Given this context, the ability for service providers to offer innovative new services quickly is difficult given the constraints of the existing networks and the corresponding constraints for the operation, ongoing maintenance and business approach and models of those networks. As a result, Communications Service Providers (CSPs) are actively pursuing network functions virtualization (NFV) with a software defined infrastructure (SDI) on a telco cloud to offer innovative new services faster, with scale to end-consumers and businesses, and with more effective business and operational models.
The key tenets of NFV based on the European Telecommunications Standards Institute (ETSI) and other standards specifications for using NFV to offer communications services require the utilization of common and horizontal approaches for media software that are elastically scalable on the underlying off the shelf high performance servers in the telco cloud and data centers. As such, these NFV tenets require CSPs to provide media processing in the network core and edge using standard data center server resource pools with scalability of the media transcoding virtual network functions (VNFs) on demand and in an automated fashion.
An effective way to provide media processing support for multiple communications applications is to utilize a common media processing approach in the NFV cloud that the different applications can all utilize and control using a common approach. This also implies that the applications have separated control and media processing to allow for independent scaling and distribution. These applications utilizing common, pooled media resources could range from media resource function processor (MRFP) for VoLTE and other IMS applications, session border controllers for IP interconnects and access, session routing and control, voice mail and WebRTC to name a few.
In order to successfully execute media processing in the cloud, a focused effort on achieving high performance in terms of input and output is needed. This is due to the fact that both media processing and transcoding tasks are being completed in software versus using purpose built digital signal processors. As such, using open approaches such as Open vSwitch and the Data Plane Development Kit (DPDK) with their large open source ecosystem continues to be critical in achieving these high performance requirements by taking full advantage of the underlying hardware. By combining these software-based acceleration techniques with servers using Intel® architecture, greater performance has been realized than by using them independently.
The advantages of supporting a separated architecture for control and media processing also allows for the easy distribution of media VNFs to edge data centers leaving the control VNFs of the applications to run in the core data center. The interworking between the core data center and the edge data center would be via software defined networking (SDN). This is especially important given the sensitivity of media to latency and delay.
The separated architecture of control and media processing not only allows for independent scaling of control plane and data plane processing, but also allows for distribution and hybrid deployment approaches. Hybrid approaches provide CSPs with flexible deployment options and the ability to introduce network virtualization along with their in-network deployments while providing flexible migration options from one to the other.
Optimal media processing in the NFV cloud requires the ability to horizontally and elastically scale media as a VNF under a management and orchestration (MANO) and VNF Manager framework. This enables the option to utilize the media VNF as a common resource across multiple applications for optimal resource usage. Intel® architecture and open performance acceleration approaches help achieve optimal performance for the media software in the data center. The ability to separate and distribute control and data plane processing and independently scale the two planes is a requirement to allow CSPs to deploy media processing at the edge. Hybrid deployment options combining NFV with in-network deployments provides additional flexibility to service providers.
GENBAND is a global leader in real time communications solutions with more than 700 service provider and enterprise customers worldwide. Our NFV solutions offer our customers industry-leading elastic and carrier-scale virtualized media transcoding capabilities for multiple real time communications applications including various mobile VoLTE (Voice over LTE) applications and Distributed Session Border Controllers (SBCs). GENBAND’s NFV solutions use the latest Intel® architecture and technology to allow service providers to automatically and independently scale control and bearer plane VNFs with elasticity for optimal NFV deployments in service provider cloud environments.
We will be showcasing our NFV capabilities at the NFV World Congress taking place in San Jose California, April 19 -22. I am looking forward to be presenting “High Performance Virtual Network Functions” during the Intel Network Builders Summit at the conference. If you are attending the show, I encourage you to stop by GENBAND stand # 19 to see firsthand some of the innovations in NFV we are bringing to market.
Sanjay Bhatia is Vice President of Solutions Marketing & Strategy at GENBAND. You can follow him on Twitter at @SanjayBhatia8.