CYS406 · Chapter 12

IDS and
IPS

The network's alarm system versus its guard dog — how intrusion detection and prevention systems watch traffic, spot attacks, classify themselves six different ways, and decide when to shout versus when to slam the door.

IDS vs IPS Signature vs Anomaly NIDS vs HIDS IDS Classification Deployment Locations False Positive/Negative
scroll ↓
01 / Foundations

What Is an IDS/IPS?

An intrusion detection and prevention system (IDS/IPS) is a network security appliance that inspects all inbound and outbound network traffic for suspicious patterns that might indicate a network or system security breach. If it finds something, the IDS alerts the administrator about the suspicious activity — it checks network traffic for signatures that match known intrusion patterns and triggers an alarm when a match is found.

🚨 IDS — Detects

An Intrusion Detection System sits passively, watches copies of traffic, and raises an alarm when it spots a match. It does not sit inline in the path of traffic — it observes and reports.

🛡️ IPS — Prevents

An Intrusion Prevention System sits inline (in the direct path of traffic, alongside/near the firewall) so it can actively block or drop malicious packets in real time, not just alert about them.

🧠

Mnemonic — Alarm vs Guard Dog

"IDS barks, IPS bites." An IDS is a smoke alarm — it detects and dings a bell but doesn't put out the fire. An IPS is a guard dog standing in the doorway — it can actually stop the intruder from getting through. Detect vs Detect+Block.

How an IDS Works — The Pipeline

Traffic flows from the Internet, through the router and firewall, into the IDS. The IDS preprocessor compares the traffic against a signature file database. If there's no signature match, it runs anomaly detection, then stateful protocol analysis. Any match triggers an action rule.

Internet
Router
Firewall
IDS
Signature / Anomaly / Protocol Checks
Alarm + Action

When a match is found: the alarm notifies the admin and the packet can be dropped, connections from that IP source are cut, and the packet is dropped.

Role of an IDS in Network Defense

An IDS works from inside the network, unlike a firewall, which only looks outside the network for intrusions.

An IDS is placed behind the firewall, inspecting all the traffic, looking for heuristics and pattern matches for intrusions.

💡

Firewall vs IDS — different jobs

A firewall enforces a rule set at the perimeter and blocks based on ports/addresses/protocols. An IDS goes deeper — it inspects traffic content for attack patterns, from a vantage point behind that firewall, so it can catch things the firewall's rules let through.

02 / Detection Methods

How an IDS Detects an Intrusion

An IDS uses three core detection techniques, and these are also the "Approach-based" classification of IDS covered later in the chapter.

Signature Recognition
Also known as misuse detection. Tries to identify events that indicate abuse of a system or network resource by matching known attack patterns.
Anomaly Detection
Detects intrusion based on the fixed behavioral characteristics of the users and components in a computer system — deviation from a learned baseline.
Protocol Anomaly Detection
Models are built to explore anomalies in the way vendors deploy the TCP/IP specification.

Signature-Based Detection — Deep Dive

Also known as misuse detection. It monitors patterns of data packets in the network and compares them to pre-configured network attack patterns, known as signatures. This method uses string comparison operations to compare ongoing activity — such as a packet or a log entry — against a list of signatures.

✅ Advantages
  • Detects attacks with minimal false alarms
  • Can quickly identify the use of a specific tool or technique
  • Helps admins quickly track potential security issues and start incident handling
⚠️ Disadvantages
  • Only detects known threats — the database must be constantly updated with new attack signatures
  • Uses tightly defined signatures, which prevents it from detecting common variants of attacks

Anomaly-Based Detection — Deep Dive

Alarms for anomalous activities are generated by evaluating network patterns — bandwidth used, protocols used, ports/devices connected to each other. The IDS monitors typical activity for a time interval and builds statistics for the network traffic. Example: an anomaly-based IDS monitors normal Internet bandwidth usage, failed logon attempts, processor utilization levels, etc.

