Distributed River Network Cache RiverCache: Network Speeds Up Tenfold in Remote Areas

RiverCache, jointly released by the University of São Paulo Networking Lab and network startup Inocra, is a distributed content cache network built along river routes. The network deploys low-cost cache nodes in small towns along rivers, pre-caching internet content close to users and addressing network latency issues in remote areas like the deep Amazon rainforest.

The core idea of RiverCache is to deploy cache nodes downstream. Villages in the Amazon rainforest are usually distributed along rivers, and the water distance between neighboring villages is much shorter than the straight-line distance, making traditional terrestrial network deployment extremely expensive. RiverCache uses small solar-powered cache stations (each about 35 kilograms), supplemented with power and software updates via river barges on a regular basis, allowing large amounts of content to be cached near users.

The pilot was conducted in 42 communities near Tabatinga in Amazonas State, Brazil, covering about 24,000 residents. Results show that after RiverCache deployment, average webpage load time in the communities dropped from 8 seconds to 0.8 seconds, and video start-up latency dropped from 14 seconds to 2 seconds. Marina Silva, founder of Inocra, said this kind of network design based on natural geography provides a low-cost connectivity solution for remote areas around the world that lack terrestrial fiber.