Edge computing: Computing close to the user

In short: Programs do not run in a distant data center but on servers or devices as close as possible to the users or the data.

In more detail: With edge computing, computing power is moved to the “edge” of the network: to many small sites worldwide, to mobile base stations, to factory floors or directly onto devices. The goals are shorter response times, less data transferred and operation even with a poor connection.

In Depth

Why proximity helps

Every request takes time because signals travel at a finite speed: a round trip from Germany to a data center overseas costs more than 100 milliseconds from distance alone. If the code runs nearby, this latency drops markedly. In addition, raw data, such as a camera’s video streams, does not have to go to the cloud in full but can be evaluated on site and only the result transmitted.

Typical uses

• Web: functions at the CDN edge (edge functions) that redirect requests, personalise pages or check access before they reach the origin server.
• Industry and transport: machine or vehicle data is processed locally in real time.
• Smart home and IoT: devices react even without internet.
• AI: small models run directly on a phone or camera.

Limits

Edge environments usually offer less memory, compute time and functionality than a full cloud region. Data is spread across many places, which makes consistency and updating (updates, security) harder. A central database then often stays in one region, and the advantage of proximity is lost again with every database access unless you replicate or cache deliberately.

Distinction

Edge computing complements the cloud rather than replacing it. A typical split: decisions with short reaction time at the edge; training, storage and analysis of large amounts of data centrally.

See also: CDN, Serverless and FaaS, Cloud, Anycast and load distribution