What a SOC does

Monitoring, detection, triage, and response: the blue team's daily job

By the end of this lesson you can

  • Describe the seven core services of a SOC and give one example output for each.
  • Tell an event from an alert and explain why an alert is a question, not a verdict.
  • Place the SOC's daily work in the NIST CSF 2.0 Detect Function (DE.CM and DE.AE) and show where it hands over to Respond.
  • Compare in-house, managed, and hybrid SOC models and name one strength and one watch-out for each.
  • List the four things a SOC needs to work: authority, data, people, and tools.

0107:00 at Ridgeway

Throughout this course you follow Ridgeway Logistics, a fictional freight forwarding company with about 140 staff, a head office, and two depots. Office staff use Windows laptops. A Linux web server, SRV-WEB-01, runs the customer shipment tracking portal track.example.com. A domain controller, DC-01, holds the user accounts. Everything here is an example made for the course.

It is Mira's first day as a Tier 1 analyst. She logs in at 07:00 Europe/Zurich time (05:00 UTC) and opens the queue. It already holds 23 alerts from the night: failed logins on the portal, a warning about an expiring certificate, a laptop that contacted a domain nobody has seen before, a scheduled job flagged as "suspicious PowerShell".

Jonas, the Tier 2 analyst who mentors her, puts a coffee next to her keyboard and says: "None of these is an emergency until we know what it is. Let's find out, one by one, and write down what we learn." That sentence is the whole job in miniature, and this course teaches you to do it well.

02What a SOC is for

A SOC (Security Operations Center) is the team, processes, and tools that watch an organization's systems for attacks, analyze what looks wrong, and start the response. People often call the defensive side of security the blue team, as opposed to a red team that plays the attacker in exercises. The SOC is where blue team work happens every day.

Strip away the tools and a SOC does four things:

  • Watch. Collect logs and other data from the systems that matter, all the time. This is security monitoring.
  • Decide. Look at what the monitoring raises and decide quickly what it means: harmless, wrong, or worth digging into.
  • Escalate. Hand real problems to the people who can act on them, with the evidence attached, along a path agreed in advance.
  • Improve. Feed what was learned back into rules, playbooks, and the systems themselves, so the same problem is caught sooner or never happens again.

Notice what is missing: the SOC does not usually fix servers, reset every password, or rebuild laptops itself. Those actions belong to IT owners and to the incident response team. The SOC's power is in seeing early, judging calmly, and telling the right people.

Myth

"A SOC is a room full of screens." The wall of dashboards is the picture everyone knows, but a SOC is a function: people who follow processes and use data. A two-person team with good logs, clear rules, and a phone number for the night provider is a SOC. A dark room with expensive screens and no agreed process is not.

03The services of a SOC

MITRE, the organization behind the ATT&CK knowledge base you meet in lesson 14, publishes a free book called 11 Strategies of a World-Class Cybersecurity Operations Center (second edition, 2022). In our own words, its eleven strategies are: know what you protect and why; give the SOC the authority to do its job; build a structure that fits the organization; hire and grow good people; put incident response first; use threat intelligence; collect the right data; choose tools that support the analyst's workflow; communicate, collaborate, and share; measure performance to improve it; and add SOC functions over time. A "Strategy 0" covers the fundamentals.

Those strategies describe how to build a SOC. Day to day, the work splits into seven services. Explore them below. Each shows what the service does, a Ridgeway example, and its main output.

Monitor Triage Investigate Support response Detection engineering Threat intelligence Report and improve

Monitor

Collects logs and telemetry from the systems that matter and keeps detection rules running on them. Ridgeway example: the portal server, the domain controller, and the laptops send their logs to one central place. Main output: a steady stream of events, and alerts when rules match.

Triage

Looks at each new alert and decides fast what it means and how urgent it is: close, dig deeper, or escalate. Ridgeway example: Mira checks whether 40 failed portal logins came from one address or many. Main output: a first verdict and a short note on every alert.

Investigate

Takes the alerts that need more than a few minutes and follows the trail across sources: what happened before and after, which accounts and machines are involved. Ridgeway example: Jonas checks every login of an account that suddenly signed in from abroad. Main output: a case with a timeline and a clear conclusion.

