Software-defined storage (SDS) is gaining traction in the enterprise as digital transformation initiatives move forward. Organizations see SDS as an opportunity to extend the life of existing storage assets, control storage costs, avoid vendor lock-in, reduce management complexity and add storage capacity without disrupting business operations.
SDS follows the software-defined model by separating the management software from storage hardware, which makes it possible to centrally manage and automate storage infrastructure. Resources from a shared pool are automatically and efficiently allocated according to application needs to maximize utilization. Because hardware is commoditized and all intelligence resides in management software, the risk of hardware interoperability is minimal. Administrators can define and apply policies through a single pane of glass for important storage functionality such as de-duplication, replication, thin provisioning, snapshots and backup across hardware from a variety of vendors.
There is no universal approach to or standard for SDS. Depending on the vendor you’re talking to or the article you’re reading, the definition could vary significantly. However, the Storage Networking Industry Association has identified five capabilities that SDS should have:
- Automated maintenance and management so admins can simply establish policies and monitor them
- A standard interface for centrally managing storage across a variety of devices
- Scalability through dynamic, automated provisioning that doesn’t affect storage or performance
- Consumption transparency, which allows admins to evaluate usage against resources and budgets
- The flexibility to store data from a variety of internal and external sources on different types of storage media
Most organizations have an enormous amount of data stored on legacy infrastructure. Migrating this data could be an expensive, time-consuming project. SDS allows you to continue to use legacy storage platforms, making it possible to knock down storage siloes and transition to the SDS model in phases instead of trying to do everything at once. You can even add new features to legacy systems without purchasing new hardware. But instead of buying more appliances when you run out of capacity, you automatically expand storage and compute resources to simultaneously increase available capacity and performance.
The lack of a universal approach or standard for SDS has led to myths and misunderstandings about the concept. Many people assume any solution with storage software falls under the umbrella of SDS, but many such solutions lack the basic features of an SDS platform. Another common assumption is that SDS creates a single, unified storage infrastructure. The truth is, block, file and object storage won’t perform as well when combined. Finally, while software is the star of SDS, hardware still matters. The better the hardware, the more effective the solution.
To increase the odds of a successful deployment, organizations must start by gaining a clear understanding of the current state of their storage environment. What are the costs? What are the capacity requirements? What problems need to be overcome? Develop a phased plan for shifting to an SDS platform, beginning with noncritical workloads and applications. Be sure to include data protection, backup and disaster recovery strategies to minimize risk.