A SOC (Security Operations Center) is the team and processes that watch security events, detect attacks and respond to them. How a SOC differs from a SIEM and which models exist we covered in “SIEM vs SOC: the difference for business” and “SOC as a Service or in-house SOC: how to choose”.
This post is about what happens inside: how an event becomes an incident, who does what, how a SOC’s work is measured and what the company gets.
From event to closed incident
- Event sources. Network equipment and NGFWs, servers, endpoints, web applications, cloud and email services send logs to a single platform. On endpoint protection, see “Antivirus vs EDR: what a business needs”.
- Correlation. The system matches events from different sources against detection rules and raises alerts.
- Triage. An analyst checks the alert: a false positive or a real threat, and how severe.
- Investigation. For real threats, the scope and root cause are established: what happened, which systems and accounts are affected.
- Response. Actions follow an agreed playbook: isolating a device, blocking an account, recommendations to the IT team.
- Closure and report. What happened, what was done, what to change so it does not happen again.
ENISA calls the SOC the first line: it receives all alerts, registers incidents and escalates them to the incident response team.
Who does what: analyst tiers
A typical split of roles in a SOC:
| Tier | What it does |
|---|---|
| First line (L1) | Triages alerts, weeds out false positives, escalates according to instructions |
| Second line (L2) | Investigates incidents, establishes scope and root cause, responds |
| Third line (L3) | Handles complex incidents, hunts for hidden threats, analyses new attack methods |
| Platform engineer | Connects sources, configures and updates detection rules |
The more routine work automation takes on, the more time analysts have for investigation.
Playbooks, automation and threat hunting
- Playbooks are step-by-step scenarios for common incidents: what to check, whom to notify, who decides on isolation. How to write them: “Incident response: the plan and playbooks”.
- Automation (SOAR) carries out routine steps on its own: enriches alerts with data, blocks an address, opens a ticket.
- Threat hunting is an active search for traces of attacks that did not raise any alert. ENISA lists data for threat hunting among the outputs of monitoring.
How a SOC’s work is measured
ENISA notes that SOCs usually build key performance indicators around detection quality — detection speed and breadth, coverage, false-positive rates — as well as the number of incidents handled, the ratio of alerts, events and incidents, the number of escalations and the workload per incident.
In practice, people most often look at:
- MTTD (mean time to detect) — the average time from the start of an attack to its detection;
- MTTR (mean time to respond) — the average time from detection to response;
- coverage — what share of critical systems is connected to monitoring;
- the false-positive rate — how many alerts turned out to be nothing.
Response times are worth setting separately for each incident severity level and writing into the contract.
How onboarding goes
- Source and risk analysis: identify critical assets and the systems to connect first.
- Onboarding and baseline: collect logs and learn the normal behaviour of the infrastructure.
- Detection tuning: adapt rules to the company’s specifics and agree response playbooks.
- Operations: continuous monitoring, response and regular reports.
How to assess a SOC — your own or a provider’s
- Which systems are connected, and which are not?
- How many alerts per month, and how many of them are real incidents?
- What is the average time to detect and respond for each severity level?
- What do you receive: raw alerts or notifications with a severity assessment and recommended actions?
- How often are detection rules updated?
- Is threat hunting done, and what has it found?
Checklist
- Critical systems are connected to monitoring, and coverage is known
- There are playbooks for common incidents
- It is defined who decides on isolating systems
- Response times are set for each severity level
- MTTD, MTTR and the false-positive rate are tracked
- Detection rules are reviewed regularly
- Management receives regular reports
How Bereg-IT can help
SOC as a Service from Bereg-IT covers the full cycle from event to closed incident: connecting sources, correlation and detection rules, triage and investigation, notifications with a severity assessment and recommended actions, response according to an agreed playbook and regular reports for IT and management. Onboarding follows the four stages above. Who does what is fixed before launch, and the coverage schedule, response time for each severity level and log retention period go into the contract.
The first consultation is free. Send a request and an information security specialist will get in touch.
Sources
- ENISA, How to set up CSIRT and SOC (10.12.2020) — SOC processes and key performance indicators: https://www.enisa.europa.eu/publications/how-to-set-up-csirt-and-soc