Product Features

Fast Innovative hardware platform:

The hardware platform of Huawei storage enables E2E data acceleration, improving the system performance by 50% compared to the previous generation.

● The intelligent multi-protocol interface module hosts the protocol parsing previously performed by the general-purpose CPU, expediting the front-end access performance by 20%.
● The computing platform offers industry-leading performance with 25% higher computing power than the industry average.
● The intelligent accelerator module analyzes and understands I/O rules of multiple application models based on machine learning frameworks to implement intelligent prefetching of memory space.
● This improves the read cache hit ratio by 50%.                                                                                                                                                                                                              ●  The intelligent SSD hosts the core Flash Translation Layer (FTL) algorithm,accelerating data access in SSDs and reducing the write latency by half.
● The intelligent hardware has a built-in Huawei storage fault library that accelerates component fault location and diagnosis, and shortens the fault recovery time from 2 hours to just 10 minutes.

Intelligent algorithms:

Most flash vendors lack E2E innate capabilities to ensure full performance from their SSDs. OceanStor Dorado 3000 runs industry-leading FlashLink® intelligent algorithms based on self-developed controllers, disk enclosures, and
operating systems.

● Many-core balancing algorithm: Taps into the many-core computing power of a controller to maximize the data processing capability.
● Service splitting algorithm: Offloads reconstruction services from the controller enclosure to the smart SSD enclosure to ease the load pressure of the controller enclosure for more efficient I/O processing.
● Cache acceleration algorithm: Accelerates batch processing with the intelligent module to bring intelligence to storage systems during application operations.

The data layout between SSDs and controllers is coordinated synchronously.

● Large-block sequential write algorithm: Aggregates multiple discrete data blocks into a unified big data block for disk flushing, reducing write amplification and ensuring stable performance.
● Independent metadata partitioning algorithm: Effectively controls the performance compromise caused by garbage collection for stable performance.
● I/O priority adjustment algorithm: Ensures that read and write I/Os are always prioritized, shortening the access latency.

FlashLink® intelligent algorithms give full play to all flash memory and help Huawei OceanStor Dorado achieve unparalleled performance for a smoother service experience.

E2E NVMe architecture for full series:

All-flash storage has been widely adopted by enterprises to upgrade existing IT systems, but always-on service models continue to push IT system performance boundaries to a new level. Conventional SAS-based all-flash storage cannot break the bottleneck of 0.5 ms latency. NVMe all-flash storage, on the other hand, is a future-proof architecture that implements direct communication between the CPU and SSDs, shortening the transmission path.
In addition, the quantity of concurrencies is increased by 65,536 times, and the protocol interaction is reduced from four times to two, which doubles the write request processing. Huawei is a pioneer in adopting end-to-end NVMe architecture across the entire series. OceanStor Dorado 3000 uses the industry-leading 32 Gb FC-NVMe/25 Gb RoCE protocols at the front end and adopts Huawei-developed link-layer protocols to implement failover within seconds and plug-and-play, thus improving the reliability and O&M. It also uses a 100 Gb RDMA protocol at the back end for E2E data acceleration. This
enables latency as low as 0.05 ms and 10x faster transmission than SAS all-flash storage.

Linear increase of performance and capacity:

Unpredictable business growth requires storage to provide simple linear increases in performance as more capacity is added to keep up with everchanging business needs. OceanStor Dorado 3000 supports scale-out of 16
controllers, and IOPS increases linearly as the quantity of controller enclosures increases, matching the performance needs of future business development.

Intelligent

On and off-cloud synergy:

Huawei OceanStor Dorado 3000 all-flash system combines general-purpose cloud intelligence with customized edge intelligence over a built-in intelligent
hardware platform, providing incremental training and deep learning for a personalized customer experience. The DME-IQ intelligent O&M and management platform collects and analyzes over 190,000 device patterns on
the live network in real time, extracts general rules, and enhances basic O&M.

Intelligence throughout service lifecycle:

Intelligent management covers resource planning, provisioning, system tuning,risk prediction, and fault location, and enables 60-day and 14-day predictions of
performance bottlenecks and disk faults respectively, and immediate solutions for 93% of problems detected.

Extensive intelligent software features:

Thin provisioning and data reduction improve space utilization; intelligent QoS improves service quality; and intelligent heterogeneous virtualization and data migration combine to ensure simplified system lifecycle management.

Simplified

Simple management:

Huawei OceanStor Dorado 3000 delivers SAN and NAS services and supports their parallel access. Built-in containers support storage and compute convergence. The convergence of cross-generation devices allows for joint
resource usage. As such, multiple functions are converged to simplify management and greatly reduce the TCO.

Simple configuration:

A brand-new graphical user interface (GUI) greatly simplifies the configuration process of traditional storage. This facilitates storage system configuration in just three steps and resource readiness in just five minutes, without assistance from dedicated personnel. This meets the key requirements of SMBs for simple and easy-to-use IT systems.

Simple O&M:

The active-active architecture ensures there is no LUN ownership, meaning a LUN does not belong to any specific controller. In addition, load balancing and non-disruptive upgrade (NDU) are supported. O&M personnel do not need to prepare much on the host side before an upgrade, greatly improving O&M efficiency.

Technical Specifications