Innogrid Supplies GPU-Based AI Analytics Environment to Korea Institute of SMEs and Startups
TECHWORLD ·
✦ AI Summary
Innogrid announced on the 1st that it had supplied a private cloud platform and a cloud management platform for the Korea Institute of SMEs and Startups' "GPU-based AI big data analytics infrastructure build" project.
The Korea Institute of SMEs and Startups pursued the project to upgrade the analytics environment of the "Small and Medium Enterprise Big Data Platform (SIMS)," aiming to move away from an existing closed, on-premises, physical PC-centered structure to a GPU-based analytics environment.
Innogrid applied Openstackit and TabCloudit to build GPU-equipped VMs, a monitoring system, and scaling functions, and plans to pursue future migration and expansion into AI workloads as well.
Innogrid announced on the 1st that it had supplied a private cloud platform and a cloud management platform for the Korea Institute of SMEs and Startups' "GPU-based AI big data analytics infrastructure build" project. The Korea Institute of SMEs and Startups served as the lead organization for the project.
The Korea Institute of SMEs and Startups pursued the project to upgrade the analytics environment of the "Small and Medium Enterprise Big Data Platform (SIMS)." The existing structure was centered on a closed on-premises setup and physical PCs, and the build focused on creating a GPU-based analytics environment for policy effectiveness analysis and forecasting simulations.
In response, Innogrid applied its private cloud platform, Openstackit, and its cloud management platform, TabCloudit. The configuration was based on domestically developed virtualization software that had obtained a National Intelligence Service security function verification certificate, and it covered the scope of responding to public institutions' needs to transition to virtualized environments.
In the system room, Innogrid set up the equipment and network configuration and established a foundation for linking cloud servers and storage. It also applied snapshot and replication functions, introduced dynamic scaling, and enabled real-time monitoring of resource status. This improved operational stability and also secured future scalability.
It also built a virtual machine (VM) equipped with a GPU and supported a high-performance development VM capable of using multiple GPUs simultaneously. As a result, resources could be allocated according to the size and nature of each analytics task. GPU and computing resources for each research project could also be operated more flexibly.
In addition, it established a monitoring system that can check host and virtual server resource usage status in real time. This made it possible to continuously track resource conditions and also to predict when expansion will be needed based on usage trends. The system became a foundation for allocating resources by project and determining expansion needs.
Based on this project, Innogrid plans to expand its migration business. The targets are public-sector and corporate customers seeking to transition from foreign virtualization environments such as VMware to domestic platforms. After existing workloads are moved, it plans to gradually expand into private cloud services, and then further into multi-cloud and hybrid cloud environments. It also plans to gradually expand into GPU-based AI workloads after existing workload migration.
This strategy aligns with Innogrid's technology vision, "From xPU to AI Platform." Innogrid proposed a model that connects computing resources such as GPU, NPU, and CPU with AI development, training, deployment, and operations environments in a single system, saying the goal is to support infrastructure expansion when needed.
Kim Myeong-jin, CEO of Innogrid, said that the localization of public institution infrastructure is moving beyond replacing foreign virtualization products and into expanding AI infrastructure that uses GPU resources. He also said the company will continue to support the public sector's transition to virtualization and expansion of AI infrastructure.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407676
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.