Contents |
Storage virtualization introduction
In general, virtualization of volumes breaks the physical connection between the LUN (server side) and the physical volume (array side); the virtualization engine keeps track of (maps) the connection between the LUN and the virtual volume (as the server sees it), and the connection between the virtual volume and the physical volume (as the array sees it).
The quality of the virtualization solution can vary significantly, both in the quality of storage management services that are offered, the degree of fine storage management that it allows, and the other technologies that are supported by virtualization (such as thin provisioning).
Storage virtualization is now a rapidly maturing technology, and is a very usefull building block, on which additional storage services can be provided to applications in an on demand way. Technologies such as tiered storage can be significantly enhanced by being based on a virtualization environment. New extensions to virtualization such as thin provisioning depend on virtualization being available.
Block-based storage virtualization
There are five main types of block-based virtualization engines:
- Heterogeneous array controller based included physical volumes in the array and on external volumes on external arrays (Hitachi is the only vendor)
- Homogeneous array controller based included physical volumes just in the array with the controller (e.g., 3PAR, Equalogic and Hitachi)
- Heterogeneous external virtualization appliances that maintain the maps for multiple heterogeneous arrays in the appliance (e.g., IBM SVC and Hitachi USP VM "diskless")
- Heterogeneous external virtualization appliances that maintain the maps for multiple heterogeneous arrays in the appliance and in the LAN switch (e.g., EMC Invista or Incipient together with CISCO LAN switches)
- Server-based virtualization (e.g., Veritas File System together with Veritas Volume Manager) offers a solution for operating systems and applications that support an alternative file system.
The 3PAR and Equalogic are homogeneous virtualization solutions only. Server-based storage virtualization is theoretically heterogeneous, but has many constraints on the operating systems and applications that can be supported. The rest are fully heterogeneous.
All the block-based storage virtualization architectures work theoretically and practically. The following comments are designed to help position the suitability of different architectures for different storage environments:-
- For high-performance tier-one applications where availability is important, the placing of the virtualization layer in the storage controller potentially may improve availability and performance by reducing communications paths, reducing the maintenance complexity and reducing the number of components that have to talk to each other in the case of failures.
- Placing the virtualization layer in the controller also increases the potential for greater layers of abstraction within the virtualization architecture, allowing greater control of resources. 3PAR are an example of achieving a high degree of automation from virtualization (with fewer knobs for storage administrators).
- Appliance based architectures have lower costs of initial implementation and adoption for a heterogeneous environments, especially for tier-two environments.
- For appliances, both architectures work. The use of LAN switches can improve I/O times a little, but there is additional complexity from additional components in the solution which could lead to lower availability and increased maintenance complexity.
- Server-based virtualization can create significant server overhead, and may require the use of a non-standard file manager which may introduce additional software maintenance costs.
File-based storage virtualization
File-based virtualization is very different from block-based virtualization. This part of this article is a stub, please help improve it.
Benefits of storage virtualization
The management of storage devices takes significant storage administrator time and is error-prone. By hiding the complexity of the heterogeneous arrays, storage virtualization helps the storage administrator to perform the tasks of copy services, backup, archiving, and recovery with less effort, reduced elapsed time, and with fewer errors. In addition, the commissioning and decommissioning of storage is very time consuming and takes a long elapsed time (typically from 3 to 9 months upwards). By being able to migrate virtualized storage dynamically (e.e., without any impact on any application) from the old array to a new array, the complexity, impact on users, and elapsed time can be reduced significantly.
Software costs and storage virtualization
One potential benefit of storage virtualization is the simplification of the storage management, and the reduction in storage array management software costs. However, a common practice for providers of virtualization software is to base software charges on the terabytes managed. As the cost of terabytes continues to drop by over 30%/year, user can purchase over 50% more storage each year with the same budget. Great care has to be taken in negotiating contracts to ensure that the virtualization software costs do not escalate out of control. Basing virtualization software on the number of physical storage arrays is a more practical method of metering storage virtualization software costs.
Related research: Managing geometric data growth in SANs
