Deduplication is essential to any modern data protection system and the VirtuoSO data deduplication engine is tightly integrated within SOFS. Incoming data is parsed to determine its type. Based upon the type identified, an algorithm (i.e., software intelligence) that has a content-aware knowledge base about each type of data type decides to either dedupe the data inline or via post-processes (at rest).
How objects are managed is also essential. Beyond deduplication and performance management, VirtuoSO can manage objects within different storage tiers, which is critical, because less frequently accessed objects can be placed on more cost-effective storage layers. It also can replicate designated high-value data objects for even higher availability and also move selected objects offsite (such as for disaster recovery or cost reasons).
Contrast this approach to the narrower focus of VTL systems. A VTL implements only a tape abstraction (i.e., disk appears to be tape) on top of a high-performance extent manager (which is not a file system). As a result, a VTL has no support for different storage tiers or for stubs that point to relocated objects. An example of why this is important is that a data protection platform has to protect not only mission-critical systems, but also important, though not time sensitive, data.
Putting all data on expensive, high-performance disk storage that mission-critical applications need in order to provide the necessary performance for a recovery action would be overkill for data that doesn’t demand that level of recovery performance. Bottom line: Tiering data properly saves money.
On the physical storage side, VirtuoSO uses HUS-110 from Hitachi Data Systems (HDS), although it can easily adapt to other storage products. And via its gateway configurations, VirtuoSO supports additional HDS storage products, such as the entire HUS line as well as the Virtual Storage Platform.
Mesabi Musings
We are now -- more than ever -- in a data-driven world. The onrush of new data from not only familiar applications and data types, but now from new applications and data types, can seem overwhelming. The resulting challenge to IT is two-fold: How do I manage the business on a daily, operational production basis and how do I manage all that quantity and diversity of data?
Sepaton recognized that a block-based approach has its limitations in being able to deal with all the requirements that the brave new world of data protection needs, which is why VirtuoSO features file-orientation. That enables the use of functions and features, such as hybrid deduplication, that are required to deal with volume, variety and velocity challenges.
The data protection world requires parallel development to the production world because it complements that world. With VirtuoSO, Sepaton has placed a stake in the ground. It will be interesting to see how other vendors respond.
Sepaton is a client of David Hill and the Mesabi Group.