⚙️ What is Threat Detection?
Definition:
Threat detection is the process of identifying malicious activity or security breaches in a system or network.
Goal:
- Detect attacks early
- Respond quickly to prevent damage
- Maintain confidentiality, integrity, and availability
1️⃣ SIEM (Security Information and Event Management)
Definition:
SIEM collects, aggregates, and analyzes logs and events from multiple sources to detect suspicious patterns.
Key Functions:
- Log collection → from servers, network devices, firewalls, apps
- Correlation → detect patterns across systems
- Alerting → notify security team about potential threats
- Reporting & compliance → audit trails, regulatory compliance
Examples:
- Splunk, IBM QRadar, ArcSight, Azure Sentinel
Pros:
- Centralized view of security events
- Can detect complex multi-step attacks
Cons:
- Reactive; relies on logs
- Requires tuning to reduce false positives
2️⃣ IDS (Intrusion Detection System)
Definition:
IDS monitors network or host activity to detect suspicious behavior or policy violations, but does not block traffic.
Types:
- Network IDS (NIDS) → monitors network traffic
- Host IDS (HIDS) → monitors individual systems
Detection Methods:
- Signature-based: match known attack patterns
- Anomaly-based: detect deviations from normal behavior
Pros:
- Detects attacks in real-time
- Can provide alerts for forensic analysis
Cons:
- Cannot prevent attacks (alert only)
- Signature-based IDS cannot detect unknown attacks
Examples:
- Snort, Suricata, OSSEC
3️⃣ IPS (Intrusion Prevention System)
Definition:
IPS is similar to IDS but can actively block malicious traffic instead of just alerting.
Functions:
- Detect malicious activity (like IDS)
- Drop packets or terminate connections
- Protect networks automatically
Pros:
- Real-time prevention of attacks
- Reduces workload on security team
Cons:
- False positives can block legitimate traffic
- Needs careful configuration
Examples:
- Cisco Firepower, Palo Alto Threat Prevention
🔄 Comparison Table
| Feature | SIEM | IDS | IPS |
|---|---|---|---|
| Purpose | Analyze logs/events | Detect intrusions | Detect & prevent intrusions |
| Action | Alert/report | Alert | Block / prevent |
| Data Source | Logs from multiple sources | Network/Host traffic | Network/Host traffic |
| Detection | Correlation & patterns | Signatures/Anomalies | Signatures/Anomalies |
| Real-time | Usually near real-time | Real-time | Real-time |
🧩 Analogy
- SIEM: Security dashboard → collects CCTV footage and alarms, analyzes patterns
- IDS: Security guard → notices suspicious behavior and calls the team
- IPS: Security guard + automatic lock → notices suspicious behavior and blocks entry immediately
✅ Key Takeaways
- Threat detection = identify malicious activity early
- SIEM → centralizes logs, correlates events, alerts
- IDS → detects intrusions, alerts only
- IPS → detects intrusions, actively blocks them
- For robust security → often used together: SIEM + IDS/IPS