Artificial intelligence is revolutionizing the media and entertainment industry, particularly through AI-driven media storage solutions, processing and cloud integration advancements. This includes sports teams using AI to analyze statistics and game developers using the technology to create new scenarios on the fly.
The times are changing when it comes to media and entertainment. That brings with it various storage demands when it comes to meeting modern workloads, according to Sherry Lin (pictured, front row, left), storage solution product manager at Super Micro Computer Inc.
Sherry Lin of Super Micro Computer; Skip Levens of Quantum; Paul Blinzer of AMD; and Praveen Midha of Western Digital talk with theCUBE.
“We see higher performance needs as gen AI has been commonly used in content creation and also post-production editing like animations, real-time rendering, virtual stage, virtual actor,” Lin said. “Storage [servers] must be able to handle those intensive tasks in order [to handle] AI-driven workloads.”
Lin; Paul Blinzer (back row, left), senior fellow and system software architect at Advanced Micro Devices Inc.; Praveen Midha (back row, right), director of segment and technical marketing of data center flash at Western Digital Corp.; and Skip Levens (front row, right), product marketing, AI strategy, at Quantum Corp., spoke with theCUBE Research’s Rob Strechay at the Supermicro Open Storage Summit series, brought to you by Supermicro, Intel, AMD and Nvidia, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed increasing demands for high-performance storage, the critical role of computational processing and AI-driven media storage solutions. (* Disclosure below.)
AI-driven media storage aims to fuel continuous content creation
Data, of course, is key in this new era of AI. While AI systems process and analyze data, they also create new content, according to Midha.
“In fact, existing repositories are now becoming even more valuable, because they are being used to feed these large training and influence models,” he said. “I think the fundamental assertion we are making here is that we are in this virtual data cycle where there’s existing data; we are using it to empower larger and faster models to train.”
That’s being used for inferencing, which is generating even more content. That content is being used as a feedback loop to create a new set of data on which an AI model is being trained on, Midha added.
At the end of the day, the data and the models have to sit somewhere. That’s where Quantum comes into the picture, according to Levens.
“Quantum is building a storage platform, not only for content production, but also for what this generation of AI and ML content enhancement workflows looks like,” Levens said. “We’re fortunate enough to work with some amazing customers, and each of these customers do what we call continuous content production, meaning there’s a deadline and you cannot fail.”
Those companies have communicated that they need to simplify what they’re doing, according to Levens. The company’s file system, called Myriad, is intended to address this.
“We’re really proud to use the Supermicro systems as a partner and Paul’s chips inside. So, we built the system that is NVMe, RDMA, point-to-point communication, but it integrates Layer 3 Intelligent Fabric networking,” he said. “It’s not just a file system; it’s a file system, and it serves everything in one. Call it the hot landing zone for data.”
AI-driven media storage solutions must support scalability
With the rise of AI in media and entertainment, opportunities for AI-driven media storage solutions arise. But the demands on storage must also be tended to, according to Blinzer.
“You need a solid foundation in the platform with the performance, scalability and openness, because if your content is locked in a proprietary API or a proprietary model, then it’s not your content anymore,” he said.
It’s also important to consider where the work gets done for media and entertainment. For many companies, the definition of what is edge and what is cloud starts to get blurred, according to Levens.
“These customers need to process wherever. It could be at the edge. Let’s say London captures it, moves it to an object store, but then perhaps L.A. needs to pick it up,” he said. “It’s really a matter of blending all of these. My favorite word for it is agility … if you have a system that kind of locks you into a certain way of working, that may work today, it won’t work tomorrow because this stuff is changing so, so quickly.”
Exponential growth needs to be supported by the platform, Blinzer added. That means the software and the platform must be able to scale.
“You have to have enough reserves that you are not locked in into this generation and then … left out,” he said. “So, the scalability and the flexibility to have your platform going from server cloud to the workstation to … edge on client scenarios is the key item.”
Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of the Supermicro Open Storage Summit series.
(* Disclosure: TheCUBE is a paid media partner for the Supermicro Open Storage Summit series. Neither Super Micro Computer Inc., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)