FIREWALLS & NETWORK SECURITY with Intrusion Detection and VPNs, 2. Intrusion Detection and Prevention Systems

Transcription

1 FIREWALLS & NETWORK SECURITY with Intrusion Detection and VPNs, 2 nd ed. 13 Intrusion Detection and Prevention Systems By Whitman, Mattord, & Austin 2008 Course Technology

2 Learning Objectives Describe the various technologies that are used to implement intrusion i detection ti and prevention Define honey pots, honey nets, and padded cell systems Describe the technologies used to create honey pots, honey nets, and padded cell systems Slide 2

3 Intrusion Detection and Prevention Intrusion occurs when attacker attempts to gain entry or disrupt normal operations of information systems, almost always with intent to do harm Intrusion detection consists of procedures and systems that identify system intrusions Intrusion reaction encompasses actions an organization takes when intrusion is detected Intrusion prevention consists of activities that deter intrusion Slide 3

4 Intrusion Detection and Prevention (continued) Intrusion correction activities finalize restoration of operations to a normal state and seek to identify source and method of intrusion to ensure same type of attack cannot occur again Intrusion detection systems (IDSs) work like a burglar alarm: detect t violation, activate t alarm Intrusion prevention system (IPS) can detect intrusion and launch an active response IDS and IPS systems often coexist Intrusion detection/prevention ti ti system (IDPS) describes current anti-intrusion technologies Slide 4

5 IDPS Terminology Alert or alarm: indication a system has just been attacked or is under attack Evasion: process by which attacker changes the format and/or timing of their activities to avoid being detected by the IDPS False attack stimulus: event that triggers alarm when no actual attack is in progress False negative: failure of an IDPS to react to an actual attack event False positive: alert or alarm that t occurs in the absence of an actual attack Slide 5

6 IDPS Terminology (continued) Noise: accurate alarm events that do not pose significant threat to information security Site policy: rules and configuration guidelines governing implementation and operation of IDPSs within an organization Site policy awareness: IDPS s ability to dynamically modify its configuration in response to environmental activity True attack stimulus: event that triggers alarms and causes an IDPS to react as if a real attack is in progress Slide 6

7 IDPS Terminology (continued) Tuning: process of adjusting IDPS to maximize efficiency in detecting true positives, while minimizing false positives and false negatives Confidence value: value placed upon an IDPS s ability to detect/identify certain attacks correctly Alarm filtering: running system for a while to track types of false positives it generates and then adjusting IDPS alarm classifications Alarm clustering and compaction: process of grouping almost identical alarms occurring at almost same time into single higher-level alarm Slide 7

8 Why Use an IDPS? NIST reasons to acquire and use an IDPS: To prevent problem behaviors by increasing the perceived risk of discovery and punishment To detect attacks and other security violations not prevented by other security measures To detect t and deal with the preambles to attacks To document existing threat to an organization To act as quality control for security design and administration To provide useful information about intrusions that do take place Slide 8

9 Why Use an IDPS? (continued) IPS technologies can respond to detected threat by attempting to prevent it from succeeding while IDS cannot IDPS operational categories: Host-based (operates on the hosts themselves) Network-based (functions at the network level) Wireless Network behavior analysis (NBA) Slide 9

10 Why Use an IDPS? (continued) Several IPS response techniques: Terminate network connection or user session that is being used for the attack Block access to target from offending user account, IP address, or other attacker attribute Block all access to targeted t host, service, application, or other resource Change the security environment Change the attack s content Slide 10

11 Network-Based IDPS NIDPSs reside on computer or appliance connected to network segment and monitor network traffic Compare measured activity to known signatures to determine whether an attack has occurred or is underway Protocol stack verification: NIDPSs look for invalid data packets Application protocol verification: higher-order protocols (HTTP, FTP, Telnet) are examined for unexpected packet behavior or improper use Slide 11

