Denver, Colorado - Premio in collaboration with California Institute of Technology (Caltech), will be showcasing a Software Defined Storage (SDS) proof of concept at SuperComputing Conference 2017, in Denver, Colorado. With the ever growing data storage requirements from science and research experiments, such as high energy physics at the CERN Large Hadron Collider, there is a need to have a robust distributed storage that can be scale in petabytes or even exabytes.
To solve this large scale storage requirement, Caltech and Premio have come together to collaborate on a proof of concept based on Premio’s FlacheStreams and ScaleStreams storage server hardware, along the Ceph open source file system. With this configuration, the proof of concept presents a unified, scalable distributed storage with high throughput performance that is reliable and flexible. With up to 1 PB (petabytes) of storage in a standard 42U rack, it has the capability to scale the capacity and performance as needed to in the future. Even with the sheer scale of the storage capacity, the goal of this proof of concept is to get the lowest dollar per gigabyte, hence minimizing total cost of ownership.
Professor Harvey Newman, leader of the Caltech high energy physics network and computing team said: “We are glad to continue our joint developments with Premio on cost effective, compact and scalable high throughput systems such the PoC being shown at SC17 this year. These are well matched to the needs of the worldwide high energy physics community working at the Large Hadron Collider, and of other data intensive science programs.”
Andy Lee, Product Manager at Premio, says, “In the era of the digital world, storage becomes a key component on how we store, access, and use our data. We are honored to collaborate with Caltech and are ready to turn this challenging proof of concept into reality. We believe with our our products will provide the perfect balance of performance, flexibility, availability, power consumption, and cost which are all crucial factors for this project. We are glad to be part of the research and science community with organizations such as Caltech, trying to resolve many challenges in the HPC world such as Terabit Data Transfer Node (DTN) to the future of ExaScale world.”
Premio will be showcasing their proof of concept demo with the open sourced Ceph Luminious version live at SC17 at Booth #663 November 13- 16, 2017 with the following hardware configurations:
- FlacheSAN1N10U-UN 1U 10 NVMe bays with AMD EPYC processor
- eDrawer4048S-UN 4U 48 bays storage server with AMD EPYC processor
- Micron 9200 1.6TB NVMe SSDs
- Mellanox ConnectX4 100GbE Network cards
- Mellanox Ethernet TOR Spectrum switch
About Premio, Inc.
Premio, Inc. is an industry-leading service provider for ODM/OEM computing technology. The firm goes to extraordinary lengths to solve the most formidable challenges faced by its customers. Premio does it by becoming more than just a partner to clients. Instead, Premio is determined to serve as its clients’ ‘inside outsource’ — an extension of an organization’s business, work culture, manufacturing processes and operations, modulating its solutions to answer special needs through speed, agility and precision. With state-of-the-art facilities in North America and Asia, Premio, Inc. partners with some of the world’s leading technology companies, including Intel, Microsoft, and Dell.
For more on Premio, click here.