Why Scale-out Matters

even_capacity

Recently we surveyed IT organizations and service providers running SolidFire all-flash storage systems, asking them why they chose SolidFire storage over all other storage platforms, and 87% of enterprise customers and 73% of service providers said they purchased SolidFire because of its scale-out architecture.

There are numerous benefits to using a scale-out system vs. a scale-up system (check this out if you need a refresher), and in this blog series, I will explore several of these benefits, including: nondisruptive linear scale, forklift elimination, cost avoidance, and the storage reshaping capabilities of a SolidFire all-flash array.

ucezu.jpg

Nondisruptive linear scale: You said I could expand with ease!

Tell me if this sounds familiar: You’ve outgrown your current storage platform, so now it’s time to buy something better and faster. Before you buy that new storage platform, you do the necessary due diligence, spending countless hours examining various applications and workloads running on your current storage system. You review your Excel spreadsheet to get the storage usage trend for the last few years so you can account for growth. You calculate used capacity down to the last MB — you’re ready. You give that capacity number to your vendor storage rep and wait for the latest and greatest storage system to arrive.

Fast forward 16 months, and after two expensive capacity expansions and a few days worth of unplanned downtime, your storage system is once again running low on capacity. Now application managers and developers are complaining of poor performance, along with not having the disk space to expand. Your storage vendor informs you that you’ve reached the expansion limit for your array, and the only option is to purchase another array.

The above scenario is a common side effect of infrastructures utilizing a legacy scale-up storage system. Traditionally, a storage system was designed to support a specific application’s needs. Over time, as storage capacity began to deplete, additional capacity would be added until hardware limits were reached. Performance suffered from this scale-up approach, however, due to the limitations of dual-controller architectures as the system scaled up.

SolidFire storage systems are immune to these pitfalls because they are based on a scale-out architecture that combines all-flash performance with the capacity and cost effectiveness of disk, while allowing you to granularly add capacity and performance linearly without impacting production workloads. This approach makes a SolidFire all-flash system suitable to a broad set of enterprise applications.

Scaling with SolidFire

Let’s revisit the same scenario but now with a SolidFire all-flash storage system:

You calculate the used capacity of all applications and workloads down to the last MB, giving a capacity number to your SolidFire account manager so he can provide you with an all-flash storage system that will meet your needs. As workloads are added to the SolidFire cluster, the built-in storage efficiencies begin to compress and deduplicate data, drastically reducing the amount of actual data being written and stored to flash.

With the unexpected additional capacity and performance, you begin to move other business applications to flash sooner than expected, where they too flourish. When the need for more capacity arises, you simply scale out the cluster by adding in the 1RU node of capacity and performance required — and nothing more.

low_capacity

even_capacity

A small storage purchase and zero downtime was all that was needed to add capacity and performance to the current all-flash platform. Sound too good to be true? With SolidFire, it’s not.

Stay tuned for the next scale-out topic: eliminating forklift upgrades.

2258 views 1 views Today
2258 views

 | 0 comments



Posted in Next Generation Data Center.