Thin provisioning allows you to partition a volume to the size that the application or user requests, but only consume the actual disk space that is in use at that moment in time. For example, if you have the need to deploy a 750 GB Oracle environment, we all know that on Day One it will not need anything near that amount of storage space. In fact, often the applications never grow to their projected size. Having to deploy this upfront can be costly not only because you have to actually buy the storage, but also because you have to power and cool that storage.
The few remaining thin provisioning naysayers will warn you about over-allocating and running out of actual disk space. I have never really seen that happen as today's systems provide you with so many early warnings that it makes it almost impossible to run out of disk space and not be aware that it is happening. Thin provisioning reduces the total drive count that you have to purchase to service your application, reducing capital expenditures as well as power and cooling costs.
Performance also is a cost-related issue that storage virtualization can help curtail by using wide striping. With traditional storage systems, you define your RAID groups to a finite number of drives. To get performance to a given application, you need to create a high-drive-count RAID group. This causes further capacity allocation issues, as well as increasing the likelihood of a double drive failure and potential data loss. Virtualization storage systems use wide striping to scatter data across all the available drives in a system. Typically, the virtualization engine will group the drives by type and then stripe across all of those resources. This delivers maximum performance to all applications, lowering the amount of drive spindles that need to be dedicated to one task in particular.
Intelligent data migration is a feature unique to a few storage virtualization products, delivering near perfect, tiered storage. Data movement in a tiered storage strategy always is a challenge. Intelligent data migration solves most of the issues and expands the use beyond just file data. Even databases can be tiered. This capability allows you to move data at a block level as that data ages. This means you can start your database on high performance Fibre Channel and then as parts of it age, move those parts to lower speed and lower cost disk technology. You can easily start with active data residing on 15K fibre, as blocks of data age, migrate those to 10K Fibre and then to SATA, for example. With the coming surge of SSD-based products, this type of movement could be the killer app for SSD storage.
Since this movement is transparent and seamless, with minimal performance degradation as it moves between tiers, a very aggressive migration strategy is possible, lowering the expenditure in expensive high-speed fibre drives. We will do a deeper dive on Intelligent Data Movement in a future entry.
A virtualized storage system is a very wise investment given the market right now. It is one of the few purchases that you can make in the data center that will increase productivity and user satisfaction, while at the same time significantly reduce budget expenditures as well as power and cooling costs.
George Crump is founder of Storage Switzerland, an analyst firm focused on the virtualization and storage marketplaces. It provides strategic consulting and analysis to storage users, suppliers, and integrators. An industry veteran of more than 25 years, Crump has held engineering and sales positions at various IT industry manufacturers and integrators. Prior to Storage Switzerland, he was CTO at one of the nation's largest integrators.