✅ Advantages
  • Identifies abnormal behavior and detects symptoms of attacks without clear details in advance
  • Information from anomaly detectors can be used later to define signatures for misuse detectors
⚠️ Disadvantages
  • High rate of false alarms due to unpredictable behavior of users and networks
  • Needs an extensive set of system events to properly characterize "normal" behavior patterns

Stateful Protocol Analysis

Compares observed events with predetermined profiles based on accepted definitions of benign activity for each protocol, to identify deviations of the protocol state. It can identify unpredictable sequences of commands (e.g., the same command issued repeatedly, or arbitrary commands). It also detects variations in command length, minimum/maximum values for attributes, and other potential anomalies. For any protocol performing authentication, the IDS/IPS tracks the authenticator used for each session and records the authenticator involved in suspicious activity.

🧠

Mnemonic — "Known List vs Weird Behavior vs Broken Rules"

"Signature = mugshot, Anomaly = 'that's not like you', Protocol = rulebook check." Signature/misuse detection matches a known "mugshot" (attack pattern) — fast, low false positives, but blind to new attacks. Anomaly detection notices behavior that deviates from a learned baseline — catches unknowns, but trigger-happy (high false positives). Stateful protocol analysis checks whether the conversation is following the protocol's own rulebook (e.g., login sequence, command order).

03 / Classification

IDS Classification — Six Axes

An IDS is classified based on six different criteria: approach, protected system, structure, data source, behavior after an attack, and analysis timing.

🧠

Mnemonic — "A Pretty Strange Dog Barks, Timely"

Approach, Protected system, Structure, Data source, Behavior after an attack, Timing — the six ways an IDS gets labeled. "A Pretty Strange Dog Barks, Timely" hits all six first letters in slide order.

1. Approach

Signature Detection vs Anomaly Detection (covered in depth above).

2. Protected System

HIDS (host), NIDS (network), or Hybrid (both).

3. Structure

Centralized System vs Distributed System (with an Agent System).

4. Data Source

Audit Trail vs Network Packets vs System State Analysis.

5. Behavior After an Attack

Active IDS vs Passive IDS.

6. Analysis Timing

On-the-fly Processing (real-time) vs Interval-based IDS.

Protection-Based IDS — NIDS vs HIDS vs Hybrid

An IDS is classified based on the system/network it offers protection to.

🌐 NIDS — Network IDS

If it protects the network, it's a network intrusion detection system (NIDS). It sits on the network, watching traffic destined for many hosts.

💻 HIDS — Host IDS

If it protects a host, it's a host intrusion detection system (HIDS). It's installed on an individual machine and watches only that machine's activity.

🔀 Hybrid IDS

If it protects both the network and a host, it's a hybrid intrusion detection system. A hybrid IDS combines the advantages of both the low false-positive rate of a NIDS and the anomaly-based detection of a HIDS to detect unknown attacks.

🧠

Mnemonic — "N is for Neighborhood, H is for House"

NIDS watches the Neighborhood (the whole network); HIDS watches one House (one host). One NIDS sensor can watch traffic to/from many HIDS-protected machines at once — that's why the slide shows a single NIDS box connected to five HIDS boxes.

Structure-Based IDS — Centralized vs Distributed

This classification is based on the physical/logical structure of the IDS deployment.

🏢 Centralized IDS

All data is shipped to a central location for analysis, independent of the number of hosts monitored. One IDS console analyzes everything (Network, Application, and Host-based monitoring systems all report in).

🕸️ Distributed IDS

Several IDS instances are deployed over a large network and each IDS communicates with each other for traffic analysis — a fully distributed, agent-based control model.

Behavior-Based IDS — Active vs Passive

An IDS is categorized based on how it reacts to a potential intrusion. It functions in one of two modes, based on the behavior after an attack.

⚡ Active IDS

Detects and responds to detected intrusions — it can listen and monitor, then take an active response (e.g., cut a connection).

👁️ Passive IDS

Only detects intrusions — it listens and monitors, and that's it. No automatic response.

