Sepaton launches super-duper big data-slurping deduper
Big iron for big data backup ... 16PB, if all goes according to plan
Sepaton has updated its 16-node deduping disk-to-disk backup mill with hybrid dedupe methods and has planned a capacity jump to 16 petabytes.
The S2100-ES2 generation is being followed by this new VirtuoSO product, which keeps the clustered node limit of 16 dedupe heads front-ending a disk drive array but bumps up the processing power of those heads, as well as extending the scope of the deduplication.
The net result is an increase in throughput and in the amount of backup and archive data that can be stored. The new kit will be solidly pitched at enterprise customers with their ever-increasing big data backup needs.
There is an S2100-ES3 product but that is limited to 8 nodes.
Sepaton says VirtuoSO has the world’s fastest D2D performance and largest capacity in a single system. It is "the industry’s highest-performing scale-out NAS-based data protection solution with five times the capacity of competitive solutions at initial release [and] up to 22 times the capacity planned for future releases."
The bare bones details are:
- Raw capacity scales from 36TB to a tested 3.3PB and a planned 16PB - 4,000 4TB disk drives.
- Throughput scales from 7.9TB/hour for a single node to a tested 31.6TB/hour for 4 nodes and an intended 126TB/hour for 16 nodes. This is faster than the 100TB/hour of HP's StoreOnce B6200 backup array.
- Hybrid in-line and post-process deduplication that's global across the nodes.
- Back-end HDS HUS 110 storage arrays and any virtual storage platform (VSP) managed arrays with RAID-6, 99.999 per cent data availability and no single point of failure. SSDs are used to accelerate system performance.
- NAS and CIFS access.
Each VirtuoSO processing node is a high capacity, enterprise class server. VirtuoSO automatically load-balances all backup, deduplication, replication, and restore operations globally across nodes. There is an Exar 1845 compression card per node.
The processing nodes run a 64-bit Linux kernel that supports multi-core CPUs "to deliver many times faster performance than the nearest competitive appliance solution". Eat your heart out Data Domain, Exagrid, NEC HydraStor and pals.
The Smart Hybrid Deduplication has these features:
- Hash-based inline deduplication applied as data is ingested.
- Post-process, content-aware deduplication applied after data has landed.
- Analysis of data types and change rates of incoming data streams and use of the most efficient combination of deduplication methods for maximum performance and capacity optimisation.
- It can perform byte-level deduplication of databases that store data in small segments (less than 8KB), which are too small for traditional deduping systems unless they take a performance hit. In this case post process/ContentAware deduplication at the byte level is faster and more efficient.
- Designed to use flash/ solid state disk (SSD) in its global index for performance.
- Non-dedupable compressed data and image data can be backed up directly.
- Policy-driven dedupe mode setting, including option to bypass dedupe for encrypted or compressed data types.
Sepaton claims: "VirtuoSO delivers the industry’s lowest total cost of ownership. Guaranteed."
It says its "Smart data movers" enable multi-source and multi-target replication, and also data migration for faster ingest and network efficiency through bandwidth-optimised replication.
VirtuoSO's management and reporting capabilities use a GUI with management dashboards visible from anywhere on any device
The packed VirtuoSO roadmap includes these features:
- Support for protocols such as OST, NDMP, SMB3, pNFS, a REST API and cloud storage APIs
- Slow disk technology support
- Federated Big Data backup applications
- Archive on spin-down storage
- Forensics applications
- Storage analytics integration with data protection applications
- Tiering support with data migration and an archive tier on the system or in the cloud
- Encryption of data-at-rest and secure erasure
- Cloud integration and multi-tenancy
- Client-side data reduction through deduplication and source deduplication
- Data movement without traditional backup applications
- Edge systems
No disk drive manufacturer has yet used the term "slow disk technology" in public and we're thinking of sub-5,000rpm drives sipping miniscule amounts of power. Likewise, the only archive facility on spun-down disks comes from SGI and we're expecting another one from Seagate subsidiary EVault in the future. It looks to us as if Sepaton and Seagate have been talking about spin-down-capable shingled media drives.
With VirtuoSO Sepaton is sticking a determined stake in the high-end deduping D2D backup and archive ground and saying to its competitors; "Match that if you can. We intend to be around for the long haul."
VirtuoSO is currently in limited release and will be generally available in the first quarter of 2014. Pricing starts at $344,500. ®