Node

A node is a single computer that belongs to a network and takes on a task there. The English word means “Knoten” in German, and that’s exactly how you can picture it: many computers are connected to one another, and every connection point in this web is a node. Such a computer is not a passive bystander. It receives data from other nodes, checks it according to fixed rules, and passes it on. The term itself says nothing about whether it’s a laptop, a server in a data center, or a small hobbyist device. What matters is only the role within the network, not the hardware.

Why thousands of nodes don’t need to belong to anyone

The appeal of many modern networks lies in the fact that they have no central authority. With a regular bank, the account list sits on the bank’s servers. If the bank goes down or someone alters the numbers there, there is no independent second source. With Bitcoin, by contrast, many thousands of nodes worldwide hold the same list of all transactions. This distribution is called decentralization.

An important property follows from this: an attacker would not need to manipulate one computer, but a large portion of all nodes simultaneously. That is practically almost impossible to achieve. It is just as difficult to shut down such a network. As long as enough nodes are still running somewhere, the system keeps working.

Anyone who runs their own node also doesn’t need to trust anyone. They verify every rule themselves using their own computing power. In the crypto scene there’s a saying for this: “Don’t trust, verify” – don’t trust, but check for yourself. Anyone who just uses an app, on the other hand, relies on a provider’s node.

What a node checks all day long

Technically, a node is simply a program that runs continuously and communicates with other computers. On startup, it looks for a few partners in the network, so-called peers. From them, it downloads the existing data history. With Bitcoin, that’s several hundred gigabytes, which fits on a normal hard drive but can take days to download.

After that, the actual work begins. When a new transaction arrives, the node checks fixed questions: Does the balance actually exist? Is the digital signature correct? Has the money not already been spent elsewhere? If a check fails, the node discards the data and does not pass it on.

There are different types to distinguish. A full node stores and verifies everything itself. A light node stores only a fraction and asks others for the rest – more economical, but less independent. A common misconception is confusing nodes with miners. Miners generate new blocks of data and require a great deal of electricity to do so. A simple node only verifies and runs on a device that uses barely more power than a light bulb.

The term outside the crypto world

In the news, the term is most commonly encountered in connection with cryptocurrencies. Reports like “the number of Ethereum nodes has declined” are seen as a warning sign, because the network then depends on fewer computers. Cloud service providers are also discussed in this context: if a large share of all nodes runs at the same company, decentralization is only apparent.

However, the word is used much more broadly. In data centers for artificial intelligence, a single server with multiple graphics cards is also called a node. Large training runs are then distributed across thousands of such nodes. In computer science generally, “node” also refers to a point in a data structure, for instance in a tree structure made up of branches.

In practice, this means: when reading a report, it’s worth asking which network is meant. A node in Bitcoin, a node in an AI cluster, and a node in a program’s execution flow only share the basic idea. Each is a single point within a larger context.