With legacy storage architectures, overprovisioning has long been necessary to ensure performance, availability and capacity. But, with increasing rack space shortage and pressure on data centers to reduce power consumption, overprovisioning storage to achieve performance is no longer a viable solution. Further, legacy disk based storage architectures mandate clear physical separation of storage for running different VM – server and desktop – workloads for the performance characteristics of the workloads are very different.
Tintri VMstore and its application-aware architecture lets you consolidate multiple VM workloads on a single storage system. It delivers simple, cost-effective performance on a per-VM basis, so you can stop worrying about storage placement for every VM. It automatically delivers performance for every VM as needed, and gives a clear picture of how much performance headroom is left on each node.
Traditionally, providing the right level of infrastructure for virtualized applications requires painstakingly careful planning. Although VMs can be provisioned and moved fairly easily, there are significant implications for the underlying infrastructure — particularly storage. Enterprise IT has to tackle storage placement carefully to make sure each VM gets the performance it needs, often provisioning the data for a single VM across multiple tiers. The most viable solution for many enterprises has been to vastly overprovision storage to limit the risk of performance degradation and the complexity of managing multiple service levels.
With Tintri VMstore, delivering the right level of performance is easy. The Tintri file system automates performance allocation, but administrators can also pin individual VMs or virtual disks to flash to prioritize workloads. Virtualization and storage teams don’t need to waste endless cycles planning for every new set of VMs, working through the minutiae of how to best architect for virtualization. As a result of the built-in quality of service VM workloads with different performance characteristics can all be consolidated on to a single storage system for efficient resource utilization and management simplicity.