huawei OceanStor 5310/5510/5610
New-Gen Hybrid Flash Storage Systems

Flash-Like Performance
The global cold and hot data perception and data collaboration algorithms that support self-learning in all scenarios are used to
detect changes in service models and cold and hot data, helping promptly locate hot data in all scenarios.
Elastic convergence of cache and tiers resolves challenges faced in the traditional practice and fuels the hot data acceleration,
which results in optimal data layout and simplified configuration.
Redirect-on-write (ROW) large block sequential write: Multiple small block discrete writes are aggregated into one consecutive
large block write, reducing write amplification and ensuring stable write performance in all RAID arrays.
Low-latency RDMA connections among multiple controllers and fully-balanced active-active architecture enable a single LUN
to deliver more than 90% of the system performance.
End-to-end NVMe architecture: Storage supports 32Gb FC-NVMe/25Gb RoCE at the front end and 100Gb RDMA at the back
end, which realizes end-to-end data acceleration and enables latency as low as 0.05 ms.
The global distributed file system distributes subdirectories on multiple controllers, which relieves the performance pressure of
frequent access to a large number of small file directories. Two layouts of metadata sequence tables and hash tables improve
the system OPS by 30%.