The introduction of server virtualization not only led to more efficient use of existing servers, but also enabled substantial changes in the ways IT professionals provision operating systems and applications.
Virtualization allowed for a tremendous decrease in the time that was necessary to accomplish those tasks, decoupled hardware upgrades from software upgrades and even gave rise to the methodologies known as DevOps.
Network function virtualization takes the next logical step, virtualizing capabilities and functionality that were traditionally delivered by the network. Much like server virtualization provides agility in OS and application provisioning, NFV leverages software-defined networks to provide agility in network configuration. With the ability to provision network devices, such as load balancers and firewalls, DevOps becomes NetDevOps and creates a framework to rapidly deploy entire applications from the ground up.
What're the differences between NFV vs. SDN, how do they interact, and how can businesses use both of them to improve network performance and services as they incorporate an increasing number of devices and deliver enhanced value to their customers?
What Are the Benefits of Software-Defined Networks?
SDNs are powerful tools for network configuration. By separating the network configuration operations from the data transmission, an application itself can define the network it needs by specifying a set of rules or policies. In that way, an application can ensure that appropriate bandwidth and proper security objectives are achieved. SDNs also offer valuable secondary benefits by reducing the amount of time network engineers devote to mundane changes, allowing them to focus on more meaningful activities.
Implementing SDN technology and procedures does not require the presence of NFV and provides operational benefits to network engineering departments in their own right. From a broader perspective, SDN allows for automated management of network devices. Administrators can apply configurations at a central point rather than attempting to ensure consistency across myriad devices.
How Does Network Function Virtualization Work?
The initial NFV concept was introduced by the European Telecommunications Standards Institute in the white paper “Network Functions Virtualisation: An Introduction, Benefits, Enablers, Challenges & Call for Action.” While the paper was written from a telecommunications industry perspective, it brought forth the NFV concept and illustrated a number of points that become highly relevant when deploying NFV in any business.
First and foremost, an NFV implementation will scale only if all of its constituent components are automated. The key value proposition of NFV is to increase the velocity at which IT services are deployed; manual processes are anathema to that goal. A second important point is reliability and uptime: Network devices fail, just as servers do. Virtualizing network devices allows rapid provisioning of replacements with minimal downtime.
NFV isn’t so much a technology as a methodology for deploying networks and services in a virtualized environment. It consists of three pillars: infrastructure, management and functions.
The first, infrastructure, comprises all of the physical components, servers, routers and storage that make up the network. The management pillar describes how the network is managed. (It should be noted that while NFV doesn’t specifically require the use of SDN to manage a network, doing so with traditional manual methods will erode any benefits of NFV.) The functions pillar highlights where those virtualized network devices, load balancers, firewalls and routers live.
Another way to think of NFV is through the lens of containerization and microservices. With this view, a network device becomes a microservice, implemented within a container. When NFV is executed in this manner, existing container provisioning systems and microservice creation practices can be leveraged. Network functions in a container can be updated, patched and redeployed in a continuous delivery and integration cycle. This approach can also contribute consistent security and logging configurations. Deploying NFV in this manner lends itself to the first principle of IT: Keep it simple and well understood.
NRF vs. SDN: How Do They Intersect?
Like white or red wines with finely aged cheese, SDN and NFV perform best when paired together. Virtualizing the application infrastructure from the bottom up, wrapping it with automation tools to deploy, monitor, self-heal and scale up in response to user demand is the fervent vision of technology leaders.
Server virtualization is already well established, with best practices that are widely acknowledged. Add a layer of containerization and microservices, and the application layer becomes highly agile. Virtualize the firewalls, routers and other network devices to bring agility to the network functions layer, then add a healthy dose of SDN to manage the network devices themselves. Garnish with a bit of automation, and you have the ingredients needed to build and deploy highly responsive and resilient IT services.