Today Enterprise Flash Storage arrays are being used in more applications and environments than ever before and is the No 1 choice for tier 1 storage. An all flash storage array is designed to provide the maximum number of IOPS for an application(s), whilst providing continuous up-time and data availability. Flash systems can deliver over 1 million IOPS and hold many PB’s of data, whilst integrating with your SAN using Fibre Channel, SAS or FCoE.
Performance is key when choosing flash storage, another key requisite should be reliability as flash drives suffer from wear of which there are three types:
No Wear Levelling – Nothing is done with the cells and it wears out
Dynamic Wear Levelling – Data is written evenly across the whole of the Flash
Static Wear Levelling – Works out how many times a cell has been written and dynamically moves it.
Flash stores data by use an electrical current to etch into Silicon a data bit and this causes Wear Levelling, whereby after so many programme erase / write cycles the Flash wears out and this could be 10,000, 100,00 or 1,000,000 writes depending on the type of Flash used.
Manufacturers overcome this problem in a number of ways by using sophisticated algorithms to work out how many times each cell has been used and then automatically re-map those blocks to another portion of Flash.
Flash storage array systems are certainly taking center stage with more technologies vying to become the dominant leader with Intel and Micron 3D Xpoint, HP and SanDisk memristor/ReRam, IBM PCM, along with SLC, MLC, TLC, QLC, 3D NAND, SSD, PCIe and NVMe. With SSD drive capacities announced by both Seagate and Samsung reaching 60TB and 15TB respectively.
The most common type of Enterprise Flash Storage uses SSD NAND, this is normally is MLC and provides a high capacity and a far lower cost point than SLC. Whilst this does not have the highest Write Cycle the more sophisticated Enterprise Flash Storage vendors take this into consideration and employ Static Wear Levelling to ensure the maximum data life-cycle is attained and reliability remains intact.
NVMe is an open logical device interface specification for accessing non-volatile storage media attached directly via the PCI Express (PCIe) bus. An NVMe device is on average 5x faster than a comparable SSD and 25x faster than a hard disk, whilst capacities today are 1TB.
The acronym NVM stands for non-volatile memory, which is commonly flash memory that comes in the form of solid-state drives (SSDs). NVMe, as a logical device interface, has been designed from the ground up to capitalise on the low latency and internal parallelism of flash-based storage devices, mirroring the parallelism of contemporary CPUs, platforms and applications.
The main issue with NVMe are as follows:
By its design, NVMe allows host hardware and software to fully exploit the levels of parallelism possible in modern SSDs. As a result, NVMe reduces I/O overhead and brings various performance improvements in comparison to previous logical-device interfaces, including multiple, long command queues, and reduced latency. (The previous interface protocols were developed for use with far slower hard disk drives (HDD) where a very lengthy delay in computer terms exists between a request and data receipt, data speeds are much slower than RAM speeds, and where disk rotation and seek time give rise to further optimization requirements.)
NVMe devices exist both in the form of standard-sized PCI Express expansion cards and as 2.5-inch form-factor devices that provide a four-lane PCI Express interface through the U.2 connector (formerly known as SFF-8639). SATA Express storage devices and the M.2 specification for internally mounted computer expansion cards also support NVMe as the logical device interface.
We provide flash storage array systems, support, configuration advice and evaluation systems for you to test in your environment. If you would like to know more please phone us on 01256 331614 or complete our online form.