With the inexorable rise in data, structured and unstructured, flowing into organisations today, is it any wonder that traditional file-based and block-based storage systems are struggling to cope with the demand placed upon their storage capacity, and accessibility by more than just a small number of users? Whilst vendors of these types of storage systems are slowly taking on board scale-out architectures so that they can update storage back-ends and offer support to HTML-based protocols and flash, the speed of change in demand and the advent of the cloud has opened up a range of new storage possibilities including cloud storage, SDS (Software-Defined Storage), object storage and all-flash arrays. But what does this mean to you, the administrators of company storage systems?
Like it or not, the cloud has lifted the lid on numerous IT options and that includes storage. Cloud storage is principally a public service from third-party companies, i.e. Rackspace, although there are hybrid storage solutions that are able to combine external and internal cloud storage capabilities. The biggest advantages of cloud storage are its accessibility, usually via web platforms (although only accessible through storage APIs); its scalability up and down, according to demand and growth of the organisation, and there isn’t an upper limit; the fact that stored data can be accessed and share across different multi-level platforms, by different users; it’s simple and easy to use, and provides an extremely cost-efficient data storage method. However, there are drawbacks and one in particular is the security of private and confidential data when it is stored within an external cloud environment. Another disadvantage is that cloud is really only suitable for unstructured data, not for transactional applications or structured data.
As the demand for mobility across organisations has increased, so has the need for virtualised environments and infrastructure; enter virtualised and software-defined storage architectures. One of the biggest advantages of this form of data storage is the ability to be able to mix different platforms that vary in capability, size and performance. The VSA (Virtual Storage Appliance) version stores the software on a virtual machine and distributed as an image that can run on a server, for example, the Data Ontap Edge VSA from NetApp. These storage systems are usually found in satellite offices that don’t have the resources to run a range of hardware, but can run within a cloud. The SDS model incorporates the software stack, which is operated using cost-effective commodity parts.
If you’re looking for scalability with no limits, access anywhere, custom metadata support, a secure framework that supports multiple tenants, and cost efficiency then you should be considering object storage. Basically, data elements are stored as objects that are coded with customised metadata and unique identifiers which are easy to manage within an infinite object space. The storage nodes deliver storage and processing capabilities that can increase proportionally with the addition of new nodes, which are built off-the-shelf and collated together. If you have large amounts of unstructured data that needs to be accessed quickly by a wide range of users, then this is possibly the ideal storage solution and has developed into the favoured storage system for Web 2.0 applications. However, one drawback is that, like cloud storage, it isn’t suitable for transactional applications or structured data, plus object storage needs to be accessed via HTML-based protocols, i.e. REST (Representational State Transfer), but this does mean that the database can be access via any device, from any location.
The tried and tested solid state storage has had a wake-up call in the form new storage systems that are NAND flash-enabled, which has no mechanical components and is semiconductor-based. However, these forms of storage solutions are not readily available; they are costly, due to the NAND flash, and most systems available at the moment from companies such as Nimbus Data, Pure Storage and Whiptail, lack in enterprise features. That said, they are good for high-end applications where support for literally thousands of IOPS are needed.
If from reading the above you think you must go out and invest in the latest technological data storage solution, don’t… traditional storage solutions aren’t going anywhere fast; they are growing and changing in order to adapt to the current business landscape that now has the cloud and mobile computing to contend with. Don’t throw in the towel just yet, investigate the developments your storage solution provider is working on, it could be far more cost effective!