Analysis Timing–Based IDS — Interval vs Real-Time

Analysis time is the span of time elapsed between events occurring and the analysis of those events.

⏳ Interval-Based IDS

Information about an intrusion detection does not flow continuously from monitoring points to analysis engines — it's simply stored and forwarded. Performs analysis offline.

⏱️ Real-Time–Based IDS

Information flows continuously from monitoring points to analysis engines. Performs analysis on the fly.

Source Data Analysis–Based IDS

An IDS is classified based on the type of data source it uses to detect intrusions — audit trails or network packets.

📜 Audit Trails

Help the IDS detect performance problems, security violations, and flaws in applications.

📡 Network Packets

Capturing and analyzing network packets helps an IDS detect well-known attacks.

📋

Misuse vs Anomaly Detection System architecture

Both system types share the same basic pipeline: Auditing Modules → Profiles → an inference engine → Target Systems. The difference is what the engine compares against — a Misuse Detection System runs an Interference (Inference) Engine against known-bad profiles, while an Anomaly Detection System runs a dedicated Anomaly Detection Engine against learned-normal profiles.

04 / Capabilities & Limits

IDS/IPS Capabilities, Limitations & Warning Signs

IDS Capabilities

IDS provides an additional layer of security to the network under the defense-in-depth principle. It does several things basic firewalls cannot do, and helps minimize the chance of missing security threats that could come from firewall evasions.

Monitoring and analyzing both user and system activities

Recognizing typical attack patterns

Analyzing system configurations and vulnerabilities

Analyzing abnormal activity patterns

Assessing system and file integrity

Tracking user policy violations

IDS/IPS Limitations — What It Is NOT

IDS/IPS cannot act as, or replace:

Network Logging Systems

Vulnerability Assessment Tools

Antivirus Products

Cryptographic Systems

🧠

Mnemonic — "LVAC" (Not a LAVAC)

"IDS is NOT a LVAC"Logging system, Vulnerability scanner, Antivirus, Crypto system. Four things students often mistakenly assume an IDS covers — it doesn't replace any of them.

IDS/IPS Security Concerns

Improper IDS/IPS configuration and management will make an IDS/IPS ineffective. Deployment should be done with careful planning, preparation, prototyping, testing, and specialized training.

Common Mistakes in IDS/IPS Configuration

General Indications of Intrusions

File System Intrusions

  • New or unfamiliar files/programs
  • Changes in file permissions
  • Unexplained changes in a file's size
  • Rogue files not on the master signed-file list
  • Missing files

Network Intrusions

  • Repeated probes of available services
  • Connections from unusual locations
  • Repeated login attempts from remote hosts
  • A sudden influx of log data

System Intrusions

  • Short or incomplete logs
  • Unusually slow system performance
  • Missing logs / incorrect permissions or ownership
  • Modifications to system software and config files
  • Unusual graphic displays or text messages
  • Gaps in system accounting
  • System crashes or reboots
  • Unfamiliar processes
05 / Components

IDS Components

An IDS system is built on various components. Knowledge of their functions and placement is required for effective IDS implementation.

1

Network Sensors

Hardware and software components that monitor network traffic and trigger alarms if abnormal activity is detected. Should be placed at common entry points: Internet gateways, between LAN connections, remote-access/dial-up servers, VPN devices, and on either side of a firewall.

2

Command Console

Installed and runs on a separate, dedicated system. Provides a UI for the admin to receive and analyze security events, alert messages, and log files, and evaluates security event information from different security devices (e.g., Sguil). Caution: if installed on a non-dedicated computer (firewall, backup server), it drastically slows the response to security events since those systems are busy with other tasks.

3

Alert Systems

Sends an alert message when any anomaly or misuse is detected (e.g., OSSEC HIDS alerts, Snort NIDS alerts — both shown in the Sguil console).

4

Response System

Issues countermeasures against any detected intrusion. You must also stay involved in the decision during incident response and be able to respond on your own — you need to decide how to handle false positives and when a response needs escalation. Recommendation: do not solely rely on an IDS response system for intrusion response.

