Cloudian and Lenovo today announced they’re teaming up to deliver a new HyperStore cluster designed to run big data, AI, and HPC workloads.
Each HyperStore cluster will be composed of six Lenovo ThinkSystem SR635 V3 servers equipped with AMD EPYC 9454P processors and flash drives. The cluster will come pre-loaded with Cloudian’s S3-comptable object storage system.
The combination of AMD processors and all-flash storage will allow the HyperStore cluster to read data at speeds up to 28.7 GB/s reads and write data at speeds up to 18.4 GB/s. According to Cloudian, testing shows the all-flash setup is 74% more efficient power-wise than an equivalent HDD setup.
HyperStore can serve as a data lake to power next-generation AI and advanced analytic workloads running atop distributed architectures such as PyTorch, TensorFlow, Apache Kafka, and Druid across media, finance, and life sciences, Cloudian says.
Lenovo’s EPYC-based processors “perfectly complement Cloudian’s high-performance data platform software,” says Cloudian CEO and Co-founder Michael Tso.
“Together, they deliver the limitlessly scalable, performant, and efficient foundation that AI and data analytics workloads require,” Tso said in a press release. “For organizations looking to innovate or drive research and discovery with AI, ML, and HPC, this solution promises to be transformative.”
Lenovo General Manager Stuart McRae says the combination of its all-flash EPYC severs with Cloudian’s software is capable of handling the most demanding AI and analytics workloads. “This partnership enables us to offer our customers a cutting-edge, scalable, and secure platform that will help them accelerate their AI initiatives and drive innovation,” McRae says in a press release.
Cloudian, which has more than 800 customers, says the HyperStore contains military-grade security and an “object lock” technology for ransomware protection.
Related Items:
Teradata Taps Cloudian for On-Prem Lakehouse
Cloudian Bolsters Object Store with Monitoring and Analytics
Cloudian, Splunk Unchain Compute, Storage
Source link
lol
