OceanStor 5120 Entry-Level Hybrid Flash Storage with USB Interface for SME Networking

SmartAcceleration is a key feature for performance improvement on the next-generation OceanStor hybrid flash storage.It leverages the large block sequential write mechanism of redirect-on-write (ROW) and uses a unified performance layer for cache and tier performance acceleration.This breaks the bottleneck of traditional HDDs in random IOPS performance, maximizing the performance of the hybrid flash system.

OceanStor storage systems use distributed file systems, which means the file systems are not owned by any specific controller. Directories and files in a file system are evenly distributed to all controllers by a balancing algorithm. Read and write requests are equally distributed on each controller so that one file system can fully utilize the resources of the entire storage cluster. Customers can use the file system in one namespace or multiple file systems based on their service plans. The distributed file systems apply to file sharing scenarios with coexisting mass volumes of small and large files. Data in each directory is evenly distributed to each controller for load balancing. The same controller processes the I/Os of a directory and its files to eliminate forwarding across controllers and improve performance for directory traversal, attribute traversal, and batch attribute configuration. When large files are written to a storage pool, RAID 2.0+ globally distributes their data blocks to all SSDs in the storage pool for improved write bandwidth.

Flexible Scalability

OceanStor storage systems support both scale-up and scale-out, achieving flexible scalability while maintaining high performance.

• Scale-up

  Increases storage capacity and improves processing capabilities of existing controllers.

• Scale-out

  Enables performance to increase linearly with the number of controllers.

High Reliability

OceanStor storage systems offer advanced data protection technologies to minimize risks of disk failures and data loss, and protect data against catastrophic disasters, ensuring continuous system running.

• Component failure protection

  Storage system components are in 1+1 redundancy and work in active-active mode.Normally, the two redundant components are working simultaneously and share loads.If one component fails or is offline, the other one takes over all loads without affecting ongoing services.

• RAID 2.0+

  The RAID 2.0+ underlying virtualization technology is used to automatically balance loads across disks.RAID 5, RAID 6, and RAID-TP are supported.RAID-TP can tolerate simultaneous failure of three disks.If a disk encounters a fault, all the other disks in the same disk domain help reconstruct the faulty disk's service data, achieving a 20-fold faster reconstruction speed than traditional RAID and significantly reducing the possibility of multi-disk failure.RAID 2.0+ supports dynamic RAID and flexible data layout, accelerating SSD reconstruction.

• Power failure protection

  Built-in backup battery units (BBUs) supply power to controller enclosures in the event of unexpected power failures.This enables the storage system to write cache data to built-in disks of controllers to prevent data loss.

• Global wear leveling and anti-wear leveling
  • Global wear leveling: If data is unevenly distributed to SSDs, certain SSDs may be used more frequently and wear faster than others.As a result, such SSDs will fail much earlier than expected, increasing the maintenance costs.OceanStor storage systems address this problem by using global wear leveling that levels the wear degree among all SSDs, improving SSD reliability.
  • Global anti-wear leveling: When the wear degree of multiple SSDs is reaching the threshold, the system preferentially writes data to specific SSDs.In this way, these SSDs wear faster than the others.This prevents multiple SSDs from failing at a time.
• Disk data pre-copy

  The disk data pre-copy technology enables the storage system to routinely check hardware status and migrate data from any failing disk to minimize the risks of data loss.

• Advanced data protection
  • HyperSnap supports writable snapshots.Snapshot creation and activation have no impact on performance.
  • HyperReplication backs up local data to a remote storage system for disaster recovery.
  • HyperClone generates a full copy of the source data in the local storage system for data backup or use by other applications, ensuring local data security.
  • HyperCDP achieves continuous data protection at an interval of several seconds, generating more intensive recovery points on storage devices.
  • HyperMetro enables real-time data synchronization and access between two storage systems. If data access fails in either storage system, HyperMetro implements seamless service switchover to ensure data security and service continuity.