Enterprise storage arrays offer just about every bell and whistle that an organization can possibly want. Whether it is compression, deduplication, a mix of flash and disk, all-flash, connectivity to the cloud, multi-protocol support – you name it, an array out there probably has it. Yet in the midst of all this feature functionality and options on storage arrays, storage arrays have somewhat lost their way in one respect as many are failing to deliver on arguably one of their basic tenets: plentiful amounts of storage that is cheap, highly available and highly reliable. This explains the emergence of a new class of economical, massively scalable storage arrays.
Though DCIG is still in the early stages of examining this trend, these storage arrays break from the mold of other enterprise storage arrays in that they primarily focus on delivering high levels of storage capacity that is both highly available and reliable at a low price per GB . The low price associated with these arrays (dare I call these arrays cheap?) is now at $1/GB or lower. Further, I have spoken to providers such as Exablox where its list price per GB is about a quarter a GB. Equally impressive, its arrays scale into the hundreds of TBs.
The driving force behind arrays such as these and why end-user organizations want them is that they solve some key dilemmas for them. For instance,
- Not everyone is ready to store data in the cloud. While storing data in the cloud is now as low as a penny per GB per month, that cost will likely never go away. Further, that cost does not factor in increased and ongoing WAN costs, the delays associated with pulling data from the cloud and the data retrieval costs that the cloud provider may charge when the data is retrieved.
- Unknown value of data. Every organization has lots of archival, backup and even production data that it needs to store somewhere. Further, it often does not have a firm grasp on the data’s value. While the data may not be worth a lot, the individuals responsible for it know that its value is more than risking deleting it and losing their job.
- Cheaper to store it than manage it. Even if an organization could delete half, two-thirds, or even 90% of the data that it has, these storage arrays make storage capacity so cheap, why even manage the data? It is more economical to store the data than try to manage it.
Yet perhaps what is most impressive about these arrays is that there is really nothing notable about them, per se. They deliver lots of storage at a low price per GB. They use high capacity, moderately performing SATA HDDs. They connect to existing corporate networks using standard NAS interfaces (Ethernet) and protocols (CIFS and/or NFS). They are pretty much plug and play.
What does make them notable is the way in which these providers are making this storage available at such a low price. In the case of Exablox, you can buy a baseline system for about $10,000. Yet once in place, you do not have to buy marked up SATA HDDs from them. Rather, you can go out to CDW, Newegg, Tigerdirect, or any other online, discount computer retailer and buy the largest SATA HDD they have available for a few hundred bucks when you need the extra capacity. You then plug the newly acquired HDD into the Exablox system and, Presto!, multiple TBs of new, available capacity for pennies on the dollar.
Yet as these new HDDs are plugged into the system the data is not left unprotected. The system into which the HDDs are inserted give users the option to make these drives a part of a RAID configuration so organizations do not have to fret about the data’s availability or integrity.
Granted, these arrays are not the end-all and be-all of storage arrays nor should end-user organizations view them in that context. These arrays will, at best, provide moderate to low levels of performance and, for the most part, should not be used to store any applications for which performance is a concern. But for those organizations that need lots of cheap storage for the pile of data that they do not know its value but do know they need to keep it around, this new generation of economical, massively scalable storage arrays fits the bill pretty well.