Around the Storage Block

No science fiction here: Showcasing the new AI Data Node from HPE

Learn how HPE is working with WekaIO and Scality to bring the very real benefits of AI to your business with AI Data Node, integrated two-tiered data storage for deep learning and high-performance computing.

HPE_AI Data Node_blog.jpgI enjoy watching science fiction movies. I really watched a lot of them in the 90s and early 00s and it is quite incredible to witness a concept or a vivid imagination showcased in the movie translate into reality later in future. One such movie was Artificial Intelligence (AI), directed by Steven Spielberg and released in 2001. I enjoyed the concept but didn’t expect it to become a reality within a decade. AI has already impacted most of us in one way or another. For example, most of the latest smart phones have “AI-enabled” cameras. And this is having a big impact on how images are captured—and turning everyone into a professional photographer.

The evolution of AI is not limited to personal devices. Data is the new fuel driving accelerated levels of innovation powered by AI. Rapid AI evolution today is powered by leading-edge technologies like deep learning algorithms and GPU hardware that is equipped for performing massively parallel computation. Flash storage technology is utilized to provide the necessary throughput performance but can be quite costly for capacity storage. As companies go into production with AI, data sets will grow to tens and even hundreds of petabytes, and will exceed the capacity of traditional storage appliances. To achieve scalability and performance while simultaneously controlling costs, storage system designers build separate tiers of storage for hot and cold data, utilizing archival object storage for the colder data. This dramatically lowers the total cost of ownership.

So this could not be more timely: This week at NVIDIA’s GTC 2019 conference, we are showcasing a new AI Data Node solution in partnership with WekaIO and Scality. It’s designed for HPC and AI workloads using software-defined storage applications deployed on HPE ProLiant and Apollo servers.

AI Data Node from HPE

AI Data Node from HPE is a solution which combines HPE Scalable Object Storage with Scality RING and WekaIO Matrix™ parallel file System, running on HPE Apollo 4200 Gen10 Server. With the AI data node, you can have high-performance, petabyte-scale storage solutions with integrated data lifecycle management, providing tiering management by file system and a single namespace. This solution can be implemented in a classic two-tier architecture, with one tier dedicated to high-performance flash while a second tier provides scalable object storage, typically as two separate clusters of storage servers. A second hybrid approach combines both tier elements into a scalable cluster, utilizing storage servers that are optimized for both NVMe flash capacity and scale-out bulk data storage.

Building blocks of the AI Data Node

HPE Apollo 4200 Gen10 Server: The new HPE Apollo 4200 Gen10 Server builds on years of proven leadership with an architecture designed for big data and analytics, software-defined storage, backup and archive, and other data storage-intensive workloads. HPE Apollo 4200 Gen10 Server delivers high-density storage with scalability of up to 392TB of capacity in a 2U rack form factor. It’s unique easily serviceable 2U design can accommodate up to 28 LFF or 54 SFF hot-plug drives, driving accelerated performance with a balanced architecture and NVMe connected SSDs. HPE Apollo 4200 Gen10 also features HPE iLO 5 and HPE silicon root of trust technology for firmware protection, malware detection, and firmware recovery.

WekaIO Matrix: WekaIO Matrix is one of the fastest parallel file systems for AI and technical compute workloads that ensures applications never wait for data. WekaIO Matrix includes the MatrixFS flash-optimized parallel file system, qualified on HPE Apollo 4200 Gen10 server. Matrix is a storage solution that delivers the performance of all-flash arrays with the scalability and economics of the cloud. Matrix transforms NVMe-based flash storage, compute nodes, and interconnect fabrics into a high-performance, scale-out parallel storage system that is well suited for I/O-bound use cases. WekaIO Matrix provides automatic tiering of your cold data to Scality RING object storage to provide low cost and limitless scale.

Scality RING: Scality RING running on HPE Apollo 4000 System combines object storage software and industry-standard servers to provide low cost, reliable, flexible, centralized management that your organization needs for large-scale unstructured data. The HPE Scalable Object Storage with Scality RING creates a solution with a lower TCO than traditional SAN and NAS storage vendors at a petabyte scale, while providing greater data protection for current and future large-scale storage needs. The combination of the Scality RING and hyperscale HPE Apollo servers has been deployed in numerous projects as a future-proof solution for petabyte-scale object applications. With HPE continuing our development of ultra-dense compute nodes, we expect to be able to store even more data in less space. The combined solution brings the best of market-leading Scality RING file and object storage on market-leading hardware from HPE for a best-of-breed solution with appliance-like experience, without appliance-like restrictions.

Managing your data with the AI Data Node from HPE

The combination of NVMe connected SSDs and 24 LFF HDD capacity makes HPE Apollo 4200 Gen10 along with WeakIO and Scality RING an ideal platform for the AI Data Node solution. With the data lifecycle management features built into WekaIO Matrix, colder data elements are automatically identified and tiered to S3-compatible Scality RING object storage. The entire data set is protected with distributed data protection scheme across a cluster of servers. This hybrid solution is a full-function high performance AI file store with an integrated and durable low-cost object storage tier, offering savings of up to half the infrastructure and operational costs of traditional solutions that deploy two separate storage clusters.AI Data Node from HPE.jpg

All components of the AI Data Node solution can be purchased from HPE. What’s more, the solution has been tested by HPE engineering and HPE Pointnext offers multivendor support.

In summary, AI Data Node provides a building block for production AI storage, with the performance and capacity to scale and grow with real-world operations. The AI Data Node leverages the storage performance of HPE Apollo 4200 Gen10 along with the performance of WekaIO and the capacity and efficiency of Scality RING to provide a foundational storage building block for production AI scenarios.

To learn more, download the technical white paper detailing the reference configuration for AI Data Node.

If you want more details, please contact your local HPE/WekaIO/Scality sales team who will be happy to whiteboard an architecture tailored to your specific requirements—and to help you better understand how the new AI Data Node can be an integral part of your IT infrastructure.  

See you at GTC?

If you are attending GTC 2019 the week of March 18th, please visit HPE booth 1129 and chat with one of our solution experts—and watch the AI Data Node in action.

MeeAshwin Shetty HPE Storage.jpegt Around the Storage Block blogger Ashwin Shetty, WW Product Marketing Manager, HPE Storage. On Twitter: @ashwinkshetty



0 Kudos
About the Author


Our team of Hewlett Packard Enterprise storage experts helps you to dive deep into relevant infrastructure topics.