5

Attack Signature Database

An IDS itself cannot "decide" — instead it maintains a database of attack signatures/patterns. Network traffic is compared against these signatures, and a decision can be made. If a match is found, the IDS raises an alert and blocks the suspicious traffic. Recommendation: periodically update the IDS attack signature database.

🧠

Mnemonic — "Sensors See, Console Controls, Alerts Announce, Response Reacts, Database Decides-with"

"Nick's Cat Always Runs Around"Network sensors → Command console → Alert systems → Response system → Attack signature database. Five components, in the same order the slides present them.

How They Collaborate

End-to-end flow of an intrusion detection event, from sensors on the network to the admin's desk:

Install DB Signatures
Sensors Gather Data
Alert Message Sent
IDS Responds
Admin Assesses Damage
Escalation if Necessary
Events Logged & Reviewed

In the reference architecture, sensors sit in the screened subnet DMZ while the database server and network management server live in a trusted management subnet — separating "sensing" from "managing."

06 / Deployment

Deployment of Network- and Host-Based IDS

💡

Staged IDS Deployment

Plan for a staged deployment. It helps gain experience and discover how much monitoring and maintenance of network resources is actually required — this varies depending on the size of an organization's network.

Deploying Network-Based IDS (NIDS)

An effective NIDS deployment requires a lot of attention concerning the network topology of the organization. Possible deployment options are categorized by the location of the IDS sensors. Consider all options and their advantages/disadvantages before placing a network-based IDS.

📍 Location 1 — Behind the External Firewall / in DMZ

Place an IDS sensor behind each external firewall and in the network DMZ.

  • Monitors attacks originating from the outside world
  • Highlights the inability of the firewall/policies to defend against attacks
  • Sees attacks targeting web or FTP servers located in the DMZ
  • Monitors outgoing traffic resulting from a compromised server
📍 Location 2 — Outside the External Firewall

Place an IDS sensor outside an external firewall.

  • Ability to identify the number and types of attacks originating from the Internet toward the network
📍 Location 3 — Major Network Backbones

Place an IDS sensor on major network backbones.

  • Monitors and inspects large amounts of traffic, increasing the chance of attack detection
  • Detects unauthorized attempts from outside the organization
📍 Location 4 — Critical Subnets

Place an IDS sensor on critical subnets.

  • Detects attacks on critical systems and resources
  • Focuses on specific critical systems and resources
🧠

Mnemonic — "Behind, Beyond, Backbone, Best-stuff"

"B-B-B-B" → Location 1: Behind the firewall/DMZ. Location 2: Beyond (outside) the firewall. Location 3: Backbone (major traffic arteries). Location 4: Best-stuff (critical subnets you most want to protect). Move from "closest to the internet" to "closest to the crown jewels."

Deploying a Host-Based IDS (HIDS)

Deploying a HIDS provides an additional layer of security. This type of IDS must be installed and configured on each critical system in the network — in fact, you should consider installing a host-based IDS on every host in the organization. When deploying a HIDS, it's recommended to have centralized management and reporting functions, which reduces the complexity of managing alerts from a large number of hosts.

07 / Alerts

IDS Alerts — The Four Outcomes

An alert is a graduated event that notifies that a particular event (or series of events) has reached a specified threshold and needs proper action by a responsible party. It sends a notification indicating that something is wrong and requires immediate attention and monitoring.

Type Attack Actually Happening? IDS Raises Alarm? Meaning
True Positive Attack Alert Correct — the IDS raises an alarm because a legitimate attack occurred
False Positive No Attack Alert Wrong — the IDS raises an alarm when no attack has taken place
False Negative Attack No Alert Wrong (and dangerous) — the IDS does not raise an alarm even though a legitimate attack occurred
True Negative No Attack No Alert Correct — the IDS stays quiet because there was nothing to catch
🧠

Mnemonic — "Crying Wolf vs Sleeping Guard"

False Positive = crying wolf (annoying, wastes time). False Negative = sleeping guard (dangerous, misses the real attack). Read the label backward: the first word says whether the alarm was truthful or a mistake; "Positive" always means an alert fired, "Negative" always means it stayed silent — regardless of whether an attack was really happening.

