Software-defined storage or SDS is a storage architecture that abstracts storage resources by separating storage software from its hardware platform. Earlier, Anand Jayapalan had spoken about how this leads to improved scalability, efficiency and flexibility. Even though it is not easy to intuitively understand how developers can separate storage hardware from software, the practice of abstracting one from the other is pretty common. SDS applies this specifically to data storage. Hybrid cloud environments would be one of its most relevant examples.
Hybrid cloud platforms are able to combine private and public server space through a combination of software and network orchestration, and the user of the cloud space experiences the platform as a single entity. In a similar manner, software virtualization makes use of programming and application techniques to abstract storage hardware in order to present a unified interface for end users or other platforms. Under a SDS solution, multiple dispersed servers hosting varied types and volumes of hardware spaces are able to function as a single storage entity.
Software-defined storage abstracts the hardware used for storage from any specific service, rather than abstracting several types of storage. Even though the benefits of virtualization are still prevalent, SDS takes the whole thing a step further by enabling more advanced storage services like deduplication and snapshots to apply across diverse types of hardware. Scaling in and out of storage media is much more fluid as the underlying storage hardware is abstracted. Moreover, the combined storage power of these servers is able to support a diverse range of storage types, including:
- Scale-Out File Storage: This is the most common form of SDS and tends to use abstraction to scale with traditional file and application storage that people are most likely to see on a typical computer.
- Scale-Out Block Storage: This is much like scale-out file storage when it comes to scalability, but applying to block storage. Block storage involves the packaging of information into blocks with distinctive identifiers, while enabling rapid retrieval during execution. It is ideal for cloud applications.
- Scale-Out Object Storage: Very similar to scale-out block storage, but leverages the benefits of modern object-based storage with metadata for optimal scalability of cloud computing applications.
- Container-Based Storage: As opposed to the use of dedicated virtualization software for wide-ranging scalability, containers are known to offer flexible and abstract storage solutions that are closely tied to their associated containerized micro-services and applications.
- Hyperconverged Infrastructure (HCI): Hyperconvergence implies to the practice of combining distributed storage, RAID servers, computing resources, as well as server processing power into a single platform.
Earlier, Anand Jayapalan had underlined that SDS is popular for enabling developers and administrators to coordinate a specific type of storage across hardware. It tends to facilitate improved automation and simplified management, standardized interfaces and higher scalability as well as transparency for monitoring and managing storage. SDS is generally more agile and cost-effective in comparison to traditional network-attached storage (NAS) and storage area networks (SANs) approaches towards to storage. It allows for the use of standard x86 based storage hardware.
