End of Life for All-Flash Arrays

Even though they currently dominate the market, the end of life for all-flash arrays (AFA) is coming sooner than you think. Let’s be clear AFAs are, in reality, single media storage systems that happen to only support flash. And in most cases, just one type of flash media. What is going to replace AFAs? The next generation of hybrid storage.

End of Life for All-Flash Arrays

The next generation of hybrid storage solutions will enable you to declare your independence from storage media lock-in. You, not your storage vendor, should choose the storage media that makes the most sense for the task at hand. The storage vendor’s job is to make sure their software exploits the strengths of your chosen storage media while compensating, as best as possible, for its weakness.

The ability to extract maximum performance from the performance tier (Optane or Flash) and maximum utilization of the capacity tier (QLC or HDD) means you can support all your storage needs from a single enterprise storage platform. Lowering TCO, efficiently delivering performance while taking a no-compromise approach to data protection is what StorONE is all about.

Why We Need Next-Generation Hybrid Storage

The first generation of hybrid storage arrays gave the hybrid storage concept a lousy reputation. To meet both performance and capacity demands while containing costs, vendors first-generation hybrid storage vendors use a small flash tier with hard disk drives. This first-generation solution is like they use too small of a flash tier, resulting in wildly inconsistent performance. Too much of the operational data is on the hard disk tier instead of the flash tier. The rapid movement of data from these small flash tiers to the lower tiers means that data isn’t written as sequentially to the lower QLC/HDD tier as it should.

Another source of inconsistent performance is the first generation software’s inability to extract high performance from each flash drive. Since, by definition, you want to use as few drives as possible in hybrid systems, the storage software must extract maximum performance from those drives to meet all the storage performance requirements from a few drives as possible.

End of Life for All-Flash Arrays

The first generation of hybrid storage did not immediately lead to the next generation of hybrid storage. Instead, the legacy software leads to the creation of All-Flash Arrays (AFA). AFAs are an expensive way to overcome software developer laziness. The problem is you, the customer, are left paying for that expense. AFAs are also the opposite of storage consolidation. Since the introduction of the AFA, data center storage is more fragmented than ever.  Now, since these vendors have labeled auto-tiering as evil, you are forced to buy an AFA for extreme performance workloads, an AFA for mainstream performance workloads, and even an AFA for backup and archive workloads.

End of Life for All-Flash Arrays

The AFA, as we know it, has no future. A storage system that only supports one type of storage media and forces customers to move data to an entirely different storage system manually reaches its end of life. We know different types of memory-based storage are on the way. We know that hard disk technology is still the most cost-effective storage in the industry for production workloads.  Your storage vendor should enable you to support all forms of storage media and use them in a single storage platform.

The Next Generation of Hybrid Storage Array – The Next Generation

In 2011 StorONE examined the problems with modern storage and its software. Once our analysis was complete, we created the storage system software and collapsed the storage stack to eliminate these problems.  The result is the StorONE S1:Enterprise Platform, a storage software operating environment that enables the full use of the performance and capacity potential of the latest hard drive and flash drive innovations.

To further improve performance and efficiency, StorONE also wrote a new Orchestration Algorithm that coordinates a smooth, parallel IO flow through the software. Our approach makes better use of multiple cores and CPUs while minimizing the number and power of CPUs required to provide high-performance. The combined effort is an environment that can deliver more performance from eight drives than most manufacturers can from 24 or 48 drives (Important in a hybrid infrastructure). The S1 platform delivers this performance without the need or want of a write cache.

To ensure proper data protection and data integrity, we also created a new type of data protection called vRAID. But unlike the previous RAID techniques used in the legacy storage system software, vRAID can rebuild large hard disk volumes built with 16TB drives in a record-shattering time of fewer than 2 hours and large flash volumes in less than 4 minutes.

What the Next Generation Hybrid Storage Should Provide

Ideally, the Next Generation Hybrid Storage Array should provide:

  • A large flash tier that is expandable to support “flash-only” workloads (Databases). It also enables 90% of hybrid IO to be from the performance tier
    • The flash tier increases data integrity and enables the storage software to write sequentially to the disk or QLC layer
    • Intelligent hybrid storage software that recalls data based on access type; transactional data moves to flash first, sequential data moves directly from HDD
    • Multi-protocol support for Block, NAS and S3
    • Efficient Storage Software
      • That enables 90% capacity utilization of all tiers.
      • That delivers hundreds of thousands of IOPS from eight flash drives
    • Complete Data Integrity and Protection
      • No Write Cache
      • Robust High Availability
      • Rapid RAID Rebuilds
      • Unlimited Snapshots that don’t impact performance
      • Flexible Replication (Synchronous, Asynchronous, Both)

Conclusion

A next-generation hybrid storage solution does more than make hybrid storage an acceptable alternative to an AFA. If you can deliver all the performance your environment needs at a price that is less than a third the price of an AFA and provides even more capacity, why would you ever want an AFA? It is the end of life for all-flash arrays.

To learn more, watch our on-demand webinar :

Posted in
George Crump

George Crump

George has over 25 years of experience in the storage industry, holding executive sales and engineer positions. Before joining StorONE, he was the founder and lead analyst at Storage Switzerland.

What to Read Next

Reduce Storage Costs and Risks

The primary objective of our Q2-2020 update is to reduce storage costs and risks. We’ve added two significant features that enable organizations to lower the cost of their storage infrastructure. We’ve also added two critical enhancements that better protect data both long term and during a disaster. StorONE’s Q2-2020 release is available to help you […]
Read More

The Problems of Using Primary Storage for DR

Most organizations count on their backup process for disaster recovery (DR) because of the problems of using primary storage for DR. Most backup solutions are far more flexible with what they can use as a replication target. The problem with using backup for DR is it wasn’t designed for the task. It often requires manual […]
Read More

Is Your All-Flash Array Underachieving?

Did you know that most all-flash arrays don’t deliver anything close to the performance potential of the drives installed in them? Underachieving all-flash arrays (UAFA?) are the result of vendors taking a shortcut to the market. They use inefficient open-source file systems and code for the foundation of their storage solutions. The problem is that, […]
Read More

Learn More About the Hidden Cost of Dedupe

  • This field is for validation purposes and should be left unchanged.