Why Automation is Key in Cloud/NFV Environments
Everyone knows that network functions virtualization (NFV) is important. It’s more than just moving network infrastructure functionality to a virtualized environment, which is how it looks on the surface. It’s also about orchestrating, delivering and managing end-to-end services. It does this by chaining together software-based network functions in an automated and programmatic fashion. So, how exactly does it work, and what are the benefits? According to Dialogic, automation is key in Cloud/NFV environments.
NFV is great in that it allows operators to automate many tasks that, in a traditional network architecture, were either cost prohibitive or not feasible. There are several areas within Cloud/NFV environments in which automation plays a big role. For example, automation is useful in onboarding a virtualized network function (VNF) into a data center cloud. It can also be applied to make scaling an application easier. It does this by automatically reserving, configuring and turning up virtual compute and storage capacity; it also loads additional network function resources into those virtual machines. But, again, how does all of this work?
To automate the network infrastructure cloud, for example, in NFV, software-based functionality is decoupled from the hardware, and a VNF Package is used to onboard and activate the application. The VNF Package includes descriptors, artifacts, software images, scripts and policies to enable onboarding and automated lifecycle management of an application in a cloud environment.
Next, the VNF Package is then placed in the VNF Catalog by the NFV Orchestrator (NFVO)—the brains behind the NFV—where it is consumed by the VNF Manager and the resource and service orchestration functions.
From there, the VNF Manager (VNFM) works on interfacing the NFVO, the VNF and the Virtualized Infrastructure Manager (VIM). This is where the VNF Package is important, because it comes back into play by communicating with the VIM layer in order to automate the instantiation of applications within the NFVI. In order to communicate with the VIM, the VNF manager can either directly or indirectly access it through OpenStack API calls or Heat Orchestration Templates for lifecycle management workflows. Once communication has been established, the VNFM processes information on value changes of VNF-related indicators to figure out if a scaling event needs to occur. If needed, KPIs are available for automatic scaling.
All of this ultimately creates an environment where COTS servers and white box switches provide compute, storage and networking resources. Within the new framework, virtualization technology—like KVM or VMware—in the compute nodes provides the hardware abstraction that is needed for additional NFV flexibility.
All in all, automation is a key player in this process. NFV is beneficial in that is results in lower CAPEX and OPEX and even allows operators to be more flexible and agile. These positives are made possible entirely through automation.
Edited by Maurice Nagle