Support response

Helps the incident response team and IT owners contain and fix a confirmed problem, by giving them evidence and watching for further activity. Ridgeway example: the SOC asks IT to disable a misused account and then watches for new logins with it. Main output: requests to act, with evidence, and confirmation that the actions worked.

Detection engineering

Writes, tests, and tunes the rules that create alerts, so they catch what matters without drowning analysts. Ridgeway example: adding an exception for Leon's approved inventory script after it fired every night. Main output: new or improved detection rules, with notes on why they changed.

Threat intelligence

Gathers and analyzes information about attackers and their methods, and turns it into things to look for. Ridgeway example: a partner warns about a fake invoice campaign aimed at logistics firms, and the SOC searches mail logs for it. Main output: watch lists, hunting ideas, and context for triage.

Report and improve

Tells management and IT what the SOC saw, what it changed, and where the gaps are, and measures its own work. Ridgeway example: Priya, the SOC lead, writes a monthly summary of alert types and open weaknesses. Main output: reports, metrics, and improvement tasks.

Monitoring feeds triage and investigation. Response support, detection engineering, and threat intelligence keep the cycle sharp. Reporting closes the loop.

In a small SOC one person may cover four of these services in a single shift. In a large one each may be a separate team. The services stay the same.

04Events, alerts and incidents

Three words carry most of the SOC's work, and beginners often mix them up.

  • An event is anything observable that happens on a system or network: a login, a file opened, a DNS query, a process starting. A busy organization produces a great many every minute. Almost all are harmless.
  • An alert is created when an event, or a group of events, matches a detection rule. The rule says "this pattern is worth a human look." An alert is a question ("is this bad?"), not an answer.
  • An incident is what you have when the answer is yes, and the activity meets the organization's agreed criteria for one: a real attack, a real compromise, or a real breach of policy that needs a coordinated response.

So at Ridgeway, Leon logging in to his laptop is an event. Twelve failed logins to the same account within a minute may match a rule and become an alert. If Mira and Jonas find that someone outside the company then logged in successfully, the case meets Ridgeway's criteria and becomes an incident. Lesson 3 follows this path step by step.

WordWhat it isRidgeway exampleWho acts
EventSomething observable happenedA user logs in to the portalNobody, it is stored
AlertA rule matched, a question to answerTwelve failed logins to one account in a minuteTier 1 triage
IncidentConfirmed harm or threat that meets the criteriaAn outside address then logs in successfullyIncident response team, with the SOC

05Where the SOC sits in CSF 2.0

The NIST CSF 2.0 (the Cybersecurity Framework published by the US National Institute of Standards and Technology in February 2024) groups security work into six Functions: Govern (GV), Identify (ID), Protect (PR), Detect (DE), Respond (RS), and Recover (RC). Each Function is split into Categories and those into Subcategories with short IDs.

The SOC's daily work sits mostly in Detect, which is about finding and analyzing possible attacks and compromises. Detect has two Categories:

  • DE.CM, Continuous Monitoring. Assets are watched to find anomalies, indicators of compromise, and other possibly harmful events. Its Subcategories cover networks and network services (DE.CM-01), the physical environment (DE.CM-02), what staff do and how they use technology (DE.CM-03), the activity of external service providers (DE.CM-06), and computers, software, runtime environments, and their data (DE.CM-09).
  • DE.AE, Adverse Event Analysis. What monitoring finds is analyzed so incidents can be recognized. Its Subcategories include analyzing events to understand the activity (DE.AE-02), correlating information from several sources (DE.AE-03), estimating impact and scope (DE.AE-04), giving information about adverse events to authorized staff and tools (DE.AE-06), bringing in threat intelligence and other context (DE.AE-07), and declaring an incident when the agreed criteria are met (DE.AE-08).

Read that list again with Mira's queue in mind. Collecting the logs is DE.CM. Triage is DE.AE-02. Checking the portal log against the domain controller log is DE.AE-03. Escalating to Jonas is DE.AE-06. The moment of declaring an incident, DE.AE-08, is the handover point to Respond.

NIST's incident response guide, SP 800-61 Revision 3 (April 2025), organizes incident response on these same six Functions. The course "Incident Response and Digital Forensics" (Track 10) teaches that side in depth. This course stays on the SOC side of the handover.

