Magazine Article | February 13, 2013

Making The Case For Storage QoS

By Chris McCall, VP of Marketing, NexGen Storage

Similar to the concept of QoS (quality of service) in the networking realm, storage QoS lets you accelerate, control, and guarantee storage performance levels.

I recently had the opportunity to meet with a customer who is considering deploying a virtual infrastructure. Their existing deployment consists of high-performance servers connected to direct-attached storage (DAS). Ensuring adequate application performance has always been the key driving factor for their infrastructure strategy, but with 26 servers and more coming online this year, footprint, power, and cooling are becoming huge issues.

Enter server virtualization. The space, power, and cooling savings are obvious, so I asked the customer why they hadn’t done it sooner. His response was simple: “Performance.”

Consolidating multiple applications on a shared storage system meant that each application’s workload would impact every other application’s workload. Application performance needs could change at a moment’s notice, resulting in unpredictable chaos. The existing DAS implementation avoids this mess by dedicating storage resources to one, and only one, application. So while DAS addressed the customer’s performance concerns, the challenge of managing 30 servers with individual storage is pushing them to a virtual infrastructure with shared storage.

Given the switch, how does this IT administrator guarantee the performance levels that are critical to his company’s productivity? That’s where storage QoS comes in.

Managing Mission-Critical Application Performance
VARs are constantly looking for ways to monitor and tune shared storage systems and often over-buy disk to accommodate performance spikes. The vast majority of SAN (storage area network) and NAS (network attached storage) products on the market force the IT administrator to configure performance, rather than manage it. To do this the administrator estimates a workload, then sizes the storage system by the number of drives. They’re left with a single pool of performance that every application shares, with no way to assign resources or prioritize. Administrators are forced to buy resources they don’t need.

  • If the storage system is built on spinning disk drives, the administrator sizes the system by estimating the number of disk drives required to hit the performance requirement. This almost always results in excess capacity.
  • For solid state systems, capacity tends to lag performance. The administrator has to purchase by the amount of capacity required, resulting in excess performance.
  • Rather than continuing this decades-old paradigm, IT administrators now have the ability to use innovative management capabilities in the form of storage QoS to manage performance. Storage QoS allows IT administrators to guarantee performance for each application and isolate workloads from one another on the SAN.

For example, a company can assign 30,000 IOPS (input/ output operations per second) to its business intelligence app, 25,000 IOPS to its order database, and 5,000 IOPS to its marketing file shares. No matter what is going on in the system, each application will get, at a minimum, the targeted level of performance.

By setting these guaranteed minimum levels of performance, storage QoS eliminates resource contention within the shared storage system. With this capability, when one application spikes, the critical applications still perform at acceptable levels.

Adding VDI To The Application Mix
In the above example, I referred to business intelligence and the order database, both business-critical applications that, prior to storage QoS capabilities, have typically not been virtualized in order to avoid the likely performance contention. VDI (video device interface) is another example of an application with extreme performance demands. Most IT departments run VDI on its own server to isolate the impact of boot storms. However, with storage QoS capabilities, the administrator has the ability to control and guarantee performance for both mission-critical applications and VDI from a single storage solution, allowing them to take advantage of the consolidation benefits of shared storage.

Storage QoS — The Missing Link
The storage industry has spent the last few decades focused on managing capacity. By consolidating capacity resources into shared systems, customers saw overall cost per GB decrease and management become much more straightforward. However, sharing capacity also meant sharing performance. When x86 processing power exploded and virtualization allowed multiple applications to run on a single host, the concentrated performance workloads exposed a massive management gap; specifically, how to manage shared storage performance.

Storage QoS addresses this missing link, allowing IT administrators to accelerate, control, and guarantee storage performance. This capability makes it realistic to support mixed workloads on one SAN, including mission-critical applications, without compromising performance.