Take Software-Defined VMDK Management to the Next Level
Posted in tech
While vSphere Storage Based Policy Management (SPBM) makes it easy for administrators to automate the provisioning of VMDKs to storage that meets its needs, admins still face the same performance problems when data needs change. Data virtualization and objective-based management complement SPBM to automatically align VM needs with storage supply.
To illustrate, let’s consider a case involving Jeff, an administrator for a virtualized environment at a Fortune 500 company.
When Jeff receives a ticket that a VM with a Gold storage policy is suffering from performance problems, lack of visibility leaves him few options. Since he does not have visibility into which VMDKs are the root cause of the problem, the best he can do is locate the Gold storage device under load and then guess which VMDK(s) are safe to Storage vMotion to another Gold storage device. This process can take some time, which results in frustrated end-users and creates a lot of anxiety for Jeff.
Jeff is always under pressure to ensure high VM QoS levels. Because troubleshooting is so difficult and remediation so painful, he commonly overprovisions storage by nearly double. This is beginning to put a strain on his budget, but he doesn’t see a workable alternative.
After researching online, Jeff discovers data virtualization and starts to see how it can solve his problems. By unifying different storage resources within a global dataspace, data virtualization gives enterprises the ability to non-disruptively move data across different storage types from flash, to SAN, to NAS, to cloud storage. Instead of manually verifying that storage devices labeled as Gold can support new workloads, Jeff can let DataSphere monitor how the devices are actually performing and do the work for him. He can do this simply, by defining business-oriented objectives in DataSphere that assign VM performance (IOPS, bandwidth, latency), protection (availability, durability, security) and price needs. These objectives are exported into vCenter, where Jeff assigns them to VMs just as he is used to doing. The difference is that DataSphere now automatically places the VMDKs on lowest cost storage that meets the VM objectives—and will move them, automatically, in the event that the applications need change or storage resources become contended.
This makes Jeff’s life easier in several ways. First, he saves budget, as VMDKs are deployed to the lowest cost storage that meets objectives and he no longer has to overprovision to protect VM QoS. Second, since DataSphere automates remediation in response to actual VM data access, problems are handled before they ever affect end users, so Jeff’s ticket count drops dramatically. Third, recurring fire drills become a thing of the past, leaving Jeff time and space to think more about how to enhance the system and to reclaim a personal life.
Data virtualization makes it easy to lead your organization in adopting VMware’s newest virtualization technologies to extend the benefits of virtualization into the storage serving your applications, using the resources you already have. To learn more about objective-based data management through data virtualization, contact us at firstname.lastname@example.org.