AI Data Cycle Storage
Western Digital unveiled a six-stage AI Data Cycle methodology that specifies the ideal storage mix for AI workloads at scale, sparking the subsequent wave of AI innovation. Customers may lower the total cost of ownership (TCO) of their AI workflows, increase productivity, and optimise their AI investments by using this framework to plan and build advanced storage infrastructures.
Artificial intelligence (AI) models process text, photos, audio, and video among other data types while also creating new, unique data in a continuous Data Cycle generation and consumption. Data storage systems must manage enormous volumes of data while providing the capacity and performance to support the computational loads and speeds needed for large, complex models as AI technologies evolve. Western Digital today launched an industry-leading, high-performance PCIe Gen5 SSD to support AI training and inference, a 64TB SSD for fast AI data lakes, and the world’s highest capacity ePMR, UltraSMR 32TB HDD, for cost-effective storage at scale.
Generative AI is the next transformative technology, and storage is essential. According to Ed Burns, Research Director at IDC, “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.” This is especially true when larger and higher-quality data sets become more common. Western Digital, a leader in Flash and HDD, can benefit from this increasing AI landscape with its strong market position and extensive portfolio, which covers a variety of needs across several AI data cycle phases.
“The engine of AI is data.” According to Rob Soderbery, executive vice president and general manager of Western Digital’s Flash Business Unit, storage is becoming more critical and dynamic as AI technologies are incorporated in almost every business. “Their commitment to providing unmatched value to their customers is demonstrated by the new AI Data Cycle framework, which will enable their customers to build a storage infrastructure that impacts the performance, scalability, and deployment of AI applications.”
Growing Business AI Data Cycle Storage Options for High-Processing Tasks
Western Digital’s first enterprise-class PCIe Gen 5.0 SSD, the Ultrastar DC SN861 SSD, boasts industry-leading random read speed and is expected to have the best power efficiency for AI applications. Its capacities reach up to 16TB, and its ultra-low latency and remarkable responsiveness make it ideal for large language model (LLM) training, inferencing, and the deployment of AI services. Compared to the previous generation, it offers up to 3x improvements in random read performance. Furthermore, the low power profile lowers total cost of ownership by delivering greater IOPS/Watt.
The growing demand of the AI market for high-speed accelerated computing coupled with low latency to support AI’s compute-intensive environments is addressed by the improved PCIe Gen5 capacity. The Ultrastar DC SN861 boasts a broad feature set, including support for NVMe 2.0 and OCP 2.0, 1 and 3 DWPD, and a 5-year limited warranty. It is designed for workloads that are mission-critical. SN861 E1.S, the Ultrastar DC, is now sampling. This month, the U.2 will start sampling, and in CQ3’24, it will start shipping in large quantities. Later this year, more information on the E1.S and E3.S form factors will be released.
The enhanced Ultrastar DC SN655 enterprise-class SSD series, intended for storage-intensive applications, is a perfect match for the Ultrastar DC SN861. With up to 64TB of capacity available, the new U.3 SSD solutions will enable quicker, larger data lakes as well as increased performance and capacity for AI data cycle preparation. The new DC SN655 variations are currently being sampled. When mass sales of the drives start later this year, more information about them will be made available.
The 32TB ePMR enterprise-class HDD with the largest capacity in the market is now being sampled by Western Digital for a limited number of clients. The new Ultrastar DC HC690 high-capacity UltraSMR HDD is specifically designed for big data cycle storage in hyperscale cloud and enterprise data centres. It will be an essential component of AI workflows where low TCO and large-scale data storage are critical. The new 32TB drive maintains exceptional dependability and reliability while delivering unmatched capacity with seamless certification and integration for quick deployment. It does this by leveraging tried-and-true architectures from generations of wildly successful products. Later this summer, more information on the drive will be accessible.
Every phase of the AI Data Cycle has distinct needs for computing and infrastructure. Western Digital is producing solutions that not only offer larger capacities but are also specifically designed to handle the exceptional performance and durability of next-generation AI workloads, thanks to their understanding of the dynamic interplay between AI and data storage, according to Soderbery. “Their mission is to assist clients in realising the revolutionary potential of artificial intelligence through ongoing innovation, a long-term strategy, and a growing portfolio.”
Concerning Western Digital
Western Digital is committed to utilising data to its fullest potential by maximising its potential. Underpinned by developments in memory technology, the Flash and HDD franchises bring about ground-breaking inventions and potent data storage solutions that help the world realise its dreams. They have committed to challenging carbon reduction targets authorised by the Science Based Targets programme, which is at the core of their principles. They also acknowledge the importance of combating climate change.
