IBM AI storage solutions improve efficiency and sustainability

The role of AI storage solutions is more critical than ever in today’s evolving artificial intelligence landscape.

As businesses must manage increasingly complex data sets and power-intensive workloads, innovations in storage technology are essential to support high performance computing and AIaccording to Sam Werner (photo), vice president of product management at IBM Corp. The transformative potential of intelligent storage systems is reshaping how organizations optimize AI workloads and improve overall performance.

IBM’s Sam Werner speaks with theCUBE about the impact of innovative storage solutions on AI workloads and the future of scalable and efficient data management.

“You need to be able to store all of this data in a very efficient way,” Werner said. “You have to have low power and cooling costs because you have to make way for all the very power-hungry graphics processing units needed for AI and supercomputing. First of all, you need to have very efficient storage. This needs to be super fast because you don’t want those GPUs sitting idle. They are very expensive. »

Werner spoke to CUBE Research Dave Vellante And John Furrier has SC24during an exclusive broadcast on theCUBE, SiliconANGLE Media’s live streaming studio. They discussed innovative storage solutions, their impact on AI workloads, and the future of scalable and efficient data management. (*Disclosure below.)

AI storage solutions for the modern era

IBM is evolving AI storage solutions with innovations such as Content-Oriented Storagewhich integrates intelligence directly into the system to optimize performance and reduce inefficiencies, according to Werner. This approach allows businesses to optimize data in real-time without duplicating data sets between systems, providing a significant advantage for scaling AI and high-performance workloads.

“We are completely changing this paradigm,” Werner said. “Rather than bringing all your data together and creating this vector database, what if we did all of that in storage, put accelerators in the storage products and actually did the vectorization as the data changes? We know when the data changes here, right? We can therefore vectorize it in real time.

Active File Management, another key feature of IBM Storage Scalesimplifies access to data across various storage systems, Werner noted. By providing a unified global namespace, businesses can manage unstructured data without copying everything into a single storage pool, streamlining operations and improving accessibility.

“Rather than copying everything into a single Uber storage pool… we have something in Scale called Active File Management,” he said. “Scale has a global namespace. We give you a single namespace that you can use anywhere and grow forever. Then we can communicate with all your unstructured data sources using our active file management and cache the data.

As organizations face increasing operational costs and sustainability challenges, energy efficiency is also an important priority for IBM. With its AI storage solutions, including the IBM Storage Scale System 6000, IBM delivers high performance while consuming significantly less power than competing systems, according to Werner.

“I challenge anyone watching this to go compare the kilowatts required to run our Scale System 6000 versus our competitors,” he said. “Look at the capacity you need to support… Nvidia has released the SuperPod or BasePOD specs. Look at our energy consumption compared to the competition. There are at least half of us and, in many cases, depending on the competitor, even more. It’s huge.

As AI workloads increase, IBM continues to prioritize ease of use and automation. Recognizing that the skills shortage is a challenge for many organizations, IBM is working to integrate AI capabilities into its systems to simplify management and improve efficiency, Werner said.

“We want to make sure that it’s not just about supporting AI; it’s about integrating AI into our products,” he said. “Automatic recovery, automatic problem diagnosis, telling you what problems need to be fixed without requiring a lot of skills to take care of them. »

Here’s the full video interview, part of SiliconANGLE and theCUBE Research’s coverage on SC24:

(*Disclosure: TheCUBE is a paid media partner for SC24. Neither Dell Technologies nor WekaIO Inc., the primary sponsors of theCUBE event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE .)

Photo: SiliconANGLE