SK Telecom said on the 18th that its proposal, "technical elements and interworking architecture for AI data centers," won final approval as an international standard at the Study Group 11 (SG11) meeting of the International Telecommunication Union (ITU) Telecommunication Standardization Sector (ITU-T) held earlier this month.
With an international standard established for the interworking architecture of AI data center systems, expectations are rising that the foundation for building and operating global AI infrastructure will be further strengthened.
The core of the new standard is systematizing the interworking architecture and signaling requirements among internal systems in AI data centers. As AI services spread, data centers are evolving into complex infrastructure where various functions—beyond large-scale computation—such as power, cooling, storage, security, and resource operations, work in tandem. Accordingly, the need for an international standard that organizes signal exchange and interworking methods among internal systems has grown.
The standard divides AI data centers into three layers—service layer, management layer, and infrastructure layer—and defines each layer's roles and functions, along with the signaling requirements for interworking between layers. In effect, it organizes into an international standard a structure in which different systems exchange status and control information and operate organically.
SK Telecom secured approval for a new related standardization item at the ITU-T in May 2024 and, after about two years of research and international cooperation discussions, led to this adoption.
Choi Dong-hee, head of AI strategy planning at SK Telecom, said, "It is meaningful that our accumulated AI data center technological capabilities and operational know-how have been recognized by a credible international organization," and added, "We will contribute to the expansion of the global AI ecosystem."