Clustered ONTAP QoS - Performance Isolation for Multi-Tenant Environments

A NetApp storage cluster running clustered Data ONTAP can be subdivided into distinct storage virtual machines (SVMs), each governed by its own rights and permissions. SVMs make multi-tenancy possible by securely isolating individual tenants—for instance, in a service provider environment—or individual applications, workgroups, business units, and so on. Each application or tenant typically has its own SVM, and that SVM can be managed directly by the application owner or tenant using the SVM.


Any time you put numerous workloads on a storage system or storage cluster there is the possibility that excessive activity from one workload can affect other workloads. Storage QoS workload management available in clustered Data ONTAP 8.2 allows you to define service levels by creating policies that control the resources that can be consumed by individual files, volumes or LUNs—or entire SVMs—to manage performance spikes and improve customer satisfaction. QoS gives you the ability to set not-to-exceed performance capacity limits (defined in terms of a maximum value for IOPS or MB/s) on a group of files or volumes or LUNS within a SVM, or on the entire SVM. This capability will enable enterprise IT administrators and service-providers to consolidate many workloads or tenants on a cluster without fear that the most important workloads will suffer or that activity in one tenant partition will affect another. As a result, you can push your storage infrastructure to higher levels of utilization, increasing your overall efficiency and decreasing capital costs.


NetApp storage QoS works with both SAN and NAS storage, and it runs across the entire NetApp FAS product line from entry to enterprise. You can incorporate third party storage into your NetApp storage environment using NetApp V-Series. With V-Series as a front end to SAN arrays from EMC, HP, HDS and others, your existing storage gets the full benefit of QoS limits and other NetApp capabilities.

Storage QoS Use Cases

There are many possible use cases for Storage QoS. The following two examples illustrate some of the ways that QoS can be used.


Create Storage Performance Limits for Enterprise Workloads

The traditional approach to ensuring QoS for workloads with different service-level needs, has been to house each application in its own dedicated infrastructure silo. However, this leads to lower resource utilization and higher costs.


Storage QoS makes it possible to run multiple workloads on a more cost-effective, shared infrastructure by creating limits for each significant workload. This can be accomplished by assigning each application workload to its own SVM and assigning a performance capacity (IOPS or MB/s) limit on that SVM, or by assigning a performance capacity limit on the group of volumes or LUNs associated with an application workload.


Limit Maximum Storage Performance for Cloud Tenants

In a multi-tenant cloud environment—whether private or public—the first tenants on a particular resource may see a level of performance in excess of their service level agreement (SLA). This can create a perception of performance degradation as additional tenants are added and performance decreases.


Storage QoS allows you to avoid this problem by assigning a performance capacity limit to each tenant in accordance with their SLA. That way, a tenant can’t exceed the set performance limit, even when resources are available, and are therefore less likely to notice changes in perceived performance over time.


Another advantage of storage QoS is that it makes it simple to establish tiers of service based on SLAs. For instance, Bronze service may correspond to a limit of 10,000 IOPS, Silver to a limit of 20,000 IOPS and Gold to a limit of 40,000 IOPS.


Figure ) In private or public clouds each tenant is assigned a storage virtual machine (SVM) and a performance limit. You can provide different service levels by establishing tiered limits (Bronze, Silver, Gold).


For more information on QoS, see this lab validation from analyst ESG:


Mike McNamara