SAN JOSE, Calif., June 7, 2024 — Fueling the next wave of AI innovation, Western digital introduced a six-step AI data cycle framework that defines the optimal storage mix for large-scale AI workloads. This framework will help customers plan and develop advanced storage infrastructures to maximize their AI investments, improve efficiency, and reduce the total cost of ownership (TCO) of their AI workflows.
AI models operate in a continuous loop of data consumption and generation – processing text, images, audio and video, among other types of data, while simultaneously producing new, unique data. As AI technologies become more advanced, data storage systems must provide the capacity and performance needed to support the workloads and computing speeds required for large and sophisticated models, while still managing immense volumes of data.
Western Digital has strategically aligned its Flash and HDD technology and product roadmaps with the storage requirements of each critical stage of the cycle, and today introduced a new industry-leading high-performance PCIe Gen5 SSD to support training and AI inference; a high-capacity 64TB SSD for fast AI data lakes; and the world’s highest capacity 32TB UltraSMR ePMR hard drive for cost-effective storage at scale.
“There is no doubt that generative AI is the next transformational technology, and storage is a critical enabler. The implications for storage are expected to be significant, as the role of storage and access to data influences the speed, efficiency and accuracy of AI models, especially as larger and better data sets quality are becoming more prevalent,” said researcher Ed Burns. Director at IDC. “As a leader in Flash and hard drives, Western Digital has the opportunity to benefit from this growing AI landscape through its strong market position and broad portfolio, which addresses a variety of needs over the of the different stages of the AI data cycle.
“Data is the fuel of AI. As AI technologies are integrated into virtually every industry sector, storage has become an increasingly important and dynamic component of the AI technology stack,” said Rob Soderbery, executive vice president and director general manager of Western Digital’s Flash business unit. “The new AI Data Cycle framework will enable our customers to build storage infrastructure that will impact the performance, scalability and deployment of AI applications, underscoring our commitment to delivering unparalleled value to our customers .
Growing suite of enterprise AI storage solutions for compute and storage-intensive workloads
The new Ultrastar DC SN861 SSD is Western Digital’s first enterprise PCIe Gen 5.0 solution with industry-leading random read performance and best-in-class projected power efficiency for AI workloads. With capacities up to 16TB, it delivers up to 3x better random read performance than the previous generation with ultra-low latency and incredible responsiveness for training large language models (LLM), inference and deployment of AI services.
Additionally, the low power profile delivers higher IOPS/Watt, reducing overall total cost of ownership. The increase in PCIe Gen5 bandwidth addresses the AI market’s growing needs for high-speed accelerated computing combined with low latency to serve compute-intensive AI environments. Built for mission-critical workloads, the Ultrastar DC SN861 offers a rich feature set, including support for NVMe 2.0 and OCP 2.0, 1 and 3 DWPD, and a 5-year limited warranty1. The Ultrastar DC SN861 E1.S is now in sampling mode. The U.2 will begin sampling this month and begin volume shipments at CQ3’24. More details on the E1.S and E3.S form factors will follow later this year.
Complementing the Ultrastar DC SN861, the expanded family of Ultrastar DC SN655 enterprise-class SSDs are designed for storage-intensive applications. New options for U.3 SSD will reach up to 64TB, providing higher performance and capacity for AI data preparation and faster, larger data lakes. These new DC SN655 variants are now being sampled. More details on the drives will be released later this year when volume shipments begin.
Western Digital is also now testing the industry’s highest capacity 32TB ePMR enterprise-class hard drive for select customers. Designed for massive data storage in large-scale cloud and enterprise data centers, the new Ultrastar DC HC690 high-capacity UltraSMR hard drive will play a critical role in AI workflows where large-scale data storage Scale and low total cost of ownership are paramount. Building on proven designs from generations of successful products, the new 32TB drive delivers unmatched capacity with seamless qualification and integration for rapid deployment, while maintaining superior reliability and reliability. More details about the ride will be available later this summer.
“Each stage of the AI data cycle is unique and has different infrastructure and compute requirements. By understanding the dynamic interplay between AI and data storage, Western Digital delivers solutions that not only deliver higher capacities, but are also tailored to support the extreme performance and endurance of data workloads. “Next generation AI,” Soderbery said. “Through our growing portfolio, long-term roadmap and constant innovation, we aim to help our customers unlock the transformative capabilities of AI. »
For more information about AI Data Cycle and Western Digital’s AI Storage portfolio, please visit: Data storage for AI.
About Western Digital
Western Digital’s mission is to unlock the potential of data by harnessing the power to use it. With the Flash and HDD franchises, powered by advancements in memory technologies, we create groundbreaking innovations and powerful data storage solutions that enable the world to realize its aspirations. At the heart of our values, we recognize the urgency of tackling climate change and are committed to achieving ambitious carbon emissions reduction targets endorsed by the Science Based Targets initiative. Learn more about Western Digital and the Western Digital, SanDisk and WD brands at www.westerndigital.com.
Source: Western Digital
