- A NOC keeps networks and services running. A SOC (security operations center) defends them against attacks.
- Work flows by tier: automation first, then L1 with runbooks, then L2 and subject experts.
- A NOC can be in-house or outsourced, in one room or fully remote.
- The NOC's ticket, alert, change and handoff records are the history of how the network was run.
What is a network operations center (NOC)?
IBM puts it simply: "A network operations center (NOC) is a centralized location where computer, telecommunications, or satellite networks systems are monitored and managed 24x7" (IBM). PagerDuty stresses that the place can be virtual: "A network operations center (NOC) is the centralized location (in-office or virtual) for an organization's network team" (PagerDuty).
IBM names telecommunications, financial services, manufacturing and energy among the industries that need networks running around the clock. Some companies keep the NOC in-house; IBM notes that "Some companies choose to operate a NOC internally and locate the infrastructure and operations hub on-premises, often within the data center" (IBM), while others outsource it to a provider that runs monitoring for many clients.
What does a NOC do day to day?
The job is to notice problems early and get them fixed. The NICE Framework work role for network operations, published by CISA, describes it as "Responsible for planning, implementing, and operating network services and systems, including hardware and virtual environments" (CISA NICCS), with tasks such as "Monitor network capacity and performance" and "Diagnose network connectivity problems". A typical shift looks like this:
- Watch dashboards and alerts for links, routers, firewalls, servers, cloud services and applications.
- Triage each alert: real or noise, which service, how many users affected, what severity.
- Run the runbook for known problems, and escalate the rest to the right team or vendor.
- Open, update and close tickets, with a timeline of what was seen and done.
- Coordinate planned maintenance and approved changes, and watch for side effects.
- Talk to carriers and vendors during outages, and post status updates.
- Write a handoff note for the next shift, and feed problems into postmortems.
How is a NOC team organized?
IBM notes that "NOCs typically operate in a tiered fashion" (IBM): routine alerts at the first level, the most severe incidents at the top, and issues move up when a technician cannot solve them in time. PagerDuty describes the levels this way (PagerDuty):
| Level | Who or what | Typical work |
|---|---|---|
| L0 | Automation | Suppresses duplicate alerts and fixes well-understood issues before a person sees them |
| L1 | First human responder | Works through attached runbooks for routine, documented issues |
| L2 | Senior engineer | Deeper triage and troubleshooting; may declare a major incident |
| Experts | Service owners and subject experts | Fix problems in systems the NOC did not build |
| Leadership | NOC manager, director | Staffing, goals, tools and reporting |
PagerDuty's description of L1 is worth copying into any NOC handbook: "L1 engineers work through attached, predetermined runbooks to resolve routine, documented issues." For how runbooks differ from playbooks, see runbook vs playbook.
What is the difference between a NOC and a SOC?
A NOC protects availability; a SOC protects against attacks. IBM writes that a SOC "is focused on network and information security, performing threat analysis and monitoring for attacks on a client network", while the NOC's mission is network uptime (IBM).
| NOC | SOC | |
|---|---|---|
| Main goal | Keep networks and services up and fast | Detect and stop attacks |
| Typical alert | Link down, high latency, disk full, failed backup | Suspicious sign-in, malware, data leaving the network |
| Main tools | Network monitoring, paging, ticketing, config backups | SIEM (security log analysis), EDR (endpoint detection), threat intelligence |
| Success measure | Uptime, time to acknowledge, time to restore | Time to detect, time to contain |
| Works with | Network, server and cloud teams, carriers | Security engineers, legal, incident response |
What records does a NOC keep?
A NOC lives in its systems of record. PagerDuty advises NOCs to "Integrate with the rest of the organization's technical systems, including systems of record like Jira or ServiceNow" and to "Conduct postmortems to recognize what went well during response and what could have gone better" (PagerDuty). Changes go through change control, which NIST SP 800-128 defines as the "process for managing updates to the baseline configurations for the configuration items" (NIST SP 800-128). For catching changes made outside that process, see config drift.
Telecom NOCs also handle regulatory outage reports. Under FCC rules, an outage means "a significant degradation in the ability of an end user to establish and maintain a channel of communications" (47 CFR 4.5), and wireline providers must notify the FCC electronically within 120 minutes of discovering an outage of at least 30 minutes that meets the rule's thresholds (47 CFR 4.9). This is general information, not legal advice; the rule has more conditions than shown here.
| Record | What it holds |
|---|---|
| Alerts and events | What the monitoring saw, when, on which device |
| Tickets | Severity, timeline, actions, escalations, resolution |
| Change records | What was changed, who approved it, the config diff, the rollback |
| Config backups | Router, switch and firewall configurations over time |
| Shift handoffs and on-call logs | Open issues, who was paged, who acknowledged |
| Postmortems and outage reports | Cause, impact, follow-up actions |
Example: a NOC ticket for a transit link outage
The ticket below is an illustration written for this page, not a real operator's record. The address block and AS number are reserved for documentation. It shows what a good NOC ticket has: a timeline in one time zone, the runbook step used, the escalation, the linked change and its rollback, and a pointer to the postmortem.
