The leaders I meet with are rarely focused on storage—they’re thinking about larger strategic initiatives; accelerating product development, moving to IT as a Service (ITaaS), and deploying private clouds. But ironically, it’s often storage that’s the bottleneck for these broader initiatives.
Now, that’s not the fault of storage itself, but rather the fault of the logical unit number (LUN) and volume-based architectures on which most storage is built. LUNs and volumes were relevant (and originally introduced) when workloads were physical. Today though, more than 80 percent of applications have been virtualized and LUNs and volumes are clumsy units of management that obscure the performance and behavior of the applications on which your business depends.
That’s why we—and our customers—have been impressed with Virtual Machine-Aware Storage (VAS). In brief, VAS eliminates LUNs and volumes; the entire file system is built to take every storage action on individual virtual machines. That allows the entire data center to operate in a common language—virtual machines. And we’ve seen VAS deliver four key categories of benefit: simplifying action, enabling powerful automation, offering granular analytics, and all on a flash platform.
When you can take action at the VM level, you can do things that are not possible with LUN and volume-based storage. For example, you can guarantee the performance of a mission critical VM. You don’t have to assign resources to the LUN in which it resides, you can simply drag-and-drop the IOPS it receives to a minimum threshold. In the image below, the Input/Output Operations per Second (IOPS) are increased for a VM and in real-time the user can see the latency drop.
And of course, admins can also lower the max Quality of Service (QoS), in case one particular VM has gone rogue and needs a tighter ceiling to avoid sapping resources.
This is only one example of VM-level action. Admins can also replicate, backup and recover, and clone VMs with a few clicks. This granularity of control allows the organization to move faster by focusing on what matters—its virtualized applications.
The control that comes with VAS also allows for more automation. One great example is service groups. Admins gather sets of VMs into groups based on shared criteria. They can apply policies to all the VMs in the group, and then as they move individual VMs between groups, systems or geographies, those policies can follow them or be over-written based on requirements.
As mentioned, LUNs and volumes prevent admin from seeing VM-level performance and analytics. When one VM starts behaving badly, it can take hours or days (and plenty of time on the phone with your vendor) to understand the reason(s). But with VAS, you have full visibility of every VM across your entire infrastructure. You can literally just scroll over a VM on your dashboard to see a visual breakdown of its latency across host, network and storage. So, should a performance problem arise, you can identify and address the root cause in seconds.
Many VAS use hybrid-flash arrays. Tintri, which has been a difference maker in our experience at CDW, has hybrid-flash arrays that deliver 99 percent of I/O from flash. But Tintri also has an all-flash array: the Tintri VMstore T5000 All-Flash series, the only all-flash array with a VAS architecture. Furthermore, the all-flash array and hybrid-flash array share a common operating system and analytics, allowing admins to manage their Tintri units all from one dashboard.
Simplifying Your Storage
By removing LUNs and volumes from your data center you remove certain complexities associated with storage, thus saving time and money.
If your organization depends on virtualization to realize its larger goals of faster development cycles, ITaaS, or private cloud, you need storage that can be an enabler. VM-aware storage can give you better oversight of virtualized applications, which opens up your time to actually run a business—your absolute main priority.