Now don’t get me wrong, I’m not denying that virtualization and cloud have played huge parts in the digital transformation journeys of communications service providers (CSPs). No; they’ve been real drivers of change. Virtual technology allows a business to be more flexible, agile and adapt to customers’ new needs and innovative business models.
But, while the benefits are great, they have given us an incredibly flexible infrastructure that is too complicated for human beings to manage because we can’t think or act that fast. Let me give you an example: I launch a new service, it goes viral, and subscribers flock to me in their droves. If I had to rely on human beings to do the necessary network reconfigurations, it would take eons, it needs to be highly automated.
Looking at the diagram below, you can see that we have these parallel phases of business transformation – the blue and red boxes showing that CSPs need to become both digital service providers themselves and digital enablers of others to reach the target state of the 2020 CSP – as described by Nik Willetts, CEO, TM Forum, in his keynote at TM Forum Live! in May 2017.
The phases of business transformation are enabled by two dimensions of technology (the grey arrows in the diagram below) which represent virtualization and cloud, then automation and artificial intelligence (AI).
We need to add AI into the mix, because it has become clear that automation alone isn’t enough. We have released the TM Forum Future Architecture Strategy Discussion Paper, which was developed as a result of work done in recent months with leading service providers – including AT&T, Bharti Airtel, BT, Telefónica, Orange and Vodafone – to establish a new, high-level vision and architecture for enabling digital business.
Relatedly, we asserted:
“In a traditional network, inventory is relatively static, but in an SDN [software-defined networking] and NFV [network functions virtualization]infrastructure, virtual machines providing communications services can be provisioned in seconds anywhere in the infrastructure. This means that a particular customer’s instance can only be identified by a real-time inventory management solution that is constantly updated.
“Intelligent automation of complex decision-making at superhuman speeds is required. For this, we must add AI [artificial intelligence]to control and operate the communications networks of the future.”
Machine learning and AI allow us to bring the pace of automation up to the pace of flexibility that we have in our infrastructure.
The problem with today’s OSS and BSS
The kind of agility we need is necessary not only in the network infrastructure, but also in operational and business support systems (OSS and BSS). The old systems are inflexible; they’re not real time and their DNA is all people and manual operations. Any changes needed to launch new services are complex, expensive and slow. These architectures and their levels of agility degrade over time as more and more changes occur, which is the precise opposite of what the business demands. A white paper by Infosys described this scenario well stating:
“Even for a well-managed group of systems, architecture decay occurs due to the various forces that act on these systems.
“What we need is a transformation strategy that addresses the problem areas of all stakeholders of the OSS/BSS ecosystem.
“A strategy that transforms the enterprise architecture to loosely coupled systems that are easier to build, test, manage, integrate, use and monitor. These agile systems should also have sufficient intelligence built-in to help analytics platforms to quickly retrieve relevant data patterns for smarter decision-making, be it for operations or for business.”
TM Forum’s Barry Graham, Program Director for Agile Business & IT, wrote a white paper, OSS of the future, which rethinks the role of OSS/BSS and is highly pertinent here (see excerpt from the paper below):
“The communications industry of today works on a timescale where network software is updated only every few months and new functionality requires extensive adaptation and testing. The OSS/BSS of the future is expected to work to a very different set of timescales for example:
- It must be possible to confidently onboard new functionality or services with minimal or no adaptation of the new functionality or existing systems.
- Onboarding, testing and verification of new functions or services should be extremely low cost (therefore automated).
- ‘Release frequency’, that is the waiting time for a new service to be deployed should be one day, and the cost overhead of a release should be reduced through automation.
- The concept-to-cash time for a new service should be as short as possible (target six months).”
Building on these efforts, TM Forum members are collaborating closely to create the open digital enablement system (ODES) – the name for our vision for the OSS/BSS of the future – which will address all of those problem points. It’s bringing together all of the work we’ve done in open APIs, platform architectures and ZOOM in a single vision (shown in the model below).
It’s this combination and single vision that make ODES stand out from any other support system architectures to date. It is:
- standards based, enabling a marketplace of commercial and open source innovation;
- a unified architecture for OSS and BSS;
- data centric with a single unified data plane which is fundamental to enabling automation and AI;
- platform-based and componentized using TM Forum Open APIs, which makes it flexible, open and agile by design;
- AI-capable and autonomous, which means it has event-driven adaptation; and
- ecosystem capable.
This approach is resonating with the industry. We spoke to Ryan Jeffery, an industry influencer and Founder of Passionate About OSS, who referenced the potential new architecture in a recent blog. He told us:
“There was quite a lot of overlap with the way I think about the OSS architecture of the future. One of main areas is in the closed loop thinking. It’s one of the spaces in OSS that can be improved upon as there doesn’t tend to be much closed loop thinking currently. What makes it really effective is the fact that it’s taking measurements and analytics, learning from them and using that to feed back into the system.”
You can read more about the specifics of closed loop thinking in this Inform article.
Join us at TM Forum Action Week in Vancouver, where on Tuesday 26 September, we will be holding an all-day open session to help shape the vision we plan to drive to consensus later this year.