IT explained · 6 min read

What is a node in networking? Devices, hops and addresses

Written for: Anyone reading network documentation or a monitoring licence and needing to know exactly what a node is and what it is not.

Nodes and hops

A mesh of network nodes joined by links, with edge devices, a switch, a router and a cloud gateway labelled by shape.A mesh of network nodes joined by links, with edge devices, a switch, a router and a cloud gateway labelled by shape.
Every addressable thing on a network is a node: devices at the edges, and the switches, routers and gateways that pass traffic between them.

In short

A node is any addressable device on a network that can send, receive or pass on data. That includes the obvious things at the edges — computers, servers, phones, printers — and the equipment in between, such as switches, routers and gateways. If it has a network address and participates in moving data, it is a node.

Key takeaways

  • A node is any addressable device that sends, receives or forwards data on a network.
  • Endpoints are a subset of nodes: the ones at the edges that terminate connections rather than relaying them.
  • Infrastructure nodes are usually monitored differently from endpoints, using SNMP or similar rather than an installed agent.
  • Some monitoring tools bill per node, which makes switches and access points billable in a way endpoint pricing does not.
  • Counting nodes is how you find the devices nobody owns, which are reliably the least patched things in the estate.

Node, defined

In networking, a node is any device that has an address on the network and takes part in moving data. The term is deliberately broad because network diagrams need a word covering everything that appears as a labelled box: a laptop, a server, a printer, a switch, a firewall, a virtual machine, an access point.

The word comes from graph theory, where a network is nodes joined by edges. Networking borrowed the vocabulary directly, which is why the links between devices are sometimes called edges and the path across several devices is described as a number of hops.

Two families of node

Practically, nodes split into two groups, and the split determines how you monitor them.

  • End nodes: devices that originate or consume traffic. Computers, servers, phones, tablets, printers, tills. These usually accept a management agent.
  • Intermediate nodes: devices that forward traffic. Switches, routers, firewalls, access points, load balancers. These rarely accept an agent and are polled instead.

Node versus endpoint versus host

These three words are frequently swapped, which causes confusion in procurement and in incident reports. Node is the widest: anything addressable. Host narrows it to devices that run services or applications and have their own address, effectively end nodes. Endpoint narrows it further, to devices you manage and defend as part of your responsibility.

So a core switch is a node but not an endpoint. A staff laptop is all three. A virtual machine is a node, a host and, if you are responsible for patching it, an endpoint too.

Why node counts show up commercially

Network monitoring tools frequently price per node, and because infrastructure counts, the total climbs faster than an endpoint count does. A thirty-person office with three switches, six access points, a firewall and two servers has far more nodes than staff.

That is worth knowing before an evaluation. Ask whether a node is anything with an IP address, whether interfaces on a switch count separately, and whether a device that is polled but not managed still consumes a licence. The answers differ enough between vendors to change which tool is cheapest for your specific shape of estate.

Finding the nodes you did not know about

A node inventory is one of the highest-value hours an IT team can spend, because unowned nodes are where risk collects. The pattern is consistent: an old access point nobody removed, a test server still reachable, a printer with default credentials, a network camera on the same flat network as the finance machines.

A discovery scan of each subnet, reconciled against your asset list, produces a short and useful list of things that are present but unaccounted for. Every one of those is either a device to manage or a device to switch off.

Common mistakes

  • Treating node and endpoint as interchangeable in a licence discussion, then being billed for infrastructure you did not expect.
  • Discovering nodes once and never again, so the inventory is stale within a quarter.
  • Monitoring intermediate nodes for availability only, missing interface errors that explain intermittent user complaints.
  • Leaving discovered-but-unowned devices on the list without deciding to manage or remove them.

Frequently asked questions

What is a node in a computer network?
Any device with a network address that can send, receive or forward data. That includes computers, servers, phones and printers as well as switches, routers and access points.
Is a node the same as a device?
Close, but node specifically means a device participating in the network with an address. A powered-off laptop in a cupboard is a device but not currently a node.
What is the difference between a node and an endpoint?
Endpoints are the subset of nodes at the edges of the network that terminate connections, and that you are responsible for managing. Switches and routers are nodes but not endpoints.
What is a hop?
One step between two nodes along a path. A packet crossing three routers to reach its destination has taken three hops, and each one adds a little latency.
Do virtual machines count as nodes?
Yes. Each virtual machine has its own address and participates in the network independently, so it counts separately from the host it runs on.

How this works in 247connect

247connect focuses on the end nodes people actually work on, giving operators fast managed access to computers and servers without a per-device meter running in the background.

More IT explainers