If you’re looking for a concise, level headed analysis of the enterprise SSD market checkout Benjamin Woo’s August 31 Neuralytix report entitled Full Spectrum Solid State. It presents a nice breakdown of enterprise SSD solution form factors, dividing the market into:
- PCI expansion boards
- Solid State storage arrays without advanced data management (which he calls SSSD’s)
- Inline storage network cache’s
- Solid State arrays with advanced data management services (which he calls SSSS’s)
- Cache extensions for storage system controllers
- Disk drive packages meant to replace legacy disk drives
The report further divides the market into solutions that are “north” of the storage network (PCI boards and SSD arrays without advanced data management) and those that are “south” of the storage network (the rest) and gives a clear, intelligent evaluation of the pros, cons, and best uses of each.
Kaminario falls into the SSSS category, according to Woo, which he says has the most potential for challenging the hard disk storage systems market head on, especially given the TCO and data center real estate benefits of flash arrays.
Woo’s market predictions fall somewhere in the middle of the spectrum, disagreeing with both the assumption that SSD will crush the hard disk market and the other extreme that SSD is only viable for functions requiring the very highest I/O performance. Like others, he sees SSD gradually usurping high-performance disk storage, while SATA hard disk systems will continue to support applications, such as many Big Data applications, that need capacity more than performance.
Perhaps the most important point Woo makes, however, is this:
“When it is all said and done, software is key. Whether it is a firmware upgrade for a storage controller, as in the case of NetApp’s FlashCache, or the integration of an entire storage operating system found in SSSS, software is ultimately what will make or break flash memory.”
Elsewhere he says, “The software will aide in the proper distribution of I/O’s and proper placement of data to provide the desired return.”
Of course we agree, since that is what SPEAR, with its robust parallelization and advanced data management features, is all about. I would add, however, that it’s not only software, but architecture that makes the difference. To reap the full benefits of SSD you need a complete architectural redesign geared directly to SSD performance. SSD solutions imprisoned in legacy hard disk architectures, like many of those dual-controller solutions on the market, are doomed to fail the scalability and advanced data management tests, no matter how many SSD’s they’re packed with. Check out Purpose Built for Speed and Safety for more information on the K-2’s architecture, which scales performance as it scales capacity. So it’s not just the software, stupid, it’s the architecture as well. And you’re not stupid.
Tags: advanced data management, Benjamin Woo, big data, cache extensions, Full Spectrum Solid State, I/O, Kaminario, legacy disk drives, NetApp, network cache, Neuralytix, PCI expansion boards, Scalability, solid state storage arrays, SPEAR, SSD, ssd form factors, SSSD, SSSS