Innogrid, a cloud computing and AI infrastructure company, said on Sept. 3 it launched Openstackit 3.5, a private cloud solution that upgrades its in-house virtualisation technology and cloud operations capabilities.

The company said Openstackit 3.5 focuses on improving stability and operating efficiency for large-scale virtualisation environments. Innogrid said it refactored the platform overall. It said the upgrade goes beyond a basic underlying technology update and enhances its in-house virtualisation technology by reflecting operational experience and requirements accumulated in real customer environments.

It said a "maintenance mode" enables VMs on nodes scheduled for work to be moved in advance to other nodes and restored after the work is completed. It also said an "instance optimised rebalancing" feature allows instances to be distributed evenly by taking into account resource usage by each compute node.

It also strengthened an "instance HA" feature for service continuity. It supports service continuity even when failures occur by automatically moving VMs running on a node to another 정상 node when a failure on a specific node is detected.

It also upgraded GPU support in line with growing demand for AI infrastructure. It strengthened GPU resource provisioning and allocation, real-time monitoring and scheduling functions, supporting customers building AI training and inference environments to use high-performance computing resources more efficiently.

Ban Il-kwon (반일권), head of the cloud infrastructure centre at Innogrid's R&D headquarters, said, "Recently, the virtualisation market is changing beyond the level of replacing a specific hypervisor, into competition in infrastructure that can accommodate stable operations and migration as well as GPUs and AI workloads in the future." He said, "Openstackit 3.5 is the result of further strengthening stability, operating efficiency and scalability by reflecting Innogrid's accumulated in-house virtualisation technology and on-site customer requirements."