FutureNet World took place last week in London, and I flew in to join one of the panels. Of course, being in London is always a treat: I did some sightseeing along the Thames and the amazing Tower Bridge, grabbed some delicious food at Borough Market, and went to the pub for a beer.
But the highlight of my trip was FutureNet World, an event aiming to define the roadmap for the telco of the future in terms of network automation and artificial intelligence (AI). FutureNet World brought together more than 50 speakers from communications service providers (CSPs), who discussed their main challenges around the digital transformation journey.
Reduced Costs, Increased Revenue: Choose Both
One of the topics that was central to the event was the role of telcos in an environment that is changing – changing not just in terms of technologies and modes of operation, but also in terms of players and business models. All the speakers agreed that there is much to be learned from internet content providers (ICPs), who offer services from gaming to cloud computing with great agility and superior customer experience. Many of the conversations touched on 5G monetization and how to make sure CSPs can grab those opportunities.
It was interesting to note that while in the past discussions on automation revolved around improving efficiency and reducing operational costs, today the focus is on automation as a means of growing revenue and enabling new revenue streams.
Another hot topic at the event was that of AI adoption, and in particular the challenges of data, including data governance, i.e., the need for CSPs to define who are the people, processes, and technologies involved in data management to ensure its availability, usability, integrity, and security.
Keeping the Optical Network Transparent
Among all the interesting subjects discussed at the event, a topic particularly close to my heart is that of automation in the lower layers of the network.
For a couple of decades, the optical network has been almost transparent to CSPs. The established practice of overdimensioning and overprovisioning the optical infrastructure upon deployment and buildout, coupled with a slow uptake of resources, kept the optical layer in the background of day-to-day operations. But this approach is no longer viable. With bandwidth demands doubling every couple of years, shrinking margins, and much more dynamic traffic patterns, CSPs need an elastic network that scales quickly and adapts to the “on-demand” paradigm, creating a network where resources are shared and used efficiently.
Automation is the only way for a CSP to maintain that transparency of the optical network and focus on end-customer needs. Here are some examples:
- Automation is required to seamlessly support modern connectivity services that need to be activated on the fly with API-driven real-time service routing and provisioning
- Automation is equally important to enable 5G ultra-low-latency slices, where the choice of optical path and resource assignment needs to take the service requirements into consideration
- Automation in the optical layer ensures network reliability and adherence to SLAs, with fine-grained, real-time network monitoring used to detect transmission degradation and trigger restoration
- Analytics and machine learning can also be used to predict failures ahead of time and drive proactive decisions in a closed-loop manner
“Automate or die” was a sentence repeated a few times at FutureNet World – I dare to say that it applies to the optical layer as much as to the services layer.
Luckily, state-of-the-art optical transport devices like coherent optical engines, as well as optical layer devices such as ROADMs and amplifiers, are highly programmable, supporting truly dynamic operation.
Additionally, modern networking equipment collects a tremendous amount of performance monitoring data that can be exposed via streaming telemetry, offering much better visibility into the network and supporting more informed, intelligent decisions.
This detailed network information, together with device programmability, is what allows automation to translate intents, i.e., desired business outcomes, and map them to specific infrastructure configurations, extracting maximum value from the network. One good example is self-tuning transponders, which monitor and track transmission conditions and autonomously find their optimal operating settings.
Another topic of discussion -and some controversy- at FutureNet World was network disaggregation and the critical role of standards and automation in running these deployments. What’s the point of adopting an open and disaggregated networking approach, if you need to add on top of it an automation layer that brings together all pieces?
The motivations behind open networking are compelling: open networking unlocks faster introduction of innovation in the network, offers access to more choice including supplier diversity, and provides means to improve network economics, all conducting to increased competitiveness.
Automation consolidates and abstracts the open network functions towards an OSS, BSS, or orchestrator, reinstating the seamless management experience of a closed system, cashing in on the gains of open at no loss. Standards and architecture blueprints for the automation of open transport networks have seen recent advances, in particular with the work of TIP MUST.
Automation Going Green
Before I come to the end of this blog, I would like to highlight a final topic that caught my attention at FutureNet World: automation and sustainability. Green-networking is clearly moving up in the CSP’s agenda – as means to reduce operating costs, but also to address the ask of an increasing number of environmentally-aware end-customers, and to enhance the service provider’s social responsibilities. Network automation has an enormous potential to contribute to sustainability, both in direct manner, for example through autonomous monitoring and control of power consumption of networking hardware, but also indirectly, enabling innovative green-tech services and applications. This is a subject that I will keep a really close eye on in the future.
I hope I have raised your interest in the state-of-the-art of network automation. If you would like to look further into these topics, I recommend checking the compelling collection of articles, videos, and podcasts open to public access that FutureNet World makes available. A replay of my panel discussion on “5G Operational Transformation: How to leverage automation in the lower layers of the network” can also be found below.