A naukri.com initiative
Openstack
6d
279
Image Credit: Openstack
OpenStack Delivers Efficiency, Cost Savings for Powering AI Workloads
- FPT Smart Cloud, a Silver Member of OpenInfra Foundation, uses OpenStack to power AI workloads in Vietnam, compared with proprietary solutions like VMware and Microsoft Azure.
- The company has built a significant OpenStack footprint for AI workloads and high-performance computing, leveraging features like flexibility for customization and mature cloud ecosystem.
- FPT Smart Cloud utilizes OpenStack for managing AI factories in Vietnam and Japan, offering services like bare metal, GPU cloud instances, and Kubernetes provisioning.
- For storage, they employ SAN storage, Ceph, and local NVME, with performance improvements seen compared to VMware.
- In another use case, FPT AI eKYC offers secure digital onboarding and authentication to Vietnamese banks, demonstrating scalability and agility with OpenStack.
- By optimizing existing AI/HPC workloads with OpenStack, FPT achieved a 7% GPU efficiency increase and multi-million dollar savings.
- Technologies like NUMA, HugePage, CPU pinning, and NVMe storage contribute to GPU optimization and network/storage capacity enhancements.
- FPT leverages OpenStack features like DPDK, SR-IOV, and vGPU live migration to improve high availability for critical AI and HPC workloads.
- The 'OpenInfra for AI' Working Group, including members like FPT Smart Cloud, aims to promote OpenStack as the go-to open-source infrastructure technology for AI workloads.
- Collaboration within the OpenInfra community focuses on enhancing infrastructure for AI workloads and addressing software gaps to support AI scenarios.
- Joining the OpenInfra Foundation can support the efforts to advance OpenStack adoption for AI workloads.
Read Full Article
16 Likes
For uninterrupted reading, download the app