Cloudian - Bringing data closer to AI than ever before.

Cloudian's storage systems, co-founded by an MIT alumnus, are helping businesses deliver data at scale to data-intensive AI models and agents.

A scalable storage system for businesses, enabling seamless data flow between storage and AI models. Photo: Midjourney

Artificial intelligence is changing how businesses store and access data. This is because traditional data storage systems were designed to handle simple commands from a few users simultaneously, whereas today's AI systems with millions of agents need to access and process massive amounts of data continuously and in parallel.

Traditional storage systems now have many complex layers, slowing down AI because data has to pass through multiple layers before reaching the GPU—the graphics processor, considered the "brain" of AI.

Cloudian – co-founded by Michael Tso (from MIT) and Hiroshi Ohta – is helping data storage keep pace with the AI ​​revolution. The company has developed a scalable storage system for enterprises, enabling seamless data flow between storage and AI models.

This system reduces complexity by applying parallel computing to storage, consolidating AI functionality and data on a single parallel processing platform capable of storing, retrieving, and processing large-scale datasets, with high-speed direct connectivity between storage and both GPUs and CPUs.

Cloudian's integrated storage and computing platform simplifies the development of commercial-scale AI tools while providing businesses with a storage infrastructure capable of keeping pace with the AI ​​boom.

“One thing people often overlook about AI is that it’s all about data,” Tso said. “You can’t increase AI performance by 10% just by having 10% more data, not even 10 times the data—you need 1,000 times the data. Storing data in a way that’s easy to manage, while embedding computations right within it so that it can be processed as soon as it’s input without having to move it—that’s the direction the industry is heading.”

Object storage and AI

Currently, Cloudian's platform uses an object storage architecture, where all types of data—documents, videos , sensor data—are stored as single objects with metadata. Object storage can manage massive amounts of data in a flat structure, ideal for unstructured data and AI systems, but previously it was impossible to send data directly to the AI ​​model without first copying it to computer memory—causing latency and high power consumption.

Last July, Cloudian announced it had expanded its object storage system with a vector database, storing data in a format ready for immediate AI use. Once data is loaded, Cloudian performs real-time vector computations of the data to support AI tools such as recommendation engines, search engines, and AI assistants.

Cloudian also announced a partnership with NVIDIA to develop a storage system that works directly with the company's GPUs. Cloudian says this new system enables faster AI processing and reduces computing costs.

“NVIDIA contacted us about 1.5 years ago because GPUs are only useful when there’s a continuous stream of data ‘feeding’ them,” Tso said. “Now people realize it’s easier to bring AI to data than to move massive data blocks. Our storage system integrates many AI functions, so we can pre-process and post-process data near where we collect and store it.”

AI-preferred storage

Cloudian is helping approximately 1,000 businesses worldwide to maximize the value of their data, including major manufacturers, financial institutions, healthcare facilities, and government agencies.

For example, Cloudian's storage platform is supporting a major automaker using AI to determine when maintenance is needed on individual production robots. Cloudian also collaborates with the U.S. National Library of Medicine to store research papers and patents, and with the National Cancer Database to store tumor DNA sequences—rich datasets that AI can process to support research into new treatments or discoveries.

“GPUs are a fantastic driving force,” Tso said. “Moore’s Law doubles computing power every two years, but GPUs can parallelize tasks on a chip, connect multiple GPUs together, and go far beyond Moore’s Law. This scale is pushing AI to new levels of intelligence, but the only way for GPUs to work at full capacity is to deliver data at a speed that matches their computing power – and the only way to do that is to eliminate all the intermediate layers between the GPU and your data.”

(According to MIT)

Source: https://vietnamnet.vn/cloudian-dua-du-lieu-den-gan-ai-hon-bao-gio-het-2433241.html