Most IT experts will agree that 2013 was the year that flash storage became a near necessity in the mainstream world of enterprise. The reason is that flash holds many competitive advantages over the tradition disk storage method, including faster speeds and reduced energy consumption. As a result many companies have turned to “hybrid systems”, where flash based solid state disks are combined with hard disk drives. Other companies have adopted flash based systems 100%.
While flash is not the only new form of memory storage, it is by far the best choice. Some people are working on other systems including phase change memory, carbon nanotubes, spin torque transfer, and memristors. None of these are close to being ready for use in enterprise and only time will tell if they will ever become a viable option.
Of course, flash storage is not without its problems. Among the biggest causes of hesitancy for enterprise has been the higher costs and other technical memory problems. Fortunately, some large scale flash providers have been able to eliminate these disadvantages.
No one would claim that flash storage is cheap. However, it is certainly on the decline. Large scale vendors have done an effective job of offsetting costs, thereby making it more attractive. One of the ways they have lowered prices has been by using consumer grade flash, which is significantly cheaper. They have also altered what is called forward error correction (FEC) to eliminate capacity penalties from RAID.
Several companies provide good examples. Dell, for example, was recently ranked as the top provider for flash/SSD technologies to enterprise. They also played a significant role in reducing costs. They claim that their flash system offers 2 times the price advantage of competitive hybrid solutions and 5x the advantage of competing all-flash solutions. In fact, their all-flash solutions costs less than a comparable 15k disk drive solution. Essentially, they are offering flash storage for the price of disk storage.
Cisco is also becoming a big player in providing affordable flash storage to enterprise. The company recently acquired Whiptail, a New Jersey-based memory card company. At the beginning of the year, they announced their new product, the UCS Invicta, an all-flash storage appliance. The system is primarily meant to be used for i/o data storage as well as transfer capabilities.
HP is another large-scale vendor that claims to have closed the cost gap for flash storage. Last years, the company extended its 3Par enterprise level storage into a flash only system. Its software was already familiar to those who had been using the 3Par system for storage, making the transition easy. It boasts fast speeds and latency even for storage arrays that have not been specifically built around the SSD drive. While still not incredibly cheap, their pricing is much less than flash storage used to be.
As flash based systems began to be more common, cost was not the only problem people were afraid of. One of the biggest fears was that of NAND flash wear out. Simply put, this wear out process is the result of repeated program and erase processes that gradually wear out NAND flash memory. NAND uses what is called floating gate transistors to store an energy charge for long periods of time. When data is erased, the oxide layer within the floating gate transistor is broken down and the ability to store memory is lost. Large-scale vendors have made efforts to eliminate this problem and in some cases have been successful.
With the rollout of any new system, there are small hiccups. However, as flash array storage becomes more reliable and affordable, there is little doubt that it will become a powerful tool for enterprises.Back to top