⚠️

Why false negatives are the scarier problem

A false positive wastes an analyst's time. A false negative means a real attack slipped through completely unnoticed — which is exactly why "not fine-tuning the IDS for false negatives and false positives" was listed as a common configuration mistake.

08 / Selection & Tools

Choosing a Good IDS/IPS & Real-World Tools

Characteristics of Good IDS Solutions

01. Run continuously with less human intervention

02. Must be fault tolerant

03. Resistant to subversion

04. Minimal overhead on the system

05. Observe deviations from normal behavior

06. Not easily deceived

07. Tailored to specific system needs

08. Copes with dynamic system behavior

Selection Criteria for an IDS/IPS Solution

IDS products must meet certain criteria to be deployed in an organization — compare different technology types, then select the one that best meets the requirements.

Requirement Category What to Evaluate
General requirements What the IDS products must meet post-deployment; org size affects how many products are needed
Security capability requirements Depends on the organization's environment, policies, and current security/network infrastructure
Performance requirements Capacity to handle network traffic/packet monitoring (NIDS) or event monitoring (HIDS)
Management requirements Must comply with the organization's management policy to be used effectively
Life cycle costs Estimated lifecycle costs should fit within the available budget

Intrusion Detection with Snort

Snort is an open-source network intrusion detection system, capable of performing real-time traffic analysis and packet logging on IP networks. It can perform protocol analysis and content searching/matching, and is used to detect a variety of attacks and probes — buffer overflows, stealth port scans, and OS fingerprinting attempts. It uses a flexible rules language to describe traffic it should collect or pass, plus a detection engine that uses a modular plug-in architecture.

Other Intrusion Detection Tools

Suricata

A robust network threat detection engine capable of real-time IDS, inline IPS, network security monitoring (NSM), and offline pcap processing.

AlienVault OSSIM

Open-source security information and event management platform.

SolarWinds Security Event Manager

Commercial SIEM/log-analysis platform.

OSSEC

Open-source host-based intrusion detection system (HIDS).

Zeek

Network security monitoring framework (formerly Bro).

Sagan Log Analysis Engine

Real-time log analysis engine, often paired with Snort/Suricata rule sets.

Snort = the name to remember

If the exam asks for a named open-source NIDS tool, Snort is the headline answer — it's the one the slides single out for its own dedicated slide (real-time traffic analysis + packet logging + rules language + plug-in detection engine).

09 / Exam Prep

Exam Tips & Tricks

💡

IDS = detect, IPS = detect + block

IPS is placed inline; IDS typically sits off to the side inspecting a copy of traffic. Don't say IDS "blocks" traffic — that's IPS's job.

Signature = known threats only

Signature-based detection has low false positives but zero-day blindness. Anomaly-based catches unknowns but has more false alarms. This trade-off is a favorite exam question.

⚠️

False Negative is worse than False Positive

A false negative means a real attack was missed entirely. A false positive is just a false alarm — annoying, but not dangerous.

🔑

Six classification axes, in slide order

Approach → Protected system → Structure → Data source → Behavior after attack → Analysis timing. Each axis has exactly 2–3 categories.

📋

NIDS sees the network, HIDS sees one host

Hybrid IDS = low false-positive rate of NIDS + anomaly detection strength of HIDS, aimed at catching unknown attacks.

🌐

Four NIDS deployment locations

(1) Behind firewall/in DMZ, (2) outside the firewall, (3) on network backbones, (4) on critical subnets. Each has its own advantage worth memorizing.

🧩

IDS is not antivirus, not a vuln scanner, not crypto, not just a log system

A favorite "what IDS is NOT" trick question — remember the LVAC list.

