Benefits of NFV
Earlier in this chapter the limitations associated with using the traditional network equipment were listed. Network functions virtualization directly addresses most of these restrictions and brings many additional benefits. It offers a framework to completely transform the way networks are architected, deployed, managed, and operated, while offering many layers of improvement and efficiency across all of these. Figure 1-18 lists a few of the benefits that NFV offers that are discussed in the sections that follow.
Figure 1-18 Some Benefits of Network Functions Virtualization
Because NFV uses regular COTS hardware, network operators have the freedom to choose and build the hardware in the most efficient way to suit their needs and requirements.
Hardware offered by traditional network vendors has very limited options for its computing, memory, storage, and networking capacities, and any modification leads to a hardware upgrade that costs time and money to the operators. With NFV, providers can now choose between many different vendors and have the flexibility to select the hardware capacities that are optimal for their network architecture and planning. For example, if the Internet gateway being used is running out of capacity to store the full Internet table and needs a memory upgrade, in most current implementations they can achieve this only through a controller upgrade or full device upgrade. In NFV, the provider can allocate more memory to the VM hosting this VNF.
Faster Service Life Cycle
New network services or features can now be deployed more quickly, in an on-demand and on-need basis, providing benefits for end users as well as the network providers.
In contrast to physical hardware, the VNFs can be created and removed on the fly. The lifecycle of VNFs can be much shorter and dynamic compared to physical devices, since these functions can be added when needed, provisioned easily through automated software tools that do not require any on-site activity, and then torn down to free up resources as soon as the need is over. This is in contrast to the deployment effort needed when a new function has to be added to an existing network, which would have required an on-site physical installation, which can be time consuming and costly. The ability to rapidly add new network functions (deployment agility) is one of the biggest advantages of NFV. Services now can also be commissioned or decommissioned with the touch of a button without the need of a delivery truck, drastically reducing deployment times from weeks to minutes.
Scalability and Elasticity
New services and capacity-hungry applications in today’s networks keep network operators, especially cloud providers, on their toes to keep up with the fast-increasing demands of consumers. The service providers have been playing catch-up with these requirements, for scaling up the traditional network equipment’s capacity takes time, planning, and money. This problem is solved by NFV, which allows capability changes by offering a means to expand and shrink the resources used by the VNFs. For instance, if any of the VNFs requires additional CPU, storage, or bandwidth, it can be requested from the VIM and allocated to the VNF from the hardware pool. In a traditional network device, it would require either a full device replacement or a hardware upgrade to alter any of these parameters. But since VNFs aren’t constrained by the limitations of customized physical hardware, they can offer this elasticity. Therefore networks do not need to be substantially overprovisioned to accommodate changes in capacity requirements.
Another way the NFV can implement elasticity is by offloading a VNF’s workload and spinning off a new instance to implement the same network function and split the load with an existing VNF. This too is not possible with traditional network equipment.
Leveraging Existing Tools
As NFV uses the same infrastructure as data centers, it can reuse and leverage the deployment and management tools already being used in data centers. Using a single centralized pane of glass for management of virtual network and virtual servers gives the advantage of quicker adaption for new deployments without the need for developing new tools and as a result eliminates the cost of deploying, familiarizing, and using new set of tools.
Rapid Development and Vendor Independence
Because NFV provides the means to easily deploy a different vendor’s solution without the heavy costs associated with replacing an existing vendor’s deployment, it keeps network operators from being locked into a particular vendor. Operators can mix and match vendors and functions, and choose between them based on feature availability, cost of licensing the software, post deployment support model, roadmaps, etc.
New solutions and features can be put into production rapidly, without waiting for the existing deployed vendor to develop and support them. Such rapid deployment is further facilitated by NFV’s inherent support for using open source tools and software.
Validation of New Solutions
Service providers often prefer to validate new solutions, services, and functions by deploying them in test setups, prior to introducing them in their production networks. Traditionally, they have had to replicate a subset of their production environment for in-house testing, which increased their operational budget. With NFV, building and managing such a test setups has become much more cost effective. The NFV-based test-setups can be dynamic and thus scaled and changed to meet the test and validation scenarios.
Amorphous Service Offering
An NFV-based deployment is not confined to a one-time design and deployment. It can adapt to market specific needs and offer a targeted set of services to match changing demands. Through a combination of elasticity and deployment agility, it’s possible to rapidly shift the location and capacities of network functions and achieve workload mobility. For example, providers can implement a “follow the sun network” by using constantly moving virtual machines based on time of the day, and spinning up or expanding new VNFs to meet the network’s requirements for services and capacity as they change during peak and off-peak usage or when major events take place in any geographic region.
Operational Efficiency and Agility
With common hardware hosting different VNFs, tasks associated with running the business, such as inventory management, procurement process, can be centralized. This reduces the operational overhead compared to segregated deployments of different network services using multiple hardware devices.
NFV is inherently automation friendly, and can increase the benefits that can be achieved through use of Machine to Machine (M2M) tools. For instance, its possible for an automation tool monitoring a device to determine the need for more memory in a network function. With NFV that tool can go ahead and request allocation of that memory—without involving any human intervention.
Network maintenance related activates can also significantly benefit from NFV by reducing possible downtimes. NFV allows for spinning up a new VNF, temporarily shift the workload to that VNF, and free up existing VNF for maintenance activities. This makes it possible to achieve In-Service-Software-Upgrade (ISSU), 24/7 self-healing networks, and minimize the operational loss of revenue due to network outages.