Intrusion Detection, Access Control & Other Security Tools Principles of Information Security Chapter 7 Part 1 References NIST Intrusion Detection Systems ◦ http://csrc.nist.gov/publications/nistpubs/800-31/sp80031.pdf#search=%22NIST%20SP%20800-31%22 http://en.wikipedia.org/wiki/Intrusion-detection_system http://netsecurity.about.com/cs/hackertools/a/aa030504.htm http://sectools.org/ids.html http://www.softpanorama.org/Security/integrity_checkers.shtml http://www.sans.org/resources/idfaq/index.php http://www.sans.org/resources/idfaq/integrity_checker.php http://en.wikipedia.org/wiki/Host_based_intrusion_detection_system http://sourceforge.net/projects/tripwire/ http://sourceforge.net/projects/integrit/ http://freshmeat.net/projects/integrit/ 2 Topic Objectives Upon completion of this chapter you should be able to: ◦ Identify and describe the categories and operating models of intrusion detection systems. ◦ Identify and describe honey pots, honey nets, and padded cell systems ◦ List and define the major categories of scanning and analysis tools ◦ Describe the tools used in each major category of scanning and analysis tools ◦ Discuss various approaches to access control, including the use of biometric access mechanisms 3 Role of Technology in Information Security Technical solutions, properly implemented, can enhance the confidentiality, integrity and availability of an organization's information assets. 4 Intrusion An attempt to gain entry into a system or disrupt normal operations of a system Usually, but not always, with malicious intent 5 Incident Response Identification of, classification of, response to, and recovery from an incident (intrusion). Includes ◦ Intrusion prevention Efforts to deter intrusions via policy, security programs, technology based countermeasures ◦ Intrusion detection Use of procedures and systems created to detect intrusions ◦ Intrusion reaction Actions taken when an intrusion is detected including limiting losses and initiating procedures to return to normal operations ◦ Intrusion correction Finalizing restoration to a normal state, identifying source and methods used for intrusion, and taking action to prevent the same type of intrusion from occurring again. 6 Intrusion Detection Systems (IDSs) Similar to burglar alarms Detect violations of configuration and activates an alarm ◦ audible (noise) ◦ visual (lights, etc.) ◦ silent (email msg, pager alert) NIST SP 800-31 Intrusion Detection Systems ◦ http://csrc.nist.gov/publications/nistpubs/800-31/sp80031.pdf#search=%22NIST%20SP%20800-31%22 7 IDS is not a firewall Firewall is software/hardware that prevents access ◦ Firewall examines each packet, checks if the packet satisfies the “rules” for acceptance, and either drops it or lets it in ◦ Like a guard at the gate – you must show the guard the ID to get in, or you are turned back Firewall rules are not all that sophisticated, they are the most basic and most crude level of sorting packets. ACL tables should have 30-50 rules, typically 8 Add the camera to the security guard IDS is software that monitors the network quietly, logs activities, and alerts if it sees anything “wrong.” ◦ Like a burglar alarm, or an alarmed security camera at the gate. ◦ IDS will have rather sophisticated rules because it has signatures of many different types of attack, e.g. SYN flood, etc. “Secure perimeters” will have both an IDS and firewall 9 Key Terminology for IDSs Alert or alarm ◦ indication that an attack/intrusion has been detected and may be in progress False Negative ◦ a failure to detect an actual attack ◦ a major problem for an IDS False Positive ◦ a false indication of an attack occurring ◦ can lead users to become insensitve to alarms Noise ◦ activity that is not particularly significant, such as unsuccessful intrusion attempts 10 Key Terminology (2) Site Policy ◦ organizational rules and guidelines with respect to use of IDSs Site Policy Awareness ◦ ability of an IDS to dynamically modify site policies ◦ SMART IDSs can adapt reaction to activity Confidence Value ◦ measure of ability of IDS to detect and identify an attack correctly Alarm Filtering ◦ similar to packet filters --- can filter alarms based on source/dest IP addresses, OS, confidence value, alarm type or severity ◦ used to filter out known false positives from actual attacks 11 Reasons to Use an IDS Deterrence ◦ deter attacks by increasing perceived risk of discovery and punishment Detection ◦ detect attacks and violations not prevent by other means such as scanning tools or vulnerability assessments ◦ detect preambles (precursors) to attacks such as footprinting or fingerprinting activities Documentation ◦ document existing threats to organization for justification of defensive measures and tools Quality Assurance ◦ Verify adequacy and improve security design and administration Information ◦ Determine causes and methods used in attacks to prevent the same attack strategy being used again. ◦ Obtain forensic evidence for legal action. 12 Types of IDSs Network-based (NIDS) ◦ Monitor network traffic to look for recognizable attack patterns Host-based (HIDS) ◦ Focus on server or host assets, such as files Application-based ◦ Focus on one application, either on one or more hosts/servers. ◦ E.g., distributed database system 13 Detection Methods Signature-based ◦ Recognize characteristics of a particular type of attack Statistical anomaly-based ◦ Recognize anomalies in operation or performance that may signal an attack. 