Organizations are becoming increasingly virtualized within their data center infrastructures which is leading them to aggressively virtualize the storage arrays in their infrastructure to complement their already virtualized server environment. As they do so, it behooves them to distinguish between, and have a clear understanding, of each virtual component that makes up their newly virtualized storage infrastructure. The need to clarify this terminology comes clearly into focus as organizations evaluate the multi-tenancy and virtual storage array capabilities found on many high end storage arrays.
Multi-tenancy and virtual storage array capabilities are becoming more prevalent on high end storage arrays used within virtualized infrastructures as these two features enable organizations to:
- Securely share an array’s physical resources between internal departments, external organizations or some combination of both
- Capitalize on the benefits that virtualization offers by segmenting a single physical array into multiple logical, virtual arrays so each individual virtual storage array may have its own properties and be managed independently.
While these two features sound very different, articulating the specific differences between them becomes very difficult very quickly. These two features share a number of attributes in common which can lead one to confusing one term with the other or even concluding that these two terms are synonymous.
To distinguish between these two features, it is best to first articulate a few attributes that the multi-tenancy and virtual storage array features share in common. These include:
- Specific physical resources assigned to each tenant or virtual storage array
- Specific management and security restrictions
- Multiple parties each having access to, and control over, their own storage resources
However there are some important distinctions between them that organizations should keep in mind as they evaluate high end storage arrays and their support for these two specific features.
First, a high end storage array that supports multi-tenancy may not necessarily support the virtual storage array feature. It is possible for a high end storage array to support multi-tenancy as it may allow an administrator to create multiple groups on the storage array that each have their own security permissions and access to specific physical resources. This capability is sometimes referred to as partitioning. However, just because a physical storage array supports this partitioning or multi-tenant capability does not mean that each of these partitions is a virtual storage array.
The operating systems on the hosts assigned to a specific partition or tenant on the physical storage array will see the storage array’s unique identifier on each of the storage volumes presented by the physical storage array. This unique storage array identifier will be same regardless of which partition or tenant presents the storage volume.
If the physical storage array supports a virtual storage array feature, the unique storage array identifier that exists on each virtual storage volume presented by each virtual storage array will be different. So even though all virtual storage arrays reside on the physical storage array, the host will not be able to come to that conclusion by simply examining the storage array ID of the storage volume presented to it.
Second, a physical storage array that supports the virtual storage array feature will be capable of doing multi-tenancy as long as it can support two or more virtual storage array instances. However, a physical storage array that supports multi-tenancy does not necessarily support the virtual storage array feature. This second point is somewhat of a re-iteration of the first point. However I make it to help further highlight the differences and similarities between these two features.
For example, if one reads that a certain physical storage array supports a virtual storage array feature and can create two or more virtual storage array instances, the physical storage is by definition multi-tenant capable. However, if a physical storage array supports multi-tenancy but each partition or storage volume within the partition cannot present its volumes using a storage array identifier other than the underlying physical storage array’s pre-existing one, the physical storage array does not support a virtual storage array feature. It is only multi-tenant.
Third, the virtual storage array and multi-tenant capabilities are not mutually exclusive as the virtual storage array feature may have multi-tenant capabilities. While this is again a bit of a nuanced point, organizations should recognize that if a physical storage array has a virtual storage array feature, within each virtual storage array they may be able to implement multi-tenancy. Using this multi-tenancy feature, they can partition resources within a virtual storage array to specific hosts and grant administrator privileges within this partition to specific individuals.
Multi-tenancy and virtual storage arrays are features that organizations are increasingly likely to encounter as they virtualize their storage infrastructure. As they encounter these two features, they should be careful to distinguish between them to ensure they are getting the right set of features for the specific problem that they are trying to solve. While there is admittedly some overlap between these two features in terms of the attributes they possess, there are also enough differences between them that having one feature without the other may preclude an organization from managing their virtualized storage infrastructure in the most efficient and effective manner possible.