Software-defined storage is definitely gaining popularity, however there are still misconceptions about what it is and how organizations can take advantage of this new storage technology. One of the most important things buyers must understand is that unlike traditional storage – which often came as a single, expandable box – you can’t simply purchase software-defined storage. It’s an architecture. You need to build it. You purchase software and hardware independently and assemble the components in the way that best meets your business needs.
Here are a few other common misconceptions about software-defined storage (SDS):
- It’s just like old-school storage but with special software! Traditional storage solutions limit you to proprietary hardware. This means you purchase software preinstalled and tightly coupled to hardware. As a plus it’s guaranteed to work properly (usually), but as a minus it is inflexible. Some of today’s software-defined storage solutions are incrementally better, but you have to read the fine print. They will sell you commodity hardware – or even let you choose from a limited list of options – but the goal is still a tightly controlled environment where the software is mated to a particular hardware profile. True software-defined storage solutions decouple software from hardware and allow customers to use whichever off-the-shelf server and disk components best fit their needs.
- It’s in the cloud, which means I don’t have to buy any hardware! Although many software-defined storage can be used with both private and public cloud solutions, it’s not one-size-fits-all. The public cloud is usually slower than on-premise storage and can get very expensive at large scale. The cost to store the data, retrieve the data, and protect the data can come with a dizzying array of additional fees. Look for software-defined storage that allows you to mix and match server infrastructure and run instances of the same software in a public cloud. Using a combination of on-premise infrastructure and public cloud services enables you to fine tune the economics and treat the cloud in an optimal way: as a DR or cold storage tier.
- I don’t need any trained people to run it! Storage used to be a domain exclusively reserved for highly-trained specialists. If you didn’t have the requisite certifications then it was, for all intents and purposes, black magic. Modern storage platforms change all that. They’re designed with the post-storage-specific-admin world in mind. That’s not to say you don’t need storage skills, but it is fair to say you don’t need a dedicated storage team. Done right, anyone can configure and use it. Software-defined storage solutions simplify storage infrastructure so that anyone who knows how to use cloud computing services like AWS can also use storage. This way, app developers, DevOps, server and virtualization admins, and IT ops teams can all work with storage without the need for a dedicated storage admin to step in.
At the end of the day, software-defined storage is not a single thing that can be purchased, it is an architecture made out of many elements. Here’s a very distilled list of the basic components you’ll need to buy:
- SDS licenses – the brains and heart of your new architecture. This will be the perpetual or subscription (or both) licenses to the storage software, the brains and heart of your new system.
- Commodity server hardware – the brawn of your new architecture. You’ll mostly select x86 servers (but we support ARM, and even GPUs are on the near-term horizon!) with 6 or more hard drives.
- Commodity Ethernet network – the nervous system of your new architecture. Most customers already have the requisite capacity. But you’ll want to make sure you have the right LAN (and WAN if you’re clustering across multiple datacenters and clouds) characteristics based on your performance and availability needs.
When choosing the best storage solution for your business, you have to consider the specific and varied needs of your organization – now and in the future. For example, do you need the performance characteristics and IOPS profile delivered by flash/SSD media? Is there a need for object storage to serve a global content repository or active archive? Is API programmability needed to support custom applications or rapid deployment models and self-service portals? Do you require multi-site data protection to ensure disaster recovery? Everyone’s needs are different. Before taking the leap to purchase your own software-defined solution, be sure that your chosen SDS has flexibility to meet your wide-range of requirements. Look for the following:
- Flexibility: support for a comprehensive feature-set across server and media types (e.g., x86, ARM, HDD, flash), storage protocols (e.g., block, file, object), efficiency capabilities (e.g., deduplication, thin provisioning), and protection tools (e.g., replication, snaps/clones)
- Elasticity: the ability to scale-out or shrink the environment – incrementally and non-disruptively over time – by simply adding or removing servers and disk
- Automation: the simplification of storage processes, such as provisioning, rebuilds, tiering, and load-balancing, and refreshes
- Programmability: the delivery of seamless storage capabilities to application programmers via rich RPC and REST APIs
- Adaptability: support for any modern compute environment including public cloud infrastructure and modern apps like Hadoop, OpenStack and Docker
The value of software-defined storage lies not just in coupling low-cost commodity infrastructure, it also includes the ability to eliminate islands of storage by having a single platform that truly lets you define storage as you need it, when you need it. Have questions? Want to learn how software-defined storage can help you? Just click below.