ARGO STORAGE
|
Peculiarities of the solution - Erasure Coding with 2/3 parity blocks. Optional redundancy for critical data can be increased to 3 copies of data blocks - Fault tolerance: from 2 disks in each disk node, up to 2/3 disks in the storage (depending on configurations)
|
- Distributed architecture: no bottlenecks typical of legacy systems - High degree of parallelism and automatic load balancing between disk nodes - Managed performance up to 8 million IOPS depending on configurations - Processing of blocks, files, and object data by a single unified system. - Controllers and storage nodes are standard x86 servers with verified configurations, no vendorlock
Basic information ARGO's distributed storage technology uses Ethernet interconnect technology with extremely low overhead. STORAGE ARGO includes its own version of the operating system, application service sets and a file system that provides reliable and long-term data storage. Thin provisioning, tiered caching enables efficient use of storage capacity and flexible support of different workloads. |
|
Functionality |
|
| Number of controllers | from 2 |
| Number of disc units | of 8 |
| Access Protocols | SMBv1/2/3/3/3.1, NFSv3/4, S3 |
| Access to the same dataset using different protocols | ✓ |
| Storage pool size | Up to 3 × 1023 Petabytes |
| File system size | Up to 256 Zettabytes |
| Number of file systems in one storage pool | 264 |
| Number of snapshots/clones | 2 × 1014 |
| No performance degradation when filling the pool | ✓ |
| No performance degradation when working with a large number of snapshots and clones | ✓ |
| Application-aware recording to HDD | ✓ |
| Asynchronous replication | 1:1, 1:N |
| Excessive coding when storing data | N+2P / N+3P |
| Ability to replicate data blocks | RF = 1|2|3 |
| Background Data Integrity Check | ✓ |
| Transactional model of working with data | ✓ |
| Duplication of metadata and control blocks with umm | ✓ |
| Hardware components are redundant | ✓ |
| Hot swap drives reserve on storage nodes | ✓ |
| Supported HDD/SSD drive types in a single node | 512E/N/4KN |
| NVMe support | ✓ |
| Data deduplication | inline |
| Data compression | inline - lz4, gzip[0-9], lzjb, zle |
| Variable block size | 512 Bytes - 1 Mbyte |
| Caching functionality | read and write cache |
| Caching levels | RAM ( 256GB ) / SAS SSD |
| Multithreaded data prefetching | ✓ |
| Functionality of self-recovery of corrupted data | ✓ |
| Isolated control and monitoring environments | ✓ |
|
System environment and interconnect |
|
| Network interfaces on controllers | 10/25/100 Gbit Ethernet |
| Network interfaces on storage nodes | 10/25 Gbit Ethernet |
| Support for Jumbo Frames | 9000 Bytes |
|
Security, integration, interface |
|
| Operating system support | Windows Server, Linux, HPUX, AIX, Solaris |
| Integration with VMware | VAAI |
| NDMP support | ✓ |
| Support for quota and resource reservation | at the user and group level |
| Control interface | Web-UI |