The demand for inexpensive online storage is booming as consumer media is increasingly digitized and kept in the cloud. According to Gartner, 36 percent of consumer content will be stored on the cloud by 2016, up from just seven percent in 2011. As this trend continues to skyrocket, the service providers that host this media will be required to balance ever-increasing demands for storage with profit margins and squeezed budgets. To find a happy medium, many service providers are looking for greater flexibility in data center architecture and control over hardware costs.
Software-defined storage is one way to solve these problems. A software-defined approach to data center architecture moves features found in hardware to the software layer and frees providers from the dependency on server “appliances” with software tied in.
The Move to Software-Defined Storage
Although the term “software-defined” may seem like a recent trend, everyday electronic devices have been software-defined for years. For example, software installed on a PC can be installed on many different hardware platforms, allowing the end-user to customize both the hardware and software to meet their needs. An operating system like Linux can be installed on any PC, which allows for greater freedom to allocate budgets precisely as needed for specific tasks – whether an intense graphic design setup or a simple Web browser.
Even though there are definite benefits in flexibility, data centers have yet to embrace software-defined technologies. As is often the case, the upfront costs required to switch existing data center infrastructure to software-defined is delaying widescale adoption of the trend. However, given the sheer level of infrastructure in data centers, the outlays needed to switch systems can be very high indeed. Yet, there is no denying the increased rates of online content storage.
Data Centers Today
Existing data centers are comprised mainly of appliances – server hardware with proprietary, mandatory software baked in. The software is designed for the hardware and vice versa, and comes packaged together in the same box. This can be a benefit for understaffed data centers that lack the tools and skills required to set up an in-house custom server deployment. Yet since hardware inevitably fails, appliances traditionally include extra layers of identical hardware to anticipate and prevent failure. These multiple copies of expensive components create higher costs in energy usage and add layers of complication to each appliance. Since the actual cost per appliance is much higher than commodity servers, cost estimates often balloon outward when companies try to scale up their data centers.
Data center administrators are beginning to consider software-defined storage approaches because of the aforementioned problems with appliances. “Software-defined” is not a new concept, but to the data center ecosystem where appliances with mandatory, pre-set software are the norm, a software-defined approach is almost revolutionary.
Software-defined solutions typically generate the following benefits:
Not all data centers are the same. A major bank with branches in several countries will have different storage needs from a telco servicing one small area or a cloud services host provider. While appliances might meet most of these needs, the benefits realized by abstracting the software from the hardware can provide substantial gains in economy of scale.
Software-defined storage gives administrators the freedom to customize specific components and software to best support their growth goals. This approach does require technically trained staff, but the flexibility afforded by software-defined storage delivers a simpler, stronger and more tailored data center for the company’s needs.
Convenience comes with a price. Traditional appliances offer convenience in the form of an all-in-one package. However, the added costs stemming from several layers of complicated software and back-up hardware can lead to significant cost hikes for a data center that needs to scale rapidly.
On the other hand, software-defined storage frees the software from the hardware and allows administrators to choose inexpensive commodity servers. When coupled with lightweight and efficient software, commodity servers can deliver substantial savings for online service providers looking to accommodate their users’ growing storage demands.
Market demands can cause changes in corporate priorities, budgets and network environments. An appliance-based, rigid storage environment locked into configurations determined by an outside vendor severely hampers an organization’s ability to react nimbly to market demands, much less anticipate them proactively.
Ever-increasing demands for cheap storage, and a continual reliance on expensive, inflexible appliances in their data centers will force companies to allocate significant funds to develop the storage capacity they need to meet rising customer demand.
Software-defined solutions offer companies the ability to “future proof” their data centers. Since the hardware and software are separate entities, each may be upgraded independently to a more appropriate option as the market dictates, at minimal cost.
The Globalization of Storage
Software-defined storage can also benefit international companies in a global way.
Since cloud services need to be readily available in locations all over the world, service providers must use global data centers to minimize load time. However, global availability presents a number of challenges. For one, load is active in the data center in a company’s region. This creates a problem, since all of that global data must be in sync. Also, companies are often required to restrict global data storage from either leaving certain countries, or being stored in others. Additionally, global data centers must be resilient to power outages and other localized issues that put a server farm offline. Finally, if a local data center or server does go down, global data centers must quickly reroute data to their available servers and minimize downtime.
While there are existing solutions to solve today’s problems, they do so at the application layer. Attempting to solve these issues that high up in the infrastructure hierarchy presents significant costs and issues with complexity. Instead, solving these issues directly at the storage level can save both time and costs.
This is just the beginning for software-defined storage. Many companies are beginning to explore the future of data center implementation in response to ballooning market demands. For data center administrators facing these types of challenges, a software-defined approach to storage could be the logical answer.
Edited by Blaise McNamee