 0 2 0 5 0 0 0 10 236 3 2 2 2 5 10 7.5 4 2.8 237 3 0 2 0 5 0 0 0 10 75

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 238 3 2 2 2 5 10 7.5 4 2.8 239 3 1 2 0 5 9.35 2.5 0 6.1 240 2 0 2 0 2 0 0 0 10 241 3 1 2 1 5 9.35 2.5 2 5.4 242 3 2 2 2 5 10 5 4 3.7 243 3 1 2 1 5 9.35 2.5 2 5.4 244 3 0 2 0 5 0 0 0 10 245 3 1 2 0 5 9.35 5 0 5.2 246 2 2 2 2 2 10 10 10 0 247 3 1 2 0 5 9.35 5 0 5.2 248 3 1 2 0 5 9.35 2.5 0 6.1 249 3 2 2 2 5 10 7.5 4 2.8 250 3 0 2 0 5 0 0 0 10 251 3 0 2 0 5 0 0 0 10 252 3 2 2 2 5 10 7.5 4 2.8 253 3 1 2 1 5 9.35 2.5 2 5.4 254 3 2 2 0 5 10 7.5 0 4.2 255 3 0 2 0 5 0 0 0 10 256 2 1 2 0 2 9.35 7.5 0 4.4 257 3 1 2 0 5 9.35 5 0 5.2 258 2 1 2 0 2 9.35 7.5 0 4.4 259 3 2 2 0 5 10 7.5 0 4.2 260 3 0 2 0 5 0 0 0 10 261 3 1 2 0 5 9.35 5 0 5.2 76

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 262 3 1 2 0 5 9.35 5 0 5.2 263 3 2 2 0 5 10 7.5 0 4.2 264 3 0 2 0 5 0 0 0 10 265 3 1 2 0 5 9.35 5 0 5.2 266 3 0 2 0 5 0 0 0 10 267 3 2 2 0 5 10 7.5 0 4.2 268 3 3 3 5 5 10 10 10 0 269 3 2 3 2 5 10 7.5 4 2.8 270 3 1 3 1 5 9.35 2.5 2 5.4 271 4 1 3 0 16 9.35 2.5 0 6.1 272 4 0 3 0 16 0 0 0 10 273 3 1 3 0 5 9.35 5 0 5.2 274 3 0 3 0 5 0 0 0 10 275 4 1 3 0 16 9.35 2.5 0 6.1 276 4 0 3 0 16 0 0 0 10 277 4 2 3 0 16 10 5 0 5.2 278 3 1 3 0 5 9.35 5 0 5.2 279 4 0 3 0 16 0 0 0 10 280 3 1 3 1 5 9.35 2.5 2 5.4 281 3 3 3 5 5 10 10 10 0 282 3 2 3 2 5 10 7.5 4 2.8 283 4 1 3 0 16 9.35 2.5 0 6.1 284 3 0 3 0 5 0 0 0 10 285 4 0 3 0 16 0 0 0 10 77

Configuration hosts Vulnerable Hosts Operating Systems attack paths Network paths V Dimension E Dimension A Dimension VEAbility 286 4 0 3 0 16 0 0 0 10 287 4 2 3 0 16 10 5 0 5 288 3 1 3 0 5 9.35 5 0 5.2 289 3 1 3 0 5 9.35 5 0 5.2 290 4 0 3 0 16 0 0 0 10 291 4 1 3 0 16 9.35 5 0 5 292 4 2 3 2 16 10 5 1.25 4.6 293 4 3 3 5 16 10 7.5 3.125 3.1 294 3 1 3 0 5 9.35 2.5 0 6.1 295 4 1 3 1 16 9.35 2.5.625 5.8 296 3 0 3 0 5 0 0 0 10 297 4 2 3 2 16 10 5 1.25 4.6 298 3 0 3 0 5 0 0 0 10 299 4 1 3 1 16 9.35 2.5.625 5.8 300 4 1 3 1 16 9.35 2.5.625 5.8 301 3 0 3 0 5 0 0 0 10 302 3 0 3 0 5 0 0 0 10 303 3 1 3 0 5 9.35 2.5 0 6.1 78