NewsOpenAI / ChatGPT / Artificial Intelligence

Huawei Releases Quick Storage Solution for AI Model Trainers OceanStor A310

Huawei enters the artificial intelligence storage field using its OceanStor A310, demonstrating the potential solution to industry-vast data processing challenges.

Last week, Huawei’s artificial intelligence (AI) storage model OceanStor 310 went live. This marked an effort to solve some industry problems linked to large model applications. 

OceanStor 310 is designed for the large AI models age and seeks to offer a storage solution for industry model training, basic model training, and implication in segmented scenario models.

Huawei Debut AI Storage Tool

The OceanStor 310 should be imagined as a highly effective librarian in an extensive digital library, speedily fetching pieces of data. This is compared to another system, for instance, IBM’s ESS 3500, a more sluggish librarian.


 Quick information fetching by OceanStor 310 means faster working by artificial intelligence applications, resulting in the prompt making of intelligent decisions. The immediate data access is what results in the OceanStor 310 being outstanding.

The OceanStor 310’s supremacy seems to be linked to its capability to hasten data processing for artificial intelligence.

Compared to IBM’s ESS 3500, the most recent all-glitzy collection by Huawei allegedly feeds Nvidia’s graphic processing units (GPUs) nearly four times quicker on a per-rack unit basis. 

That is based on the methodology utilizing Magnum GPU Direct by Nvidia, which entails the direct sending of information from an NVMe storage resource to the graphic processing units in the absence of a storage host system.

📰 Also read:  Ava Labs Lays Off 12% of Staff to Support Growth of Avalanche Ecosystem

Features of OceanStor 310 

OceanStor 310 showed a performance whose write bandwidth is 208 gigabits per second (Gbps), while the sequential read bandwidth is 400 gigabits per second. 

Nevertheless, the extent to which these numbers are affected by open-source and closed-source structures is uncertain. Concerning its mechanism, the OceanStor A310 is developed as a profound learning information lake storage solution to provide limitless horizontal adjustability and outstanding performance for combined workloads.

In a statement, Evangeline Wang, a Product Management and Marketing Expert at Huawei, said they know that regarding artificial intelligence apps, enhancing the efficacy for AI model trading remains the most significant problem. 

She added that continuous information feeding to the CPU and GPUs is the storage system’s most significant problem during AI training. The storage system must offer excellent performance.

To address this problem, up to 96 processors, NVMe SSDs, and a memory cache can be supported in every OceanStor. Up to 4096 A310s can be clustered up, which entails sharing a universal file system capable of supporting standard application protocols.

Significance of Training AI Models 

The OceanStor 310 seeks to use SmartNICs and an immensely parallel design to reduce the time required to transmit data. A benchmark study comparing Huawei’s solution with its direct rivals established that in sequential writing and reading, Huawei’s A310, with its minute modes, was the quickest.  

The OceanStor 310’s introduction comes amid the artificial industry seeking effective data processing and storage solutions. This initiative by Huawei seeks to solve several present problems and might play a role in attaining properly organized artificial intelligence model training.

📰 Also read:  AI Tokens: What Are They And How Do They Work?

Nevertheless, the U.S. ban on Huawei, mainly because of concerns linked to national security and the firm’s supposed links to the Chinese government, includes another layer of intricacy to the firm’s undertaking in artificial intelligence storage solutions.

The likely ramifications on OceanStor 310 might be prominent. By providing a solution to some data storage and processing inadequacies, the firm is trying to challenge the rest of the suppliers. It also attempts to push the artificial intelligence industry towards another age of probable efficacy and innovation.

Editorial credit: Koshiro K /

Tokenhell produces content exposure for over 5,000 crypto companies and you can be one of them too! Contact at if you have any questions. Cryptocurrencies are highly volatile, conduct your own research before making any investment decisions. Some of the posts on this website are guest posts or paid posts that are not written by Tokenhell authors (namely Crypto Cable , Sponsored Articles and Press Release content) and the views expressed in these types of posts do not reflect the views of this website. Tokenhell is not responsible for the content, accuracy, quality, advertising, products or any other content or banners (ad space) posted on the site. Read full terms and conditions / disclaimer.

📰 Also read:  AI Tokens: What Are They And How Do They Work?


Stephen Causby

Stephen Causby is an experienced crypto journalist who writes for Tokenhell. He is passionate for coverage in crypto news, blockchain, DeFi, and NFT.

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button
Skip to content