Fusion-io has upgraded its ioTurbine flash caching software product to give customers more choice about where they want to apply caching intelligence in the server/storage infrastructure.
ioTurbine software runs on a physical or virtual server to provide intelligence to speed up application performance via better use of flash storage using tight coupling with input/output (I/O) routines to redirect read requests to server-based flash.
The product allows customers to control sharing of PCIe flash storage between virtual machines (VMs) and assign priority to the virtualised applications. It can adjust I/O and cache size when administrators create new VMs, shut down old ones or migrate VMs between different hosts.
Previous versions of ioTurbine operated at both hypervisor and virtual machine (guest) level. Now customers can also choose to apply caching at the level of the hypervisor only, the guest only or directly on a physical server.
Fusion-io vice-president for product marketing Lee Caswell said the changes brought more choice for the customer. “ioTurbine can now look at the ‘colour’ of data and cache at the level of the physical disk, logical volume, file or virtual machine,” he said.
“It sets the stage to intelligently look at data being stored and apply caching to it,” he added.
It enabling customers to choose where to apply caching is aimed at – for example, cloud operators that only want to control the hypervisor environment that customers occupy, but not their VMs. At the same time, the ability to manage VM caching, but not at hypervisor level, would be appropriate for cloud customers.
Flash caching products are a software adjunct to the wave of server flash hardware that has hit the market in the past couple of years.
While server PCIe format flash has become popular for speeding applications without the latency involved in data transiting to shared storage, flash caching software has helped add further intelligence to the PCIe flash deployment.