SK Telecom Launches AI Cloud Manager to Maximise GPU Usage

SK Telecom (SKT) has launched its AI-based B2B solution, ‘SKT Enterprise AI Cloud Manager,’ which optimises AI infrastructure resources and enables the integrated management of AI development environments. AI Cloud Manager is a commercial product developed from SKT’s expertise in managing and operating large-scale GPU resources.

• Make Telecom Talk My Trusted Source
• 
• 

Also Read: SK Telecom to Monitor Fibre Network Status With AI

AI Cloud Manager

AI Cloud Manager is a GPU cluster-specialised solution based on an AI job scheduler, designed to enhance GPU resource efficiency. By managing numerous GPU resources as if they were a single computer, the solution maximises GPU performance and reduces AI development learning time. SKT noted that when multiple jobs are requested in a cloud environment, the AI Job Scheduler efficiently allocates limited resources to ensure smooth task execution.

Real-time Monitoring and Resource Optimisation

As large-scale data learning requires GPU resources, optimising these resources is essential for AI cloud solutions. For example, GPU usage and performance are monitored in real-time for each project, and unused GPUs are detected and reassigned to other projects, SKT explained. Additionally, resource efficiency is improved by prioritising the allocation of cloud resources to high-priority projects.

Beyond optimising GPU usage, AI Cloud Manager offers a comprehensive MLOps (Machine Learning Operations) environment, supporting all stages of AI development—from data preprocessing and model training to deployment and inference.

According to SKT, the platform supports large-scale parallel learning, enhancing speed by distributing data across multiple computers. Its web-based interface (Web UI) allows multiple developers to collaborate on projects without needing specific software installations.