Huawei | OceanStor Dorado 3000 All-Flash Storage System

All-Flash Storage System


The Huawei OceanStor Dorado 3000 is an affordable, all-flash storage system. Its cutting-edge hardware platform, FlashLink® intelligent algorithms, and end-to-end (E2E) NVMe architecture combine to boost performance by 50% over its predecessor and reduce latency to as little as 0.05 milliseconds. In order to make storage more intelligent during application activities, intelligent algorithms are embedded into the system.

Simplified operations and maintenance (O&M) are another benefit of the active-active (A-A) architecture and graphical user interface (GUI) design. SMBs in the telecommunications, financial services, government, manufacturing, and other sectors can put their trust in Huawei OceanStor Dorado 3000 all-flash storage because of its superior performance in virtualization, office automation (OA), and branch offices. The storage system’s reasonably priced services make it a good fit for the information technology needs of small and medium-sized businesses.


Product Features


Innovative Hardware Platform

Huawei’s storage hardware platform provides end-to-end (E2E) data acceleration, double the speed of the system’s performance compared to its predecessor.

  • Front-end access performance is increased by 20% thanks to the clever multi-protocol interface module, which takes over the protocol parsing duties previously handled by the general-purpose CPU.
  • The computing platform outperforms the market standard by 25% in terms of processing speed.
  • The intelligent accelerator module uses machine learning frameworks to study and comprehend the I/O rules of various application models, allowing for intelligent prefetching of memory. This results in a 50% increase in the hit ratio of the read cache.
  • Flash Translation Layer (FTL) is the central technique in the smart SSD, which speeds up data access in SSDs and cuts write latency in half.
  • The fault recovery time is reduced from 2 hours to just 10 minutes because to the intelligent hardware’s in-built Huawei storage fault library, which speeds up component fault location and diagnosis.


Intelligent Algorithms

Most flash manufacturers don’t have the in-house E2E capabilities necessary to guarantee their SSDs’ full potential. With its own controllers, disk enclosures, and operating system, the OceanStor Dorado 3000 is able to run the cutting-edge FlashLink® intelligent algorithms.

  • The many-core balancing algorithm uses all of a controller’s processing cores to their full potential.
  • The method for service splitting removes the strain on the controller enclosure by moving reconstruction tasks to the smart SSD enclosure, allowing the controller to focus on I/O processing.
  • The caching approach speeds up application processes by utilizing an intelligent module to speed up batch processing.

Synchronization between SSDs and controllers ensures a consistent data layout.

  • The large-block sequential write algorithm combines several smaller data blocks into one large one before writing it to disk. This eliminates write amplification and keeps performance consistent.
  • Algorithm for independently dividing metadata, which efficiently mitigates the performance hit introduced by trash collection and ensures consistent operation.
  • Algorithm for adjusting I/O priorities in order to reduce access latency for read and write I/Os.

Huawei OceanStor Dorado is able to reach unprecedented performance because to FlashLink® intelligent algorithms that make full use of every flash memory.

E2E NVMe Architecture for Full Series

Businesses have embraced all-flash storage as a method to modernize their IT infrastructures, but the demands of always-on service models keep raising the bar for IT system performance. The bottleneck of 0.5 ms latency cannot be overcome by traditional SAS-based all-flash storage. However, NVMe all-flash storage is a future-proof architecture since it allows for direct communication between the CPU and SSDs, hence decreasing the distance data must travel. As an added bonus, the write request processing time has been cut in half while the number of concurrent operations has increased by a factor of 65,536. Huawei was an early adopter of E2E.

All models utilize NVMe technology. The Huawei-developed link-layer protocols and the leading 32 Gb FC-NVMe/25 Gb RoCE protocols at the front end allow the OceanStor Dorado 3000 to perform instantaneous failover and plug-and-play, enhancing both reliability and operations and maintenance. For end-to-end data acceleration, it employs a 100 Gb RDMA protocol. As a result, we may achieve latency as low as 0.05 ms and transmission speeds 10 times that of SAS all-flash storage.

Linear Increase of Performance and Capacity

Storage needs to offer simple linear increases in performance when more capacity is added to keep up with the ever-changing needs of businesses as they expand at unpredictable rates. The performance requirements of a growing business can be met by scaling up to sixteen controllers on an OceanStor Dorado 3000 system, with IOPS growing linearly with the number of controller enclosures.



On and Off-Cloud Synergy

The Huawei OceanStor Dorado 3000 is an all-flash system with an integrated intelligent hardware platform that blends general-purpose cloud intelligence with tailored edge intelligence to deliver incremental training and deep learning for a one-of-a-kind client experience. DME IQ is an intelligent O&M and management platform that monitors a live network in order to collect and analyze data from over 190,000 device patterns in order to derive general principles that improve fundamental O&M.

Intelligence Throughout Service Lifecycle

Intelligent management allows for 60-day and 14-day predictions of performance bottlenecks and disk errors, respectively, and gives rapid remedies for 93% of problems observed. This includes resource planning, provisioning, system tuning, risk prediction, and fault identification.

Extensive Intelligent Software Features

Intelligent heterogeneous virtualization and data migration together with thin provisioning to optimize space efficiency, intelligent quality of service to enhance service quality, and intelligent data reduction to simplify system lifecycle management.


For Inquiries