What is Software-Defined Storage? Breaking Down the Buzz
11-05-2021 | By Robin Mitchell
The concept of software-defined stuff is taking over the world, with software-defined cars and software-defined networking making headlines. However, software-defined storage is another term that is beginning to rise, but what exactly is it and how does it leverage hardware like SSDs, HDDs, tapes, flash memory, RAM, and RAID arrays?
What is Software-Defined Storage?
When I first encountered the term ‘‘software-defined storage’’, I found it difficult to understand. Upon further investigation, the term seems more of a marketing buzzword, similar to the Internet of Things (IoT). When the term IoT came out, it left me confused as it sounded like a new paradigm of programming and device that somehow existed in their own bubble, but the truth was that IoT devices were simply, wait for it, internet-connected devices.
Depending on where you research, Software-Defined Storage, or SDS, will be described as a scalable solution that enables platforms to utilise open-source software and off-the-shelf hardware to create a storage solution that can save money and enable software systems to interact with storage intelligently. Now, for the real definition of SDS!
Software-Defined storage, simply put, is a storage solution that provides advanced storage mechanisms, but these mechanisms are defined in software. Another way to think of it is by replacing drive controllers with software. For example, RAID controllers allow drives to be used in different configurations, but an SDS would remove the RAID controller and implement this function in software instead.
What advantages does SDS have?
So, it is clear that SDS is something that must be important because major players such as IBM are investing in SDS, but what advantages does SDS provide? The first advantage of SDS is that removing dedicated hardware controllers means there is no need to purchase specialised hardware. This means a RAID drive system can be constructed from off-the-shelf parts, which can be a more cost-effective solution.
The second advantage SDS presents is that systems can be more easily expanded by removing the need for dedicated hardware controllers. For example, a RAID drive whose purpose is to backup data can only work with so many drives. A new RAID system may be required if more storage is needed, including the cost of a new case, disks, and RAID controller. Furthermore, the software is still needed to tie the two separate RAID drives together. HOWEVER, an SDS can work with as many drives that can fit on a network, and the software controls how these drives are handled.
The second advantage carries onto the third advantage of SDS; reconfigurability. A drive system controlled in software instead of hardware can be configured in many different ways, which can be upgraded over time. If a RAID controller is found to contain a bug or limitation, simply using a new RAID controller may not be an option.
The fourth advantage of SDS is the ability for targeted storage. Storage varies in technology and size, with hard disks being great for storing large files that are often accessed, while tape drives are ideal for large backups. However, the result of these different abilities is varying costs, so trying to manage data storage depending on cost is complex. An SDS would instead be able to automatically route data to store it on the most economical system, depending on its access rate and type.
Who is involved with Software-Defined storage?
A range of services is developing SDS systems in the IT and computing field. IBM is one contender that provides SDS solutions for various applications, including AI, big data, and servers. IBM has delivered SDS solutions and services to multiple clients, with The University of Queensland and Coventry Building Society being two examples.
Dell is another company that offers SDS solutions for clients with the use of their Dell EMC ECS. According to Dell, the EX range is scalable storage racks that integrate hardware and software with bays usable with SSDs and HDDs. Furthermore, the use of such racks can enable the use of SSDs as metadata cache that is required frequently while using HDDs for storage of larger files that are less frequently accessed.
The SDS industry is still in its infancy, and the advantages of SDS are still to be realised. However, the ability to replace the hardware with software does present real advantages. The question is whether this concept can be transported to the consumer sector and find its way into everyday computing devices.