Low-Carbon Computing Platform from Old Smartphones
Google/DeepMind
University of California San Diego
Researchers at UC San Diego, with Google's support, are building a computing platform from retired smartphones. By extracting motherboards from old phones and clustering them, they aim to create a low-cost, low-carbon datacenter using 2,000 Pixel phones. This approach reduces embodied carbon by reusing hardware that would otherwise be discarded.
Researchers at the University of California San Diego, with support from Google, are developing a phone cluster computing platform that repurposes retired smartphones into a general-purpose computing resource. The process involves removing the motherboards from old phones (which contain the core compute functionality and account for about 50% of embodied carbon) and clustering them into groups of 25-50 devices, which together match the performance of a modern server. The phones run a general-purpose Linux OS instead of Android, and jobs are orchestrated with Kubernetes. The university plans to deploy a 2,000-phone datacenter using Pixel smartphones, which will support hundreds of students and researchers for cloud computing, reducing the need for new hardware. Early tests show that a 20-phone cluster can handle grading for a class of 75+ students with latencies below AWS's t3.micro instance. The full system is expected to launch in Fall 2026.
- Сокращения
- OS = Operating System — операционная система
- vCPU = Virtual Central Processing Unit — виртуальный центральный процессор
- GPU = Graphics Processing Unit — графический процессор
- SPEC = Standard Performance Evaluation Corporation — Standard Performance Evaluation Corporation
Source: Google Research —
original