Streaming

Streaming is a way of transmitting data over the internet. Instead of loading an entire file completely first, it is broken down into many small pieces. These pieces arrive one after another, and the first ones are already used while the later ones are still on their way. With a movie, this means: you see the picture after two seconds, even though the last minute hasn’t reached you at all yet. The opposite of this is the classic download, where the file lands completely on the device before anything happens. Streaming therefore feels immediate, but usually leaves nothing permanent on the hard drive.

Why no one waits for the loading bar anymore

Waiting time is the most expensive currency on the net. A movie file in good quality can be several gigabytes in size. As a download, that would take many minutes depending on the connection. With streaming, the same movie starts within seconds. This exact difference is what made services like Netflix, Spotify, or YouTube possible in the first place.

For providers, this has a second advantage. The file is never fully present on the user’s device, only in short excerpts in the cache. This makes copying harder and enables subscription models instead of individual sales. The music industry has completely transformed because of this: before 2010, downloads were sold, today most revenue comes from streaming subscriptions.

The principle also plays a role in artificial intelligence. When a chatbot responds, the text usually appears word by word. That is streaming: the program sends off each finished piece of text immediately. The entire answer might take ten seconds, but the first word only half a second. Users therefore rate the response time noticeably better, even though nothing has changed about the actual computing time.

Buffers, chunks, and shifting quality

Technically, streaming relies mainly on a cache, the buffer. Your device always loads a bit of a supply ahead, usually a few seconds of video. Playback happens from this supply, while more is loaded in the background at the same time. As long as new supply arrives faster than it is consumed, everything runs smoothly. If the buffer runs empty, the picture stalls — that’s the infamous spinning loading circle.

Modern services additionally adapt to the connection. The video is available on the server in several quality levels, broken down into segments of a few seconds each. The player continuously measures how fast the data is arriving. With a poor connection, it simply fetches the next chunk in lower resolution. That’s why a YouTube video on a train suddenly gets blurry instead of freezing completely. Experts call this adaptive streaming, meaning streaming that adjusts itself.

A special case is live streaming, for example during a soccer match. Here there is no supply that could be made arbitrarily large, because the event is happening right now. The buffer stays small, and that’s exactly why live broadcasts stutter more often. The delay between the stadium and the screen ranges from a few seconds to almost a minute, depending on the technology.

From the living room to the quarterly figures

In everyday life you encounter streaming constantly, often without the word ever coming up. Music apps, series portals, Twitch, video calls, radio in the browser — all streams. Even the video feed from a doorbell camera to your phone works on the same principle.

In business news, the term usually appears as a business model. Then it’s about subscriber numbers, licensing costs for movies, or the so-called streaming war between Netflix, Disney, and Amazon. A typical figure is the time users spend per day with a service.

A common misconception is that streaming is automatically more efficient than downloading. The opposite can be true: if you stream an episode of a series five times, you transmit the data five times over. A download would only need it once. For the data centers behind it, streaming is therefore an enormous cost factor — and a noticeable share of global data traffic.