Scale out NAS Backup for Scale out Storage

Backup seems like a simple concept, one that for most companies involves a fairly simple process, copying data to a dedicated storage area. But most companies still struggle with it and for the largest enterprises and government agencies the process is becoming increasingly complicated. Moving the sheer volumes of data these organizations must, in the timeframes they have available, requires a special solution. For most of the past decade, that solution has come from Sepaton.

Sepaton’s ‘DeltaScale’ disk-based backup systems deliver some of the fastest backup and recovery performance on the market (up to 80TB per hour) in a modular, scalable, single-system architecture – there’s no sprawl of backup appliances. Using byte-level deduplication (dedupe), Sepaton’s systems provide some of the highest, most consistent data reduction ratios regardless of data type, enabling multiple-PB, single-system capacities.

Now scale out primary storage systems are driving the need for an even larger, more scalable backup solution, one that’s designed to support all backup applications. And the demand for IP-based network protocols in virtualized data center environments is driving the need for this capability in a NAS-based system.

VirtuoSO is a new NAS-based architecture for Sepaton’s high-performance, deduplicating backup systems. SO stands for “Scale Out”, a scale out NAS in this case that’s designed to keep pace with today’s scale out primary storage systems, in capacity, performance and compatibility. As a NAS solution this system can leverage standard file system protocols to support any backup software application that’s designed to back up data to a NAS storage target.

The VirtuoSO data protection solution leverages a proven scale out hardware architecture that can support from 36TB to 16PB of backup data in a single storage pool, again, no backup appliance sprawl. The system uses 3TB disk drives to reduce data center footprint and supports up to 96 x 10GbE ports for extremely high throughput.

OptiScale Architecture

The VirtuoSO system uses an object storage architecture with a file system presentation layer that Sepaton refers to collectively as its “OptiScale” architecture. Object storage is becoming the default architecture for the largest scale out storage systems in the Big Data and hyperscale web environments because it’s more efficient and enables performance to scale as capacity expands. It does this by writing data to discrete objects (similar to files) which are identified by a unique ID number and accessed through a simple index lookup.

As it grows, a traditional block-based storage system can get extremely complex since the storage controllers have to manage the placement of associated data (files) on arrays of unrelated data blocks. Managing the storage and retrieval of these blocks, as well as the LUNs and RAID groups involved, generates a large amount of metadata. This structure and volume of metadata can slow down file access and create metadata processing bottlenecks, especially in environments that have very large file counts.

An object-based architecture is less complex and involves much less metadata. This makes it ideal for very large disk backup systems that must be able to grow with the scale out primary storage systems they’re backing up and support large numbers of files.

Scale out NAS

Putting a scale out NAS on top of the object-based storage system further leverages these advantages. With a POSIX-compliant, distributed file system, VirtuoSO supports backup applications that can write to NFS or CIFS storage targets. This essentially makes the product backup-application-agnostic freeing Sepaton from having to maintain development for current and future software releases and develop specific drivers to support each backup application that may emerge down the road.

Sepaton has made their name with one of the highest performing deduplication solutions on the market, using ‘ContentAware’, byte-level deduplication. This modified post-process dedupe technology produced performance similar to in-line deduplication without the scalability issues that most of those dedupe solutions had.

But Sepaton believes that there’s really no one best deduplication technology, it varies with the data being processed. For example, in-line dedupe is best for smaller files and small change-rate data sets. But for multi-streamed workloads, such as databases or transaction-based data sets that often have very small record sizes, it isn’t the best method. In these scenarios, where data objects are typically less than 8KB, in-line technologies won’t find much duplication. For these data sets, post-process dedupe is the most effective technology.

Smart Hybrid Deduplication

With VirtuoSO, Sepaton has leveraged both in-line and post-process deduplication to produce the best data reduction ratios possible. Called ‘smart hybrid deduplication’, the system can apply whichever dedupe technology is most effective, based on the data type or on preset policies. And, any portion of a file that does not deduplicate effectively in-line will be passed to the post-process method. Similarly, it can run post-process deduplication on a data set that has already been stored with in-line dedupe to maximize data reduction.

It’s common for other backup products to recommend skipping deduplication altogether for some data types, such as multiplexed or source-deduped data. Sepaton’s hybrid deduplication technology can give users a solution that produces better dedupe ratios, even in these difficult situations.

Future proof

Object storage is an ‘extensible’ architecture, meaning it has the flexibility to expand easily and efficiently, so it doesn’t become more complex as it grows. This enables the system to provide consistent, predictable performance throughout the expansion process. The result is a storage architecture that can scale along with the primary storage systems it’s backing up, eliminating the cost and disruption of a ‘forklift’ upgrade.

The use of a scale out NAS layer enables VirtuoSO to support any backup software application that can write data to a standard file system. This means that the Sepaton system is ‘application-agnostic’ and will remain compatible with all major backup software revisions down the road.

Sepaton’s OptiScale architecture is designed with the flexibility to integrate with other protocols on the front end, such as OST and REST. Its Smart Data Mover software will enable data replication between heterogeneous storage platforms, even to the cloud. And with its advanced deduplication technology integrated within the file system it can replicate these data sets in less time.

Summary

With scale out storage primary data sets are expanding at unprecedented rates and backup systems are hard pressed to keep up, both in terms of capacity and consistent performance. Deduplication can help, but data reduction ratios can be hard to predict since different data types respond differently to deduplication. For companies backing up large data sets these factors can make backup windows hard to plan for, an unacceptable situation when critical data must be protected.

Sepaton’s VirtuoSO system with its object storage architecture can maintain storage performance by removing the potential bottlenecks of traditional storage as data sets grow. As a scale out NAS system with integrated, smart hybrid deduplication it can provide consistent, predictable backup performance at almost any scale, for virtually any backup application.

Sepaton is a client of Storage Switzerland

Eric is an Analyst with Storage Switzerland and has over 25 years experience in high-technology industries. He’s held technical, management and marketing positions in the computer storage, instrumentation, digital imaging and test equipment fields. He has spent the past 15 years in the data storage field, with storage hardware manufacturers and as a national storage integrator, designing and implementing open systems storage solutions for companies in the Western United States.  Eric earned degrees in electrical/computer engineering from the University of Colorado and marketing from California State University, Humboldt.  He and his wife live in Colorado and have twins in college.

Tagged with: , , , , , , , ,
Posted in Product Analysis

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 22,245 other followers

Blog Stats
  • 1,563,235 views
%d bloggers like this: