Not Every All-Flash Array is the Same – AccelStor Briefing Note

All-Flash arrays are more than just the go-to option for data centers looking to solve performance problems. For many organizations they are becoming the mainstream primary storage system. But not every all-flash array is the same, and their differences impact just how far the flash array can scale, how much performance it will deliver and how much it will cost.

The Types of All-Flash

Generally there are three types of all-flash arrays. Initially all-flash arrays derived from DRAM appliances. These systems use proprietary hardware and interfaces in order to achieve the absolute maximum level of performance. But the use of the proprietary hardware made these systems expensive and difficult for the vendors to transition to new technologies as they became available. These systems are simply too expensive for the mainstream use cases and their performance is often overkill.

The second type of all-flash array uses traditional array hardware and enclosures. It replaces hard disk drives with solid state drives (SSDs), but does little optimization to the software to better handle the nuances of flash technology.

There is a third type of all-flash array coming to market that leverages advanced software technology with optimized hardware. The software accounts for an all-flash reality and optimizes the way data is written to and read from flash storage. AccelStor is an example of a company delivering this type of technology.

Introducing AccelStor – NeoSapphire All-Flash Arrays

AccelStor’s first set of products are its NeoSapphire All-Flash Arrays. The company provides a range of solutions and configurations to cover a wide variety of use cases. AccelStor’s portfolio includes an entry level system, mainstream systems, high density systems and NVMe systems. AccelStor has all-purpose systems for a variety of mainstream workloads, virtualization heavy workloads, HPC workloads and extreme high performance environments. The company claims some of its high-end systems achieve 700k sustained IOPS (random R/W mix).

The heart of the AccelStor product portfolio is their FlexiRemap Technology. The software is a system-level remapping layer in the operating system kernel that maximizes parallelism by spreading the IO workload across multiple SSDs. It also includes a flash-optimized journaling technique for power loads protection. FlexiRemap will coalesce inbound random write traffic shaping it into a sequential write which further increases performance.

FlexiRemap also replaces the need for a RAID type of data protection strategy. AccelStor claims it provides more efficient and effective data redundancy without the performance and lifespan penalties common in traditional RAID approaches. It can automatically redirect workloads when a drive fails and it enables improved system lifespan since it integrates with FlexiRemap’s global wear leveling techniques.

Thanks to the efficiency of FlexiRemap, the solution does not need memory caching or NVRAM, which means it has excellent power-loss data protection and won’t suffer from NVRAM bottleneck when the system is under duress.

In addition to the flash optimization FlexiRemap brings, the system also provides all of the enterprise features IT professionals expect from a mainstream system; like thin provisioning, deduplication, snapshots, clones and snapshot backups. It also includes asymmetric active/active High Availability for mission critical workloads.

StorageSwiss Take

The most common method for improving the performance of an all-flash array is to increase the processing power within the system. The problem with this approach is that it is both expensive and does little to eliminate the real culprit of performance loss, latency. More processing power just makes the time time between latent moments go faster.

Another approach is to create smarter software that better organizes writes and optimizes flash performance, thereby reducing latency. With better software the flash vendor isn’t forced to use more expensive CPUs and they are solving the core issue, latency reduction. AccelStor’s approach improves overall performance and data protection while still providing all the features that IT professionals expect in a mainstream system.

Twelve years ago George Crump founded Storage Switzerland with one simple goal; to educate IT professionals about all aspects of data center storage. He is the primary contributor to Storage Switzerland and is a heavily sought after public speaker. With over 25 years of experience designing storage solutions for data centers across the US, he has seen the birth of such technologies as RAID, NAS and SAN, Virtualization, Cloud and Enterprise Flash. Prior to founding Storage Switzerland he was CTO at one of the nation's largest storage integrators where he was in charge of technology testing, integration and product selection.

Tagged with: , , , , , , , , , , ,
Posted in Briefing Note

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 22,245 other followers

Blog Stats
%d bloggers like this: