borzaya - Fotolia
Storage IOPS contention can be a workload performance deal breaker. With VMware Storage I/O Control, administrators can allocate storage to VMs so that they don't compete for resources and create IOPS bottlenecks. Storage I/O Control is certainly not new to VMware vSphere 6.5, but the Storage I/O Control mechanism that vSphere 6.5 uses has been redesigned to operate using VMware vSphere APIs for I/O Filters framework.
The vSphere APIs for I/O Filters (VAIO) framework processes the I/O traffic that passes through the ESXi hypervisor. VMware and third-party developers can create various filters -- code segments -- that intercept and process I/O requests between the guest OS and the virtual disk. This means I/O is always filtered before it is issued or committed to disk. In effect, VAIO-based filters are the basis of VMware's policy-based management mechanism. Third-party vendors have made caching and replication filters available for some time. With the redesign of Storage I/O Control to use VAIO, VMware's quality of service filter now provides policy-based management for storage latency.
Once an administrator enables Storage I/O Control for a data store and selects a proper latency and throughput level, he can apply Storage I/O Control configuration preferences in a VM Storage Policy dialog box and assign that policy to specific VMs or disk files, such as a .vmdk files.
An example of Storage I/O Control operation occurs with common business operations such as month-end report generation. This usually requires inactive reporting applications to suddenly spring into operation on the last day of the month and crunch numbers to generate business reports. The sudden spike in Storage I/O Control can easily affect other workloads in the enterprise while the reports are being generated. However, administrators can use Storage I/O Control to create a storage policy and I/O limit rule for the VMs that run the reporting workload and any associated databases. If you limit I/O, it will cause the corresponding workloads to take longer to complete demanding tasks, but the reduced I/O demands will ease contention with other workloads and prevent changes in performance.
Storage policies are treated as dynamic entities -- you can create new policies and apply them to VMs and disk files as needed, and modify existing policies on demand. If a policy doesn't effectively ease performance issues, or the policy limits are excessive and adversely affect the workload to which the policy is applied, you can revise the policy simply by changing the IOPS limit. Storage policy changes take place almost immediately.
Don't let unaligned I/O kill storage performance
Monitor IOPS storage consumption with QoS Management
Eliminate resource contention in hyper-converged environments
Dig Deeper on Using monitoring and performance tools with VMware
Related Q&A from Stephen J. Bigelow
Regression tests and UAT ensure software quality and both require a sizeable investment. Learn when and how to perform each one, and some tips to get... Continue Reading
Learn the meaning of functional vs. nonfunctional requirements in software engineering, with helpful examples. Then, see how to write both and build ... Continue Reading
Just because software passes functional tests doesn't mean it works. Dig into stress, load, endurance and other performance tests, and their ... Continue Reading