Organizations can leverage both SDN and NFV solutions to move toward a more software-centered networking model. Rather than being competitive technologies, these solutions can often be rolled out side by side in a complementary design. SDN is a “big picture” solution that determines network policies guiding the delivery and use of network resources, notes Equinix. Meanwhile, NFV comprises a “wide range of specific functionalities,” such as routers, firewalls and SD-WAN. Both solutions are centrally managed, and both can deliver cost savings and enhanced flexibility for network administrators. 

With SDN, much of the cost savings comes from the ability to automate network configurations and changes — which in turn allows organizations to redeploy staff to other projects, reducing operational expenses. Because personnel costs account for such a large chunk of network spending, Equinix notes, even a small reduction in this area can lead to a significant cost benefit. NFV addresses the other side of the cost equation, eliminating the need to purchase specialized hardware for individual network functions. Along with cutting equipment costs, this reduces requirements for power and cooling, as well as the data center footprint. 

Both solutions allow data center administrators to quickly pivot in response to changing demands. The programmable interfaces of SDN solutions enable simple provisioning of new network devices and reconfiguration of existing devices. NFV gives IT shops the flexibility to quickly deploy and decommission functions to support proof-of-concept trials, and to locate functions at the network edge to optimize network security and performance. 

For years, organizations have leveraged server virtualization to use software to create multiple virtual machines on individual physical servers. This ability to increase the use of physical infrastructure leads to benefits including reduced hardware costs, faster provisioning and deployment, a reduced management burden and significant energy cost savings. Legacy applications sometimes prevent organizations from moving to an entirely virtualized compute environment, but most enterprises have by now widely adopted server virtualization. Once a novel solution, the technology is now approaching universal adoption, with 92 percent of businesses running virtual machines in their data centers, according to Spiceworks. An additional 5 percent of organizations plan to adopt the technology by 2021. 

The advent of the public cloud has further separated compute power from on-premises hardware, with many organizations investing in Infrastructure as a Service (IaaS) solutions that require essentially zero management on their part. And increasingly, organizations are looking into containerization to make their workloads nimbler. Containers, often managed via an open-source system such as Kubernetes, are essentially very lightweight virtual machines that allow data center administrators to deliver code without additional “overhead” such as guest operating systems. This can lead to an even more efficient use of data center resources than traditional server virtualization, requiring fewer racks, less energy, fewer software licenses and less maintenance.

For a time, many observers in the IT industry made a distinction between SDS and storage virtualization — with storage virtualization referring to the pooling of multiple storage arrays into one or more logical containers, and SDS referring to a software-defined delivery of features such as deduplication, snapshots and replication. Today, storage virtualization is more commonly understood to be one of the central components of SDS. 

SDS solutions don’t separate storage from hardware but rather add a software layer between the physical storage and any data requests, giving organizations more ability to manipulate how and where data is stored. This added control allows data center architects to build out SDS infrastructure using commodity storage hardware, leading to improved flexibility and reduced costs. Organizations that decouple storage software from their hardware have greater agility to expand storage capacity when needed, enabling more strategic planning that improves outcomes and controls costs. By emphasizing storage software over hardware, these organizations also enable a more flexible posture for upgrading or decommissioning hardware.

Other key features and benefits of SDS include automation, standardized application programming interfaces for management and maintenance, the ability to scale out infrastructure without impeding performance, and transparency into which storage resources are available and at what costs. 

For many organizations, the first large step toward SDx is an investment in HCI. These solutions incorporate compute, storage and networking in commodity hardware, making them a simple way for businesses to reap the benefits of software-defined data center infrastructure. One chief draw of HCI is that it is modular and highly scalable, which allows organizations to deploy just a few nodes to start and then increase their investments over time in response to changing demands. Because HCI typically takes far less time and effort to stand up than traditional IT infrastructure, organizations can rapidly expand their environments when needed. 

The simplicity of HCI can also provide cost savings, in part because of decreased management burdens. Some organizations report that their total cost of ownership can be 30 percent lower for an HCI cluster than for comparable traditional infrastructure. Additionally, HCI’s scalability enables organizations to fund their investments essentially through an operational expense model — allowing them to make only a small upfront investment and then pay for additional resources as needed. This approach mirrors the subscription or pay-per-use financing model of public cloud resources; and, indeed, HCI is a central component of many organizations’ hybrid cloud environments.