14 **Network-Based IDS (NIDS) Monitors traffic on a network segment ◦ Sends alerts when an attack pattern is recognized. ◦ May use the monitoring port or switched analysis port (SPAN) of a hub or switch Specially configured connection capable of viewing all traffic traversing the device Necessary for switches in particular Snort is an example of a NIDS 15 NIDS Signature Matching Must compare observed activity to known attack patterns ◦ Uses special implementation of TCP/IP stack to reassemble packets for use with verification and comparison techniques Protocol stack verification ◦ Look for invalid (malformed)TCP/IP data packets ◦ Exploits of packet structure ◦ Example: DoS and DDoS frequently use malformed packets to exploit vulnerabilities in the protocal Application protocol verification ◦ Application layer protocols (HTTP, FTP, etc.) are examined for unexpected behavior, or improper use ◦ Exploits of protocol use ◦ Example: DNS cache poisoning ◦ Can significantly affect system throughput. May be necessary to use separate systems for protocol stack verification and application protocol verification 16 **Pros & Cons of NIDS Pros ◦ Conservation of equipment through strategic design and placement ◦ Little or no disruption to normal network operations because of passive monitoring ◦ Not usually susceptible to attack and may not be detectable by attackers Cons ◦ Easily overwhelmed by network traffic -> fail to recognize attacks ◦ Must be able to monitor all network traffic. May not be possible if switches to not have adequate monitoring ports. ◦ Cannot analyze encrypted packets (e.g., ssh) ◦ Cannot reliably determine whether an attack was successful ◦ Cannot recognized all forms of attacks. E.g., malformed packets 17 **Host-Based IDS (HIDS) Monitors activity only on a single server or host AKA, system integrity verifiers ◦ benchmark and monitor status of key system files and detect changes or deletions to these files. ◦ capable of monitoring system configuration files and databases such as Windows registry, .ini, .cfg and .dat files. ◦ based on change or configuration management record sizes, locations, attributes of system files triggers alert when changes occur ◦ monitor system logs for predefined events Very reliable. Integrit and Tripwire are examples of HIDS 18 **Pros & Cons of HIDS Pros ◦ Can detect events missed by NIDS ◦ Files on host system decrypted and available to HIDS for analysis (compared to NIDS) ◦ Not affected by switched networking ◦ Able to detect inconsistencies in use of applications and system programs Cons ◦ Administration overhead because of host-based configuration ◦ Vulnerable to direct attacks and attacks against the OS ◦ Not optimized for multi-host or multi-device scanning ◦ Susceptible to some DoS attacks ◦ Can require large amounts of disk space ◦ Can impose a performance penalty on host systems 19 **Application-Based IDS (AppIDS) Variation of HIDS Examines one application for abnormal events ◦ users exceeding authorization, invalid file executions, etc Pros ◦ Able to track users (accountability) ◦ Ability to view encrypted data while unencrypted Cons ◦ May be more susceptible to attack because of not being as well protected as network and host based systems ◦ Less capable of detecting software tampering. Should be used in conjunction with NIDS and HIDS 20 Signature-Based IDS AKA, knowledge-based IDS looks for traffic patterns that match known signature --attack patterns ◦ footprinting and fingerprinting activities ◦ exploits that use a known sequence of activities Cons ◦ Signatures must continually be updated as new attack strategies develop ◦ May not detect attacks that are purposely slow and methodical --- e.g., sneaky or paranoid attacks 21 **Statistical Anomaly-Based IDS (stat IDS) AKA, behavior-based IDS based on frequency with which certain network activities occur Establish baseline performance by collecting statisical summaries of normal network traffic Periodically sample network traffic and use statistical methods to compare sampled activity with the baseline. ◦ Measured activity outside baseline parameters - exceeds the clipping level & triggers alert Pro ◦ IDS can detect new types of attacks Cons ◦ Require more overhead and processing capability Must constantly compare patterns of activity against baseline ◦ May not detect minor changes to system variables and may generate false positives ◦ May not be suitable for networks with wide variations in usage and traffic patterns 22 **Log File Monitors (LFM) Similar to NIDS Look for patterns and signatures in log files, including system log files, network log files, IDS log files. Able to examine log files from multiple systems Resource intensive to support collection, movement, storage and analysis of large quantities of data 23 IDS Response Be careful that the response does not make the problem worse, e.g. ◦ shut down network connection (leads to loss of business, i.e. “we attacked ourselves successfully” ◦ go after/attack the attackers (it is illegal!) ◦ make a huge negative advertisement (e.g. reporters, police, etc.) Active response ◦ A clear action is initiated when certain types of alerts are triggered --- data collection, modify environment, block intruders Passive response ◦ Report information and wait for administrator to take some action - administrator becomes the active component 24 Examples of IDS Response Options Audible / visual alarm SNMP traps and plug-ins ◦ send message to SNMP management console E-mail message ◦ Problem - intruder may block email system Page or phone message Log entry ◦ IDS system log file on separate device Evidentiary packet dump ◦ Allow further analysis on data for possible legal use Action against intruder ◦ traceback to identify source of attack ◦ ill-advised may not be legal Launch program, reconfigure firewall Terminate connection or session 