“Data Sovereignty,” “Security,” and “Scalability” in a Sovereign AI Environment NTT docomo Business
NTT docomo Business is expanding its IoT-oriented NaaS, “docomo business SIGN,” which comes standard with security features, with an eye on the rapid spread of physical AI.
The company believes that IoT will gradually evolve into physical AI.
Physical AI refers to AI that controls physically operating machines such as robots, but physical AI will not progress with AI and physically operating machines alone.
Kazuo Komine, Executive Officer and General Manager of IoT & Physical AI Services at NTT docomo Business, says, “Manufacturing machines such as robots needs to be cost-effective, and compact size is desirable. Therefore, it’s not realistic to put all functions into the machine. We believe that the mainstream trend will be to limit the machine’s functions, such as shifting necessary functions to the AI infrastructure outside the machine.”
There are other reasons for shifting functions to the AI infrastructure.
As physical AI becomes more widespread, the number of cases where multiple machines work in coordination will increase. Controlling this coordination from the AI infrastructure side is more efficient.
“In preparation for this era, we are promoting an ‘As a Service’ model that allows for flexible use of physical AI functions. Security is paramount for physical AI, so we want to provide wide-area management of a large number of machines on top of the AI infrastructure while maintaining strong security,” says Mr. Komine.
As part of this, they are offering a sovereign AI environment combining MEC and GPUaaS, usable from a single GPU unit, as a menu item in “docomo business SIGN.”
The company emphasizes “data sovereignty,” “security,” and “scalability” in advancing the social implementation of physical AI.
While traditional on-premise solutions ensure data sovereignty and security, the high initial investment costs make scalability difficult. While cloud computing offers scalability, it presents challenges in ensuring security due to concerns about data being stored overseas and the complexity of security measures.
The sovereign AI environment is characterized by its combination of data sovereignty, security, and scalability.
Mr. Komine explains, “In terms of data sovereignty, we provide services without leaving our carrier network. Because it’s a carrier network, it’s a closed network, but we further enhance security by implementing threat detection. Customers can use it on demand, only when and as much as they need, making it easy to scale.”
Furthermore, a new “Pro” menu has been added to the “docomo business SIGN” lineup to support the operation of tens of thousands of IoT lines and global use. This enables integrated management of domestic and international IoT lines, automation of line operations, and system integration, achieving both advanced operational management and security measures.
Business models for IoT and physical AI generally start small, increasing the number of machines as the business gains momentum, and such expansion can be flexibly handled with Pro.
“As IoT and physical AI become widespread, managing tens or hundreds of thousands of machines will become impossible with manual labor. Pro is a service that addresses these challenges for our customers,” says Mr. Komine.
One of the key points is its compatibility with the next-generation eSIM standard for IoT, “SGP-32.” This allows for remote switching of networks and other settings without requiring operation on the machine itself.
For example, when exporting products from Japan to overseas, it’s possible to configure the system to switch to the carrier of the destination country before shipment.
When an instruction is sent from the eSIM management server, the instructed eSIM queries the network distribution server for information on the local carrier to be used in the destination country. The network distribution server responds, and the local carrier information is downloaded into the eSIM. As a result, the eSIM information switches to the local carrier. While the product is in Japan, it will be out of range and unable to communicate, but once it enters the local carrier’s area in the destination country, it will be able to communicate.
NTT docomo Business plans to promote Pro primarily to large global companies such as automobile manufacturers, construction equipment manufacturers, and camera manufacturers.
※Translating Japanese articles into English with AI
