Aiserver Manufacturer
1 / 5
OceanStor Dorado 18500/18800/8000 is a high-end storage product with high storage performance and reliability for core business of enterprises, providing enterprises with excellent data service experience. It adopts a new generation of hardware platform and extremely stable SmartMatrix fully interconnected architecture, which can tolerate 7 out of 8 controllers to fail without interrupting business. It ensures that core business is continuously online.
With built-in innovative hardware, FLASHLINK® intelligent algorithm and full range of end-to-end NVMe architecture, it can reach 40 million IOPS and the latency can be as low as 0.05ms; the industry's first intelligent acceleration module makes storage "smart" and gets better with use. OceanStor Dorado 18500/18800/8000 is widely used in database, virtualization and big data analysis scenarios in operators, finance, government, manufacturing and other industries.
| Model | OceanStor Dorado 8000 V6 / Dorado 18500 V6 Description |
|---|---|
| 02354JLD | Dorado 8000 V6 Controller Enclosure(4U,Four Ctrl,NVMe,AC\240V HVDC,4TB Cache,SPE74C0400)/Dorado 18500 V6 controller enclosure (4U, four controllers, NVMe, AC\240V high-voltage DC, 4TB cache, SPE74C0400) |
| DV6-SMARTIO4*32FC-H | 4 ports SmartIO I/O module(SFP28,32Gb FC)*4 |
| H-RDMA-BE | 2 ports 100Gb RDMA I/O module(QSFP28,for Back-End)*2 |
| 02355WYM | D8V6-SSD-NVMe-7.68T*184 |
| 02355XUP | D8-NVMe-DAE36-2U-ACp*6 |
| 14130858 | SN2F01FCPC*16 |
| 04052182 | HSC-100G QSFP28-2M*12 |
| 88036XNH | D8V6-LBS-Super*1 |
| 88036SMP | D8V6-DDCM-C-SUP-RES2*1 |
OceanStor Dorado 6800/18500/18800 high-end all-flash storage system is a new benchmark for storage performance and reliability created for core business of enterprises, providing enterprises with excellent data service experience. With innovative hardware platform, FLASHLINK intelligent algorithm and full-series end-to-end NVMe architecture, it can reach 40 million IOPS and 0.05ms extreme latency; it adopts the extremely stable SmartMatrix fully interconnected architecture to ensure that core business is online 7*24 continuously; 3-layer intelligent management, predict performance in advance and warn of faults, and operation and maintenance are always one step faster.
Component-level reliability: Wear and anti-wear balance logic ensures disk durability.
Architecture & product-level reliability: Tolerance of controller single/dual/multi-point (8 bad 7, 4 bad 3, cross-engine) failure, disk frame-level failure, three-disk failure.
Solution & cloud-level reliability: The industry's only SAN & NAS integrated AA active-active, SOCC, two locations and four centers, no network Cloud backup, zero business interruption, zero data loss.
3-layer intelligent management:
Advanced Integration: SAN & NAS & object fusion/generation heterogeneous fusion. Efficient resource utilization. Intelligent ransomware detection, detection efficiency up to 99.99%.
The storage innovative hardware platform realizes end-to-end data acceleration, which can reach 40,000,000 IOPS. The intelligent multi-protocol interface card is used to carry the protocol analysis work that the general CPU of this application is responsible for, improving end access performance by 20%. The computing power of the computing platform with industry benchmark performance is better than the industry standard by 25%.
Based on machine learning framework, combined with intelligent hardware and intelligent caching algorithm, actively analyze and master multiple applications. Using the I/O law of the model, the intelligent prefetching of memory space is realized, and the hit rate of reading cache is increased by 50%.
Smart cache+SCM intelligent multi-level cache, intelligent identification of data hot and cold grading acceleration, OLTP full read delay reduce by 60%. Intelligent SSD carries the core FTL (Flash Translation Layer) algorithm to accelerate the data access in SSD disk, shortening writing delay by 50%. Built-in fault gene bank in the intelligent hardware accelerates the fault diagnosis and location of components, reducing the time from 2 hours to 10 minutes.
Intelligent Algorithm: Based on the self-developed design, the industry-leading FLASHLINK® intelligent algorithm is developed: Many-core equalization algorithm fully releases processing power in the controller; Service diversion algorithm offloads data reconstruction from the controller to the intelligent hard disk frame; Cache acceleration algorithm improves the performance of batch processing scenarios; Block sequential writing algorithm aggregates multiple discrete blocks into a unified large block to reduce write amplification; Metadata independent partition algorithm controls performance loss caused by garbage collection; I/O priority scheduling algorithm ensures fast access times.
Full range of end-to-end NVMe architectures: Shortens the transmission path via direct CPU-to-SSD communication. Concurrency is increased by 65,536 times, protocol interaction is reduced from 4 times to 2, and write request processing efficiency is doubled. Front-end adopts 64Gb FC-NVMe/100Gb RoCE protocol, back-end utilizes 100Gb RDMA protocol, allowing latency to reach 0.05ms.
Globally shared distributed file system OceanFS: NAS performance is maximized through multi-core parallel processing, metadata intelligent pre-fetching, and intelligent layout, putting NAS performance 30% ahead of the industry.
Linear expansion of performance and capacity: Scale-out architecture supports linear expansion to a maximum of 32 controllers, and IOPS increases linearly with the engine to a maximum of 40,000,000 IOPS.
