In 2012, seven of the world’s leading telecom operators selected the European Telecommunications Standards Institute (ETSI) to define a new approach for designing telecom networks based on software. At the time, I held an executive role at one of those operators advocating against adding proprietary hardware appliances and single silo services to an already complex web of legacy equipment in the core network.
The resulting ETSI white paper spurred the industry — operators and vendors alike — into action as the concepts of software-defined networks (SDN) and network functions virtualization (NFV) gained traction. The industry began to consider how to transform from legacy, hardware-based networks to exploit these visions of agile, programmable and software-based infrastructures.
However, without an accepted industry-wide roadmap, operators have adopted disparate approaches to this transformation that range from quite ;tactical and practical’ to ‘highly strategic and ethereal’. Many have tapped the telecom ecosystem for help, stimulating a proliferation of evolving standards and sometimes disjointed solutions. The journey to virtualization and network transformation has been circuitous, and the rewards have been so far largely elusive.
This is troublesome for operators who anticipated and need solutions to many of their fundamental issues, such as continued revenue declines despite growing subscriber bases, and the corresponding margin pressures. They require network solutions that future-proof their network and provide service agility for new innovative services, while also seeking elegant evolution paths without sacrificing network reliability. The costs to deliver the transformation while maintaining historical services and network reliability during this protracted transition period has exacerbated the pressure on capital and operational expenses. Operators want to navigate this transition more nimbly and efficiently and many are looking for guidance, particularly around operational excellence in the new model.
Operators certainly realize the promise of infrastructure modernization to achieve operational agility, flexibility and cost effectiveness. The decoupling of network software from proprietary hardware and server boxes liberates innovation as the software is no longer constrained by the physical limitations of the hardware layer. It enables rapid and continuous software experimentation and when properly architected and abstracted, introduces new paradigms for change management and reliability. Costs can be reduced and performance improved.
In its 2012 white paper, ETSI noted the communications industry will look and feel familiar to the IT industry one day. In fact, there are lessons for our industry to be learned from the transition of the IT industry from mainframes and mini computers, to cloud and social networks. For example, hard-wired networks with single functions boxes — like the mainframes of the 1960s and 1970s — are tedious to maintain, slow to evolve, and prevent operators from offering dynamic services compared to their modern analogs. Just look at the complexity of today’s network, as illustrated below, with many single-function, hardware-based applications. This kluge parallels the messy data center environments of the past. This is exactly what operators hope to move away from, but it has tremendous inertia. Ironically, as the telecom industry continues its transition to software-based infrastructures, this network complexity albatross could become the barrier to future success. The solution is in the journey.
This became clear to me a few years after retiring from AT&T, when I was advising a vendor offering a unique approach to simplifying the core network through an industry-leading virtualized software platform. As operators struggle with the deployment of NFV and SDN, they are turning increasingly to implementation partners such as portfolio players and system integrators to provide a broader view and implementation assistance. This can be helpful, but nonetheless, the journey from legacy hardware/ software silos to some future software environment remains arduous. A journey from an already instantiated and embedded software platform is obviously much less painstaking
Vendors that can provide bundled solutions of software along with industry data sources will enable unique opportunities for operators, ones that cannot be realized with traditional hardware-based approaches. The efficacy of a transition launched from a high performing legacy software platform toward a future software-defined end state was so exciting that I came out of a blissful retirement to draft onto this journey.
The transition to NFV and SDN is expected to continue well beyond 2020. It is both time and resource intensive, and many benefits are hard to realize in the near term. Immediate cost savings will be difficult to achieve because the technology and standards are still evolving. The transformation effort is complex as legacy and next-generation networks will coexist for many years. Trends such as the Internet of Things (IoT), containers and network slicing, and the transition to 5G will all add new complexities.
More importantly, this transformation isn’t just about technology, as it touches every part of the operator’s business. To be successful, operators must consider the implications of the transformation on employees, operational processes and product development. We need to guard against the ice man repairing the refrigerator phenomena. Platforms are a software game. Network transformation is both a network and a software challenge. Having spent many years in both of these disciplines, I offer the following for your consideration:
- This is a long-term migration requiring specialized skills
- Adopt an approach to bridge legacy and new network technologies
- Revisit your architecture: Migrating existing complexity into the NFV world will result in ‘virtual chaos’ with no cost savings.
- The signaling core is a perfect place to look for dramatic reductions in cost and complexity and a great place to start.
The bridge from a software platform to a software future is far shorter and less perilous then the chasm between the drawing above and the promise of an elusive future. May you have the hindsight to know where you’ve been and the foresight to know where you are going!