Internet traffic does not run at office hours. As one part of the world winds down for the evening, another is waking up and starting to stream, browse and route, and the exchange carrying that traffic has to perform the same at four in the morning as it does at four in the afternoon. Keeping it that way is the job of the LINX Network Operations Centre (NOC), the team that monitors the exchange, responds to incidents and supports members around the clock.
Watching the Network, One Alert at a Time
The NOC monitors services, responds to incidents and member tickets, provisions new services and coordinates planned maintenance with engineering. Two tools sit at the centre of that. Icinga does the alerting, so when a port drops it raises the alarm and opens a ticket straight away. Grafana is where engineers look deeper, at how a port is performing and how much traffic it is carrying. If something looks wrong, the NOC picks it up with the member or checks the physical connection on-site themselves or via remote hands.
The Difference an In-House Team Can Make
The NOC came in-house around a decade ago, and it changed the relationship with members. Members now speak directly to engineers who know the network, the systems and the processes inside out. Responses are quicker and the support is more personal. The names on both sides have become familiar over the years, and when a ticket starts going back and forth the team is happy to move it onto a call and sort it there to resolve any issues, queries or concerns.
Two in the Morning is a Working Hour
The internet doesn’t sleep, and neither does the NOC, the team works 24/7 365. Members sit in different time zones, and a ticket can land at two in the morning as easily as two in the afternoon. Often that is deliberate. A member upgrading from 10GE to 100GE, or 100GE to 400GE, wants the change made when their traffic is lowest, so the NOC goes to site to move the cabling and handles the configuration remotely. Where the team cannot attend in person, the NOC engages a remote hands team at the data centre to do the physical work instead.
Types of Issues the NOC Handles
Fibre breaks are among the most common incidents handled by the NOC. These are typically caused by damage to cables between sites on a LAN, as opposed to within the data centres, especially in up-and-coming metros. Monitoring systems usually detect the outage first, allowing the team to quickly identify the affected members and services. The NOC then raises the issue with the supplier, monitors progress, escalates where necessary to ensure a timely resolution, and keeps members informed through the appropriate mailing lists and communication channels.
Hardware failures also occur from time to time. The failure of a line card carrying member-facing ports is treated as a major incident because of the potential impact on connected members. During these incidents, the NOC works closely with engineering teams, assists with coordination of on-site activities, and provides regular updates to affected members.
The team also deals with issues arising from member-side changes. A common example is when a member replaces or reconfigures a router, resulting in a new MAC address being presented to the exchange. Because LINX applies MAC filtering on member connections, the new MAC address may no longer match the expected value, leading to connectivity issues. In most cases, members can update their MAC address details themselves through the LINX Portal. Where this is not possible, the NOC can validate the change and make the necessary updates on the member’s behalf to restore connectivity.
Processes That Hold Under Pressure
None of it holds together without documentation. The NOC maintains provisioning and support processes owned across the team rather than by one person, practised in a test environment by new starters and reviewed regularly. The test is simple. A process has to work for an engineer in the middle of the night, and for someone in their first week.
Why the NOC Matters
The LINX NOC exists because one of the world’s largest internet exchanges needs people watching it, all the time. Supporting members, responding to incidents, quietly improving something in the background. The job is the same every hour of every day, and that is to keep the network stable and dependable.
< Go Back