Operationalising NFV: A guide to market best practices
It’s now well understood that the entire telecoms industry is technologically and operationally migrating to virtualised network functions utilising network functions virtualisation (NFV) as the architectural framework. Many operators have already deployed NFV and those that haven’t are advanced in their planning. It’s no longer a question of if the telecoms industry will transform to virtualised environments but the questions of how and when are being determined by the unique situations of individual telecoms operators.
Of course, any conversation about timescales requires a common view of what constitutes NFV transformation and this also is open to interpretation. This transformation is happening right now but full virtualization of operators’ entire infrastructures remains at least several years away and we are entering a period in which virtualized and traditional infrastructure will exist side-by-side. This in itself presents significant challenges that are potentially far greater than operating exclusively virtual or entirely physical networks.
The hybrid era, during which operators will further operationalise their virtual infrastructure, requires a top down approach that sees full engagement with virtualisation from the chief executive down through the organisation, encompassing each of its functions. This top down vision is augmented with a practical approach to ensuring virtual and traditional systems can operate effectively in parallel during the transition.
The enablers of this vision are management and network operations systems (MANO) backed by real-time analytics. The analytics are critical because, without real-time insight into what capacity and support infrastructure is available where and when, dynamic services such as those that will be delivered by 5G cannot be delivered with any efficiency. Of similar importance, the hand-off between statically provisioned, function-specific hardware and flexible, virtual infrastructure needs to be continually analysed to ensure first service availability and quality and second optimised resource utilisation.
Automated healing and real-time analytics are also required to provide the operator with the visibility it needs to run and control the network. Inevitably this will be done automatically because of the sheer volume of interactions required to help the network self-heal and change functions as demand dictates. The goal is to deploy systems that can ingest, correlate and calculate to ensure the right measurements are provided to orchestration. Automation is vital because the traditional approach of passing events through a help desk to fix issues will not be fit for purpose in the world of real-time service delivery.
Typical examples of areas in which NFV is being deployed are virtualised customer premise equipment (vCPE), virtualised evolved packet core (vEPC), software defined wide area networks (SD-WAN) and virtualised IP Multimedia Subsystems (vIMS). These areas have been selected because they either offer rapid and clearly identifiable operational cost savings, or because they are new and require operators to invest to offer a service, such as vVoLTE. These areas present relatively simple to achieve early wins in which operators can demonstrate the benefits of virtualisation. They are also teaching operators the best practices for operationalising NFV.
One operator that has used NFV to enable it to launch VoLTE services is Ooredoo Kuwait. The operator has deployed a single, unified cloud to support its NFV and IT applications. The deployment is part of Ooredoo’s UNIFY initiative which aims to utilise software defined data centre architecture and SDN and NFV to virtualize operations. Demonstrating the collaborative nature of NFV deployments, Ooredoo Kuwait partnered with its VNF and cloud vendors to deploy an IMS VNF in a test environment within three months. The VNF was then – following successful testing – seamlessly transferred to Ooredoo’s production IT environment enabling it to conduct its first VoLTE call.
Another example of vIMS deployment is at Manx Telecom, an operator located on the Isle of Man. The operator has deployed a virtualized fixed line IMS solution which is fully integrated into the core of its existing telephony infrastructure and enables it to deliver its entire suite of multimedia residential and business communications. Taking a top down approach, Manx Telecom and its partners, applied agile workflow philosophies from the very beginning with the initial design phase running in parallel with early integration testing. The project lifecycle was just ten months and along the way the operator was able to address major obstacles inherent to IMS integration and the deployment of NFV.
North American operator MetTel have placed virtualization technology in its core network. MetTel uses VMware as an NFV platform running virtualised network functions (VNFs) created by software defined WAN (SD-WAN) specialists VeloCloud. These enable MetTel to provide more trusted bandwidth to its customers, and enable greater flexibility by allowing the operator to utilise different transport mechanisms, configured on-demand. Use cases include deployment of VeloCloud's SD-WAN with the VMware Airwatch mobile device management system to enable an enterprise customer to see what is happening on each device, right down to the individual app.
These early deployments, though, must be undertaken with the longer-term picture in mind. Investments in NFV made today must integrate with future virtualisation efforts. It’s important that a modular, multi-vendor, multi-domain approach is facilitated because a key operator requirement is to avoid replication of traditional stovepiped siloes in the virtualized environment and becoming locked-in to a single vendor.
Virtualization therefore must enable NFV solutions from various vendors to interoperate smoothly and accommodate solutions from a different range of specialists from day one if all the benefits of virtualisation are to be accrued by operators. In addition, NFV requires a horizontally scalable platform that is open, extensible and can support multiple VNFs. The virtualised environment centres around working with a large ecosystem of technology providers and this represents a change from the traditional approach of using vertically integrated, inflexible hardware.
It can be of no surprise to learn then that the virtualisation market in the telecoms sector is so ripe and that the operationalisation of virtualisation is happening now. The 2016 Accenture Enterprise Survey confirmed that 95% of respondents believe their network services will be virtualised within the next three years, with, 33% already using such solutions.
NFV provides a crucial platform for continued innovation as operators continue to transition to being software defined. It doesn’t however guarantee success by itself. Achieving this requires an operator to take what they have learnt in their initial deployments and apply it to their business-wide strategies for complete virtualisation. Then, and only then will they be successfully able to react to new market opportunities and increase value in the eyes of the customer. Then and only then will they be able to compete and deliver the same quality, speed, scale and choice of competing digital disruptors.
- » Poland and the USA declare need for stringent checks of foreign telecoms gear
- » T-Mobile opens performance testing facility for 5G, NB-IoT, and more
- » Deutsche Telekom launches 5G in five German cities during IFA media day
- » Opinion: The 5G future of operations will be automated
- » Industry 4.0: Putting CSPs at the heart of global business