Edge computing promises lower latency by reducing the distance data must travel to be processed. However, fiber optic cables transmit data at approximately 200 km/ms. A data center 100 km away adds just 1 ms to the response time.
Current 5G networks typically have latencies of 30–40 ms, with the very best private networks achieving about 10 ms. Data processing itself usually takes a few milliseconds, particularly when video compression is involved. Reducing response time by 1 ms by bringing compute closer to the device makes little meaningful difference.
Additionally, in most regions, there is likely a large city within the 1-ms range that already houses hyperscaler data centers. Those large-scale facilities offer better economies of scale, and hyperscalers are better than MNOs at selling compute services. As a result, today’s “edge” solutions largely mirror the traditional model, in which MNOs transport traffic to a peering point, where it is processed at a hyperscaler’s data center.
One exception is private 5G networks, which route traffic to the private network owner’s IT system. While this technically qualifies as edge computing, it is functionally just another form of peering, this time into a local IT network.
VIA EE Times Europe
Learn about the Internet Society’s 50/50 vision to keep Internet traffic local, efficient, and cost-effective