Ticket NOC-48213 Severity 2 Status: resolved
Service: transit link EDGE-2 to upstream AS64501 (documentation AS number)
Timeline (UTC)
02:14 Alert: BGP session EDGE-2 <-> AS64501 down; traffic shifts to EDGE-1
02:15 L0 automation: link error counters attached to the ticket
02:19 L1 acknowledges; runbook "Transit BGP down" step 1: light levels low
02:31 L1 escalates to L2: fiber suspected; opens case with the carrier
02:58 L2: EDGE-1 at 85% of capacity; applies approved change CHG-7731
(prefer EDGE-3 for 198.51.100.0/24) to spread the load
04:40 Carrier reports a cut fiber repaired
04:47 BGP session re-established; CHG-7731 rolled back at 05:05
05:20 Shift handoff note: watch EDGE-2 errors for 24 hours
Linked: alert ALR-99120, change CHG-7731 (config diff), carrier case,
postmortem PM-112 (follow-up: alert on light levels earlier)
Because the alert, the change and the postmortem are linked by ID, anyone can rebuild what happened later. The incident timeline template shows how to lay those joins out.
Common mistakes in running a NOC
PagerDuty notes that "It's no longer possible to maintain eyes-on-glass methods with engineers watching screens all day to pinpoint issues" (PagerDuty). That is the root of most of these:
- Too many alerts and no deduplication, so real problems hide in noise.
- Runbooks that are missing, untested or stored on a system that goes down with the outage.
- Tickets closed with "fixed" and no timeline, cause or link to the change.
- Emergency changes made on a router console with no change record, leaving configuration drift behind.
- Shift handoffs done by voice only, so open issues are lost overnight.
- Timestamps in mixed time zones across alerts, tickets and chat.
- Pasting customer names, circuit IDs or passwords into tickets that many people can read.
Are a NOC's records useful outside the company?
They can be, when they describe the operator's own network. AI data buyers have asked Origin for the records of companies that run and defend their own infrastructure: incidents joined to the change that fixed them (the ticket, the chat thread, the config diff or pull request, the outcome), configuration and routing history going back years, ordered logs of what happened, and the runbooks, on-call records and postmortems from the same period. For a NOC that means NOC tickets, routing and config changes, and outage records of the operator's own network. A NOC run for clients holds records about client environments, and those stay out unless the client has agreed. ISPs and telecom operators should read the ISP network operations page, which covers subscriber data and CPNI. The security incident data page and audit log examples cover the security side.
Origin is an independent data partner. Your company decides what is shared and signs any license itself, and a buyer may pay Origin when a deal closes.
- Subscriber and customer data come out first: names, accounts, circuit IDs tied to one customer, traffic contents.
- Secrets: passwords, SNMP community strings and keys in config backups are removed and rotated.
- Addresses and hostnames are pseudonymized consistently, so the sequence still reads (anonymizing pcap files and logs).
- Regulatory filings and threat intelligence received under TLP limits stay out.
Does your company hold records like these? See what your data is worth.
Common questions
What does a network operations center do?
A network operations center monitors an organization's networks, servers and services, usually around the clock. Its staff triage alerts, follow runbooks for known problems, escalate the rest, track work in tickets, coordinate planned changes and maintenance, talk to carriers and vendors during outages and hand open issues to the next shift.
What is the difference between a NOC and a SOC?
A NOC keeps networks and services available and fast: link failures, latency, capacity and outages. A SOC, or security operations center, detects and responds to attacks: suspicious sign-ins, malware and data theft. They use different tools, alerts and measures of success, and each escalates to different teams, though both work from tickets and runbooks.
What are NOC levels L1 and L2?
L1 is the first person to see an event. L1 engineers work through runbooks to fix routine, documented problems. L2 engineers take what L1 cannot fix, troubleshoot more deeply and can declare a major incident. Many NOCs add L0, meaning automation that filters and fixes known issues, and escalate past L2 to service experts.
What tools does a NOC use?
A NOC usually uses network and infrastructure monitoring, an alerting and paging tool, a ticketing system, a change management process, configuration backup tools for routers, switches and firewalls, dashboards, a status page and chat. Runbooks link these together. The exact products differ by company, but every NOC needs alerts, tickets and change records joined by ID.
Does a NOC work 24 hours a day?
Most do. IBM describes a NOC as monitoring networks 24x7. Coverage comes from rotating shifts in one location, teams in several time zones that follow the sun, or an on-call engineer paged outside business hours. Each shift change needs a written handoff note so open issues carry over.
Can a company outsource its network operations center?
Yes. Many companies hire a provider that monitors and manages their networks from its own NOC, while others keep it in-house for control. The contract should state which alerts the provider handles, response times, escalation contacts, who may make changes and who owns the tickets, logs and configuration backups it creates.
Sources and further reading
- IBM: What is a network operations center (NOC)?
- PagerDuty: NOC best practices
- CISA NICCS: NICE Framework work role, Network Operations
- NIST SP 800-128: Guide for Security-Focused Configuration Management of Information Systems (PDF)
- 47 CFR 4.5: definitions of outage and special facilities (LII)
- 47 CFR 4.9: outage reporting requirements (LII)