Names vary

CSF calls it Continuous Monitoring, vendors talk about "detection and response," and job ads say "security monitoring" or "SOC operations." The activity is the same. When you meet a new term, ask which of the four verbs it covers: watch, decide, escalate, or improve.

06In-house, managed or hybrid

Not every organization can staff a SOC around the clock. Three models are common:

SOC modelWho staffs itStrengthsWatch-outs
In-houseThe organization's own analystsKnows the business, the people, and what is normal. Can act fast inside the companyHard to cover nights and weekends. Needs hiring, training, and tools
Managed providerAn external company watches the logs under a contractCoverage at all hours. Sees many customers and many attacksKnows less about your business. Can only act as far as the contract allows. You still own the risk
HybridOwn team for part of the time or part of the work, a provider for the restLocal knowledge by day, coverage by nightTwo teams must share criteria, tools, and notes. Every handover is a chance to drop something

Ridgeway runs a hybrid. Its small in-house SOC works weekdays from 07:00 to 19:00 Europe/Zurich time. A managed security provider, written here as example.org, covers nights and weekends. Twice a day one team hands over to the other, which is why lesson 22 is all about handover notes.

Whatever the model, MITRE's book makes clear that a SOC cannot work without four foundations:

Need 1

Authority

A written mandate that says what the SOC may watch, whom it may ask for logs, and who must act on its escalations. Without it, an analyst who spots an attack at 03:00 cannot get anything done.

Need 2

Data

The right logs from the systems that matter, with correct time stamps, kept long enough to look back. No data, no detection.

Need 3

People

Trained analysts with time to think, a path to grow, and respect for front-line work. Tools do not triage by themselves.

Need 4

Tools

A place to search logs, raise alerts, and record cases that fits how analysts actually work, not the other way around.

Defender's view

Start with what you protect. MITRE's first strategy is to know what you protect and why. A SOC that monitors everything equally watches nothing well. Ridgeway's tracking portal, domain controller, and payroll deserve the best logs and the fastest triage. A meeting room display does not. That list decides which logs to collect, which alerts come first, and what counts as an incident.

Law and ethics

Security monitoring often touches what people do at work: logins, web visits, emails. That makes logs personal data in many places. Monitoring must follow your organization's written policy and local law. Before watching user activity, the SOC asks HR and legal what is allowed, and it never monitors a network or a person it has no mandate for.

07How this course works

You will do the job, not just read about it. From lesson 4 on you work in an Ubuntu virtual machine called soc-lab, with free tools: jq, SQLite, sigma-cli, and Zeek and Suricata in containers. You read logs that the lessons write for you, practice captures published with the Zeek test suite, and community detection rules from SigmaHQ.

Three habits run through every lesson:

  • Your own lab only. You never point a tool at a network you do not own or are not authorized in writing to monitor, not even your employer's or a shared apartment's.
  • Times with a zone. Ridgeway's day runs on Europe/Zurich time, but tickets use UTC in ISO 8601 form, such as 2026-10-07T05:00:00Z.
  • An analyst notebook. Keep a paper notebook or a text file. Every lesson ends with an entry, the same way Mira writes down what she checked and why. Good notes are half of SOC work.

Lessons 1 to 3 need only paper and pen. Lesson 4 builds the lab.

Practical exercise: map Ridgeway's SOC

Paper and pen, or a text file that becomes your analyst notebook. Everything is about the fictional Ridgeway Logistics. Check off each step as you finish it. Your checkmarks are saved in this browser.

Key takeaways

  • A SOC watches, decides, escalates, and improves. It is a function made of people, process, and data, not a room.
  • Seven services make up the work: monitor, triage, investigate, support response, detection engineering, threat intelligence, report and improve.
  • An event is anything that happens, an alert is a rule match and a question, and an incident is a confirmed problem that meets agreed criteria.
  • In NIST CSF 2.0 the SOC lives mostly in Detect: DE.CM for monitoring and DE.AE for analysis, handing over to Respond when an incident is declared.
  • In-house, managed, and hybrid SOCs all need authority, data, people, and tools, and all start with knowing what they protect.