🛠️

Snort is the star tool

Open-source, real-time traffic analysis + packet logging, rules-based, modular plug-in detection engine. Know it as the flagship named NIDS example.

10 / Cheat Sheet

Quick Reference — Everything at a Glance

Topic Key Point
IDS/IPS Network security appliance that inspects inbound/outbound traffic for suspicious patterns
IDS vs IPS IDS detects and alerts (passive); IPS detects and blocks (inline)
IDS placement vs firewall IDS works from inside the network and sits behind the firewall; firewall only looks outside
IDS pipeline Internet → Router → Firewall → IDS → signature/anomaly/protocol checks → alarm + action
Signature (misuse) detection Matches known attack patterns; low false alarms; can't catch unknown/variant attacks
Anomaly detection Flags deviation from learned normal behavior; catches unknowns; more false alarms
Stateful protocol analysis Compares activity to protocol's own accepted-behavior profile; flags command-order/authenticator anomalies
Six classification axes Approach, Protected system, Structure, Data source, Behavior after attack, Analysis timing
NIDS Network intrusion detection system — protects the network
HIDS Host intrusion detection system — protects one host; install on every critical/every host
Hybrid IDS Combines NIDS's low false-positive rate with HIDS's anomaly detection to catch unknown attacks
Centralized IDS All data shipped to one central location for analysis
Distributed IDS Several IDS deployed across a large network, communicating with each other
Active IDS Detects AND responds to intrusions
Passive IDS Only detects intrusions (listen and monitor)
Interval-based IDS Stores and forwards data; analyzes offline
Real-time-based IDS Continuous flow from monitoring points to analysis engine; analyzes on the fly
Data sources Audit trails (app flaws/violations) vs network packets (well-known attacks)
IDS capabilities Monitor user/system activity, analyze configs/vulnerabilities, assess integrity, track policy violations
IDS/IPS is NOT A logging system, a vulnerability scanner, antivirus, or a cryptographic system
Common config mistakes Blind spots in placement, ignoring alerts, no response policy, poor false pos/neg tuning, stale signatures, inbound-only monitoring
General intrusion indicators File system (new/missing files), network (repeated probes, odd logins), system (slow performance, gaps in logs, crashes)
IDS components Network sensors, Command console, Alert systems, Response system, Attack signature database
Network sensor placement Internet gateways, between LAN connections, remote-access/VPN servers, either side of firewall
Command console rule Must run on a dedicated system — non-dedicated hosts slow down security event response
Response system rule Don't rely on it solely — a human must decide on false positives/escalation
Attack signature database Must be updated periodically; IDS compares traffic against it to raise alerts/block
Staged deployment Roll out in stages to learn how much monitoring/maintenance is actually required
NIDS deployment Location 1 Behind external firewall / in DMZ — monitors outside attacks, DMZ server attacks, compromised-server outbound traffic
NIDS deployment Location 2 Outside the external firewall — counts/types all Internet-origin attack attempts
NIDS deployment Location 3 On major network backbones — inspects large traffic volumes, catches outside unauthorized attempts
NIDS deployment Location 4 On critical subnets — focuses on the most critical systems/resources
HIDS deployment Install on every critical system (ideally every host); use centralized management/reporting
Alert definition Graduated event notifying that a threshold was reached; requires action from a responsible party
True Positive Attack happened + alert raised (correct)
False Positive No attack + alert raised (false alarm)
False Negative Attack happened + no alert (missed attack — most dangerous)
True Negative No attack + no alert (correct)
Good IDS characteristics Runs continuously, fault tolerant, subversion-resistant, low overhead, spots deviations, hard to deceive, tailored, adapts to dynamic behavior
Selection criteria General requirements, security capability, performance, management requirements, life cycle costs
Snort Open-source NIDS; real-time traffic analysis + packet logging; rules language + plug-in detection engine
Other tools Suricata, AlienVault OSSIM, SolarWinds Security Event Manager, OSSEC, Zeek, Sagan Log Analysis Engine