TOKYO, Oct 08 (News On Japan) - NTT Communications announced on the 7th that it successfully conducted a demonstration experiment to connect two data centers using the next-generation communication infrastructure, IOWN, to handle the integrated processing of generative AI learning.
The experiment linked the GPUs (graphics processing semiconductor servers) of data centers 40 kilometers apart using IOWN, which allows for high-capacity and low-latency communication. As a result, it achieved almost the same processing time as when handled by a single data center.
Currently, more than 80% of data centers are concentrated in the Tokyo and Osaka areas. The rapid increase in the development and operation of generative AI has led to server congestion and challenges related to power supply limitations. NTT Communications aims to further develop this technology to promote the distribution of data servers to regional areas.
Source: テレ東BIZ