Virtualization (Network Function Virtualization – NFV) is the decomposition and decoupling of network functions from proprietary hardware to enable them to operate on off-the-shelf hardware to more efficiently utilize resources and increase agility in service offerings. This leads to microservices in Virtual Machines (VMs). Orchestration (Software Defined Network – SDN) adds the ability to automate the management of the network by recognizing the need to spin up or spin down virtual machines in response to traffic levels and support self-service provisioning of subscribers.
Studies by ACG Research and Analysys Mason on the benefits of virtualization of the VoLTE network showed these findings:
Services are created faster via software integration1
Time-to-market averages: 15 months to 6 months1
Overall savings of 33%2
While the promise of NFV/SDN is to reduce costs and increase service velocity it creates new challenges for service providers to maintain service quality during this transition. Monitoring dynamic virtual/ hybrid networks introduces a whole new level of complexity. There is a feeling of having even less control due to lack of visibility to transient VNFs. So the challenge is creating visibility to a virtualized infrastructure and providing seamless visibility across both virtual and physical network infrastructures.
Service assurance solutions are needed that holistically support hybrid networks with:
1) Visibility – down to VNF but with context of service and subscribers
2) Scalability – Carrier scalability for an NFV solution with real-time performance
3) TCO – Hardware is not free! The software solution must be optimized for the hardware footprint
Elastic and scalable monitoring solution dynamically allocated through the Orchestration layer. The virtual instrumentation follows the individual VNFs as they “spin up” and “spin down,” following “service chains” that compose the VNFs, and follow the interaction of the virtualized network components with the traditional physical network components.
Virtualized network functions means having holistic coverage for both north/south and east/west traffic. The “east/west” traffic coverage means having a virtualized probe that sits atop the hypervisor. With the virtual probe producing real-time information, that is needed for monitoring the “spin up” and “spin down” of VNFs, there comes the capability to provide its information as a feedback loop for the orchestration layer. To minimize the utilization of server resources the virtual probe (vprobe) must be built for software-only performance or else you will have to employ more hardware to accommodate performance monitoring (as Commercial off-the-shelf (COTS) hardware is not free!) And finally, the service assurance solution must support multi-tenancy to separate the respective traffic visibility that is inherent in virtualized resources.
Service providers cannot afford to lose visibility to virtual elements. Networks with VNFs that fluctuate in response to demand require a tool that can detect and resolve network and policy conflicts. To deploy virtualization with confidence and trust that these virtualized components deliver quality services, carriers need a solution that provides a comprehensive view of real-time traffic between migrating VNFs. Carriers need a single toolset with end-to-end correlation between physical and virtual network elements, allowing you to see the complete network picture.
1Source: ACG Research, Juli 2015
2Source: Analysys Mason, Juni 2015; savings realized with virtualized IMS adoption for VoLTE.