14 DAY TRIAL // OCZ has formally introduced the Z-Drive R4 series of PCI Express enterprise solid state drives months ago, but that same series of storage devices is making the news again, thanks to a newly gained functionality.
The SSDs have been paired with the VXL software and have apparently acquired the ability to significantly speed up virtualization of Microsoft's SQL Server 2012 database.
SQL Server 2012 is very useful for setting up cloud data centers, so OCZ optimized the VXL software on the PCI Express SSD models to enhance it.
Essentially, VXL lets Z-Drive R4 PCIe appear as a highly available network resource through the “SAN-less DATA Center” mode. The flash memory is exposed to any VM in the virtualized cluster without any of the hypervisor layer's capabilities being sacrificed.
For instance, live, on-demand business data exploration and analysis can be delivered and performed, for example.
Secondly, database file groups can be divided between virtualized all-flash drives, SAN drives (storage area networks) and flash-cached drives.
Thirdly, frequently accessed data pages can be auto-detected and hot-zone cached within SQL Server 2012.
Furthermore, dynamic SQL Server 2012 VM migration from one server to another can be accomplished through VMWare vMotion.
Add to that end-to-end mirroring, high availability (HA) and fault tolerance to server and flash failures, and all bases are covered. The Accelerating MS SQL 2012 with OCZ Flash Virtualization white paper has all the details.
“Our PCIe Z-Drives and VXL software distribute the flash between VMs based on need, making sure that no VM inefficiently occupies flash when it can be better used elsewhere in the SQL Server 2012 environment and that the flash cache itself is optimally used regardless of how many VMs are running concurrently,” said Ryan Petersen, CEO of OCZ Technology.
“Even though the flash cache resource is located in one server, it can be shared across multiple servers making VXL the only fast cache and virtualization software that can deliver this unique capability, especially when VMs do not typically reach peak workload requirements concurrently.”