12 Network-Based IDPS (continued) Some advantages of NIDPSs: Good network design and placement of devices can enable organization to use a few devices to monitor large network Usually passive devices and can be deployed into existing networks with little or no disruption to normal network operations Not usually susceptible to direct attack and may not be detectable by attackers Slide 12

13 Network-Based IDPS (continued) Some disadvantages of NIDPSs: Can become overwhelmed by network volume and fail to recognize attacks they might otherwise have detected Require access to all traffic to be monitored Cannot analyze encrypted packets, making some of the network traffic invisible to the process Cannot reliably ascertain if an attack was successful or not Some forms of attack are not easily discerned, specifically those involving fragmented packets Slide 13

14 Wireless NIDPS Monitors and analyzes wireless network traffic looking for potential problems with wireless protocols (Layers 2 and 3 of the OSI model) Cannot evaluate and diagnose issues with higher-layer protocols like TCP and UDP Some issues with implementation include: Physical security Sensor range Access point and wireless switch locations Wired network connections Cost Slide 14

15 Network Behavior Analysis System Examines network traffic to identify yproblems related to flow of traffic Uses a version of anomaly detection method Typical flow data relevant to intrusion detection and prevention includes: Source and destination IP addresses Source and destination TCP or UDP ports or ICMP types and codes Number of packets and bytes transmitted in the session Starting and ending timestamps for the session Slide 15

16 Network Behavior Analysis System (continued) Typically monitors internal networks; occasionally monitors internal/external t l network connections Most sensors, passive mode deployment only Types of events most commonly detected by NBA sensors include: Denial-of-service (DoS) attacks (including DDoS) Scanning Worms Unexpected application services Policy violations Slide 16

17 Host-Based IDPS Resides on particular computer or server (the host) and monitors activity only on that system Also known as system integrity verifiers Benchmark/monitor status of key system files Triggers alert when file attributes change, new files are created, or existing files are deleted Managed HIDPSs can monitor multiple computers simultaneously by creating a configuration file on each monitored host and by making each HIDPS report back to a master console system Slide 17

18 Host-Based IDPS (continued) Some advantages of HIDPSs: Can detect local events on host systems and also detect attacks that may elude NIDPSs Functions on host system, where encrypted traffic will have been decrypted and is available for processing Unaffected by use of switched network protocols Can detect inconsistencies in how applications and systems programs were used by examining records stored in audit logs, enabling it to detect some types of attacks, including Trojan Horse programs Slide 18

19 Host-Based IDPS (continued) Some disadvantages of HIDPSs: Pose more management issues since they are configured/managed on each monitored host Vulnerable to direct attacks, attacks on host OS Not optimized to detect multi-host scanning; unable to detect t scanning of non-host devices Susceptible to some denial-of-service attacks Can use large amounts of disk space to retain the host OS audit logs Inflicted overhead on host systems may reduce system performance below acceptable levels Slide 19

20 IDPS Detection Methods Signature-based (knowledge-based, misuse- detection) IDPS: examines network traffic in search of patterns that match known signatures Statistical anomaly-based (stat, behavior-based) IDPS: compares sampled network activity to established baseline Stateful protocol analysis (SPA) IDPS: uses profiles to detect anomalous protocol behavior Log file monitor (LFM) IDPS: reviews log files from servers, network devices, and other IDPSs for signatures indicating an attack or intrusion Slide 20

21 IDPS Response Behavior Response depends on organization s policy, objectives, and system capabilities Responses classified as active or passive Active response: definitive action automatically initiated when certain types of alerts are triggered; can include collecting additional data, changing or modifying the environment, and taking action against the intruders Passive response: report information they have collected and wait for administrator to act Slide 21

22 IDPS Response Behavior (continued) Some possible responses IDPSs can produce: Audible/visual alarm SNMP traps and plug-ins message Page or phone message Log entry Evidentiary packet dump Take action against the intruder Launch program Reconfigure firewall Terminate session or connection Slide 22