25 **Some Limitations of IDS Cannot compensate for weaknesses or holes in security mechanisms ◦ firewalls, identification and authentication methods, link encryption, etc. Do not respond instantaneously in heavy network load scenarios Cannot detect new attacks or variations of existing attacks Cannot automatically investigate attacks Do not work effectively in switched networks 26 27 **Honey Pots, Honey Nets, and Padded Cell Systems Powerful security tools that go beyond intrusion detection Honey Pots or Honey Pot Systems ◦ Decoy systems designed to lure potential attackers away from critical systems and encourage attacks against the decoy ◦ Designed to divert attackers from critical systems, collect information about their activity, and encourage attackers to stay on the system long enough to be documented. Honey Net ◦ A collection of several honey pot systems on one subnet. ◦ Contains pseudo-devices configured in ways to make it look vulnerable. Padded Cell ◦ A honey pot that has been protected so it cannot be easily compromised --a hardened honey pot. ◦ When IDS detects attackers, it transfers them to a specialized environment where they can do no harm. Use of these techniques may be construed as illegal. 28 **Pros & Cons of Honey Pot Systems Pros ◦ Attackers diverted to targets where they can do no damage ◦ Allows time to decide how to respond to an attacker ◦ Attacks can be monitored to obtain useful information that can be used to develop better defensive methods ◦ May be effective for identifying insiders who are snooping around the network Cons ◦ Legal implications not well defined ◦ Not yet shown to be generally useful technologies ◦ May provoke expert hackers to more hostile attacks ◦ High level of expertise required to use these types of systems 29 **Trap and Trace Systems Trace incidents back to source similar to caller ID ◦ If originator is inside the organization, individual can be dealt with internally or externally. ◦ If originator is outside the organization, numerous legal issues must be addressed back-hacking --- reverse hacks ◦ deemed unethical by most codes of professional conduct ◦ vigilante justice Enticement vs Entrapment ◦ Enticement attract attention to a system by placing tantalizing bits of information in key locations legal and ethical ◦ Entrapment lure an individual into committing a crime to get a conviction not legal and not ethical 30 Active Intrusion Detection LaBrea (in “real life” this is the name of a famous ancient large tar (i.e. asphalt) pit in LA full of dinosaurs bones, because they would get stuck in tar and die http://www.tarpits.org/education/guide/geology/entrap.html) ◦ Uses the unused IP address space on a network ◦ Monitors ARP requests for invalid destination IP addresses Responds as if it were the target device Completes TCP 3-way handshake, then reduces sliding window to very small size to slow down data connection. Allows time for network administrator response 31 **Access Control Mechanisms Process for controlling access to protected systems ◦ User (supplicant) seeks to use a protected system ◦ User must be identified and authenticated before being authorized to use the protected system **Identification ◦ Provide an identity (username) **Authentication ◦ Provide some form of evidence that the supplicant is in fact the stated user Password, passphrase, biometric, key, etc. **Authorization ◦ Allow the user access to authorized systems, once the identification and authentication requirements have been met. ◦ Domain based (single-sign on) systems frequently issue a "token" that is then used on other devices in the domain 32 **Authentication Process that validates the user's identity **4 general methods for authentication ◦ Knowledge something the supplicant KNOWS password, passphrase, personal identification code ◦ Possession something the supplicant HAS mag stripe cards, smart cards, tokens synchronous tokens - generate one-time password used to complete authentication asynchronous tokens - challenge response mechanisms ◦ Unique Characteristics who the supplicant IS biometrics ◦ Distinguishable Characteristics what the supplicant PRODUCES signature recognition, voice recognition 33 **Biometrics Process of using unique body characteristics of the supplicant for authentication Includes ◦ Fingerprint comparisons ◦ Palm print comparisons ◦ Hand geometry comparisons ◦ Facial recognition using photo ID ◦ Retinal print comparison ◦ Iris pattern comparison Only 3 bodily characteristics considered truly unique ◦ Fingerprints ◦ Retina (blood vessel pattern) ◦ Iris 34 **Effectiveness of Biometrics **Criteria for evaluation ◦ false rejection rate % of authorized supplicants who are falsely rejected Type I error nuisance to authorized users, but no threat to security ◦ false acceptance rate % of unauthorized supplicants who are falsely accepted Type II error clear breach of security, unacceptable to security professionals ◦ crossover error rate (CER) level at which false rejections = false acceptances most common measure of accuracy for biometric systems indicator of optimal balance between false rejection and false acceptance rates. 1% CER is superior to a 5% CER 35 **Effectiveness and Acceptance of Biometrics Many highly effective and reliable biometric systems are considered intrusive to users. Preference for using systems that are effective and acceptable to users. Table 7-3 Effectiveness vs Acceptance Effectiveness Most to Least Secure Acceptance Most to Least Accepted Retinal scans Keystroke patterns Fingerprint Signature recognition Handprint Voice pattern Voice pattern Handprint Keystroke patterns Fingerprint Signature Retinal scans 36