NTT’s IOWN DCI Rack Solution Launches in Japan: Optical Tech to Resolve AI Data Center Power, Heat, and Latency Issues

PR TIMES Japan

Overview

Itochu Techno-Solutions (CTC), APRESIA Systems, and Edgecore Networks Corporation have partnered to deploy the “IOWN DCI Rack Solution” in Japan, based on NTT’s IOWN concept. This solution fundamentally redefines data center connectivity by replacing traditional electrical signaling with advanced optical technology to address critical challenges such as power consumption, heat generation, and communication latency in the AI era. Utilizing Edgecore’s Nexvec™ platform, it enables efficient optical interconnection of resources like GPUs and memory.

In Depth

Addressing AI-Era Data Center Challenges

The rapid proliferation of generative AI has presented massive AI data centers with severe challenges concerning power consumption, heat generation, and data transmission latency. Traditional electrical wiring methods are approaching their limits in addressing these issues, necessitating a fundamental transformation of infrastructure. In response, the all-photonic technology central to NTT’s IOWN (Innovative Optical and Wireless Network) concept is gaining attention as a promising solution.

Domestic Rollout of “IOWN DCI Rack Solution”

Itochu Techno-Solutions (CTC), APRESIA Systems, and Edgecore Networks Corporation have initiated a strategic collaboration for the domestic deployment of the “IOWN DCI Rack Solution,” which is built upon NTT’s IOWN concept to resolve AI-era infrastructure challenges. This solution aims to entirely revamp the traditional electrical signal-centric connectivity and transmission within data centers using advanced optical technology.

Technical Features and Anticipated Benefits

At the core of the IOWN DCI Rack Solution is Edgecore Networks’ Nexvec™ platform. This foundation enables an architecture where critical computing resources such as GPUs and memory for AI processing can be efficiently connected via optics. This is expected to bring about the following fundamental improvements:

• Reduced Power Consumption: Minimizes electrical-to-optical conversion losses, significantly improving power efficiency in data transfer.
• Heat Suppression: Decreases heat generation from cables, lowering data center cooling costs.
• Lower Latency: Maximizes the transmission speed of optical signals, drastically shortening communication latency between GPUs.

This solution embodies the “Watt-Bit linkage” concept, a core tenet of the IOWN vision that holistically optimizes power and information. Its domestic deployment is set to bolster Japan’s AI infrastructure competitiveness and potentially serve as a standard model for next-generation data centers. Future challenges include ensuring compatibility with existing infrastructure, fostering skilled technical personnel, and optimizing initial investment costs, though the robust collaborative framework is expected to accelerate its adoption.