23 Selecting IDPS Approaches and Products Technical and policy considerations What is your system s environment? What are your security goals and objectives? What is your existing security policy? Organizational requirements and constraints What requirements are levied from outside the organization? What are your organization s resource constraints? Slide 23

24 Selecting IDPS Approaches and Products (continued) IDPSs product features and quality Is the product sufficiently scalable for your environment? How has the product been tested? What is the user level of expertise targeted by the product? Is the product designed to evolve as the organization grows? What are the support provisions for the product? Slide 24

25 Strengths and Limitations of IDPSs IDPSs perform the following functions well: Monitoring and analysis of system events and user behaviors Testing security states of system configurations Baselining security state of system and then tracking any changes to that t baseline Recognizing patterns of system events that correspond to known attacks Recognizing patterns of activity that statistically vary from normal activity Slide 25

26 Strengths and Limitations of IDPSs (continued) More functions that IDPSs perform well: Managing operating system audit and logging mechanisms and the data they generate Alerting appropriate staff by appropriate means when attacks are detected Measuring enforcement of security policies i encoded in the analysis engine Providing default information security policies Allowing non-security experts to perform important security monitoring functions Slide 26

27 Strengths and Limitations of IDPSs (continued) IDPSs cannot perform the following functions: Compensating for weak or missing security mechanisms in the protection infrastructure Instantaneously detecting, reporting, responding to attack during heavy network/processing load Detecting ti newly published attacks or variants Effectively responding to sophisticated attacks Automatically ti investigating attacks Resisting all attacks intended to defeat them Compensating for fidelity issues of data sources Dealing effectively with switched networks Slide 27

28 Deployment and Implementation of an IDPS IDPS control strategies Centralized: all IDPS control functions are implemented and managed in a central location Fully distributed: all control functions are applied at the physical location of each IDPS component Partially distributed: ib t d combines the best of the other two strategies; while individual agents still analyze and respond to local threats, their reporting to a hierarchical central facility enables the organization to detect widespread attacks Slide 28

29 Deployment and Implementation of an IDPS (continued) IDPS deployment Great care must be made in deciding where to locate IDPS components, physically and logically During deployment, each component should be installed, configured, fine-tuned, tested, and monitored NIDPS and HIDPS used in tandem can protect individual systems and organizational networks Use a phased implementation strategy so as not to affect entire organization all at once First implement NIDPSs and then install HIDPSs Slide 29

30 Deployment and Implementation of an IDPS (continued) Deploying network-based IDPSs NIST recommends four locations for NIDPS sensors: Behind each external firewall, in the network DMZ Outside an external firewall On major network backbones On critical subnets Slide 30

31 Deployment and Implementation of an IDPS (continued) Deploying host-based IDPSs Proper implementation of HIDPSs can be a painstaking and time-consuming task, as each HIDPS must be custom configured to its host May be beneficial to practice an implementation on one or more test servers beforehand Installation continues until either all systems are installed or organization reaches the planned degree of coverage it is willing to live with Slide 31

32 Measuring the Effectiveness of IDPSs When selecting an IDPS, one typically y looks at four measures of comparative effectiveness: Thresholds Blacklists and whitelists Alert settings Code viewing and editing Slide 32

33 Measuring the Effectiveness of IDPSs (continued) Once implemented, IDPSs are evaluated using two dominant metrics: Administrators evaluate the number of attacks detected in a known collection of probes Administrators examine the level of use, commonly measured in megabits per second of network traffic, at which the IDPSs fail In order to truly assess effectiveness of IDPS systems, test process should be as realistic as possible in its simulation of actual event Couple realistic traffic loads, levels of attacks Slide 33

34 Honey Pots, Honey Nets, and Padded Cell Systems Honey yp pots (decoys, lures, fly-traps): decoy systems designed to lure potential attackers away from critical systems Honey net: collection of honey pots connecting several honey pot systems on a subnet Honey pots are designed to: Divert an attacker from critical systems Collect information about the attacker s activity Encourage the attacker to stay on the system long enough for administrators to document the event and, perhaps, respond Slide 34

