NTT - Nippon Telegraph & Telephone Corporation

05/30/2024 | Press release | Distributed by Public on 05/30/2024 03:02

May 30, 2024 The Cutting Edge: NTT's Faster Data Analysis

Latency, or the time delay between data being generated and being processed, is a critical issue in many modern applications. For instance, in autonomous driving, even a small delay in data processing can be the difference between a safe journey and a potential accident. Traditional cloud computing systems, where data is sent to central data centers, struggle with this delay because of the physical distance data must travel. Additionally, processing huge amounts of data requires a lot of power, leading to higher operational costs and increased environmental impact.

NTT, in collaboration with open source software producer Red Hat, has developed a solution that addresses the challenges of latency and power consumption in data analysis, particularly important in environments dependent on real-time decision-making, such as autonomous vehicles and smart city infrastructures.

The research partners ran proof-of-concept trials for a real-time AI analysis platform in Yokosuka City and Musashino City, Japan. They demonstrated a 60% reduction in latency and a 40% decrease in power consumption per traffic monitoring camera compared to traditional methods. Their platform supports scaling up GPUs for more cameras without overloading the CPU, potentially cutting power use by another 60% for 1,000 cameras.

NTT's solution leverages edge computing, in which data processing is performed close to where data is generated, minimizing the distance data needs to travel, cutting down on delay and allowing for quicker responses to real-time data. By taking advantage of its All-Photonics Network (APN), which utilizes light for swift data transmission, and incorporating Remote Direct Memory Access (RDMA), the setup supports the rapid collection and processing of extensive sensor data at the source. This method not only speeds up data handling, but also lessens network load and cuts power consumption.

The solution is built using Red Hat OpenShift, a platform that supports container orchestration, enabling the flexible operation of processing workloads across distributed and remote locations. This flexibility means that as conditions or demands change at different data points, the system can adapt quickly without the need for a complete overhaul or significant downtime.

The nature of the research partners' edge computing framework opens the door for large-scale AI use. Implementing AI systems can be slow and complex, and may involve higher maintenance costs and the time to integrate new AI models and hardware. However, with the advent of edge computing in increasingly remote locations, AI analysis can be positioned nearer to the sensors, thereby reducing latency and enhancing bandwidth. This transforms the potential for instant, local data processing. In smart cities, sensors that monitor everything from traffic to environmental conditions can use AI to analyze data and make real-time changes in traffic flow, where emergency services could be automatically alerted to accidents as soon as they occur, without waiting for data to be relayed back and forth to distant servers.

In terms of environmental impact, reducing the power required for data processing can contribute to sustainability goals. By optimizing how and where data is processed, NTT's solution not only improves efficiency, but also supports the broader adoption of green technologies.

NTT's edge computing solution sets the stage for more intelligent, efficient, and responsive technology applications across various sectors. By bringing the processing power closer to the source of data, it paves the way for innovations that could transform how we manage and interact with our environment.

For further information, please see this link:

If you have any questions on the content of this article, please contact:
Public Relations Department of Nippon Telegraph and Telephone Corporation
[email protected]

NTT-Innovating the Future