35 Honey Pots, Honey Nets, and Padded Cell Systems (continued) Padded cell: honey ypot that has been protected so it cannot be easily compromised in other words, a hardened honey pot In addition to attracting attackers with tempting data, padded cell operates in tandem with traditional IDPS When IDPS detects attackers, it seamlessly transfers them to special simulated environment where they can cause no harm Allows organization to observe and document actions and tactics of an attacker Slide 35

36 Honey Pots, Honey Nets, and Padded Cell Systems (continued) Advantages of using honey pot or padded cell: Attackers can be diverted to targets that they cannot damage Administrators have time to decide how to respond to an attacker Attackers actions can be easily and more extensively monitored, and the records can be used to refine threat models and improve system protections Honey pots may be effective at catching insiders who are snooping around a network Slide 36

37 Honey Pots, Honey Nets, and Padded Cell Systems (continued) Disadvantages of using honey pot or padded cell: The legal implications of using such devices are not well defined Honey pots and padded cells have not yet been proven as generally useful security technologies An expert attacker, once diverted into a decoy system, may become angry and launch a more hostile attack against an organization s systems Administrators and security managers need a high level of expertise to use these systems Slide 37

38 Trap and Trace Systems Use a combination of techniques to detect an intrusion and then to trace it back to its source Trap usually consists of a honey pot or padded cell and an alarm Trace feature is process by which organization attempts to determine identity of an intruder Slide 38

39 Trap and Trace Systems (continued) If intruder is someone inside the organization, administrators are within their power to track the individual and turn him or her over to authorities If intruder is outside security perimeter of the organization, numerous legal issues arise Back hack: hacking into a hacker s system to find out as much as possible about the hacker Enticement or entrapment? Slide 39

40 Active Intrusion Prevention Some organizations do more than wait for an attack and implement active countermeasures When attacker sends ARP request to unused IP address, LaBrea pretends to be a computer at that address, allowing attacker to connect Once connected, LaBrea changes TCP sliding window size to a low number to hold open the connection from the attacker This greatly slows down network-based worms and other attacks and gives LaBrea system time to notify system and network administrators Slide 40

41 Chapter Summary Intrusion occurs when attacker attempts to gain entry or disrupt normal operations of information system, almost always with intent to do harm Intrusion detection consists of procedures and systems that identify system intrusions Intrusion reaction encompasses actions an organization takes when intrusion is detected Intrusion prevention consists of activities that deter an intrusion Slide 41

42 Chapter Summary (continued) Intrusion detection system (IDS) works like a burglar alarm: detects violation, activates alarm Intrusion prevention system (IPS) can prevent intrusion from successfully attacking the organization by means of some active response Because these systems often coexist, term intrusion detection/prevention system (IDPS) is used to describe current anti-intrusion intrusion technologies Slide 42

43 Chapter Summary (continued) IDPSs commonly operate as either network- or host-based systems Network-based IDPS functions at network level Host-based IDPS operates on hosts themselves Systems that use both approaches are called hybrid IDPSs Slide 43

44 Chapter Summary (continued) IDPSs use variety of detection methods to monitor and evaluate network traffic Three methods dominate: signature-based approach, statistical-anomaly approach, stateful protocol analysis approach Log file monitor (LFM) IDPS is similar to NIDPS Using LFM, system reviews log files generated by servers, network devices, and other IDPSs, looking for patterns and signatures that may indicate an attack or intrusion is in progress or has already occurred Slide 44

45 Chapter Summary (continued) Honey ypots: decoy systems designed to lure potential attackers away from critical systems Honey net: collection of honey pots connecting several honey pot systems on a subnet A honey pot is configured in ways that make it look vulnerable to lure potential attackers into attacking, thereby revealing themselves Trap and trace applications use a combination of techniques to detect intrusion and then trace it back to its source Slide 45