‘Composable telecom’ requires transformation and dynamic processes

By Dave Duggal, February 17, 2016 

Originally posted on the TM Forum’s Inform website

The signs were everywhere at Action Week if you looked for them. From the keynotes to the sessions and the conversations at breaks. Ideas that once seemed subversive and encountered resistance appeared to be germinating and, while not mainstream yet,  taking root. It’s increasingly understood that old monolithic approaches are now getting in the way of becoming an agile digital business. By definition, the notion of ‘composable telecom’ requires a disaggregated model of a service provider.

Transformation is a phenomenon that occurs when there is a realization that established practices no longer meet new requirements. It always confronts inertia because change involves hard work and risk to arrive at a new operating model. It’s only when the fear of change is overcome by the fear of disruption (which is translated to a desire to lead) that transformation really happens.

The signs of disruption are everywhere from Uber, AirBnB, Netflix and Amazon Web Services; it’s almost become cliché to mention them. The most dramatic evidence of this trend was presented in a 2011 Deloitte paper called The Shift Index by John Hagel, John Seely Brown & Lang Davison, which was covered in Forbes as The Most Important Business Study — Ever?. It analyzed the “topple rate” for Fortune 500 firms noting the lifespan had declined from 75 years half a century ago to less than 15 years. Now consider that none of those aforementioned market leaders existed 15 years ago. This is not irrational exuberance – whether you call it the Fourth Industrial Revolution or Third Platform, it’s clear that the revolution will be digitized.

Predicting the future

OK, back to Action Week, and my reading of the tea leaves.

The meeting opened with a strong keynote panel featuring George Glass, Chief Systems Architect, BT, and Chris Boyd, Head of Digital Transformation, Telefonica, who both made clear that while they and their organizations are huge contributors to the Forum and other standards-development organizations, they aren’t sitting still and waiting on consensus – they’ve been out in front pioneering more flexible, extensible and scalable architectures. While their approaches varied on details, both companies have been busy decomposing their capabilities and processes, and it’s great for the Forum to showcase them as role models for transformation that the industry needs to embrace.

The overall conference agenda was smartly designed with lots of sessions titled some variation of “X impacts on Y”, which promoted introspection and debate. This is cross-functional systems kind of thinking is absolutely necessary.

As we’ve noted since joining the Forum, it has great domain assets that are highly evolved and widely adopted. The only problem is that they can be siloed and disconnected – SID, TAM, eTOM, etc. Moreover, while the conceptual architectures of each are professional, they are generally hierarchical in design and statically implemented, which is what currently impedes the agility service providers want. I am not the first to note this, but change has been slow. The good news is that today’s business demands are driving recognition of this problem far more effectively than the theoretical arguments.

Making progress

I was very pleased to hear Dave Milham, Chief Architect, Service Provider Engagement, TM Forum, open up the ZOOM sessions noting that they will now be focusing on orchestration. Two years ago we contributed a metamodel as a strawman to provoke thinking about dynamic, data-driven, policy-controlled orchestration. These concepts seem to have taken root and there was a recognition that “orchestration is everywhere”. Yes! Just as Obi-Wan Kenobe explains the Force to Luke Skywalker: “It surrounds us, penetrates us, it binds the galaxy together.”

Several people referred to ubiquitous orchestration as “dark matter”. Personally, I prefer to look at it as the white space between the boxes on our architectural models – within boxes, between layers, across silos, beyond boundaries. Also, I wish we’d raise this discussion from orchestration, which has an IT connotation of static, linear and centrally controlled, to process more generally. Orchestration is just a type of process, as are choreography and collaboration. When we talk about end-to-end processes it will include some combination of all three, and the coordination will be most certainly a dynamic process.

I was invited to make a presentation to a Frameworx Business Process Framework session on this topic. I’m pleased to report – there was no dogmatic resistance to these notions. We were able to have a rich discussion about conceptual architecture versus implementation architecture and the impact of hierarchy.

In software, hierarchies are a common form of passing down attributes as a form of “inheritance”. The approach is adopted from natural sciences – we’ve all seen zoology charts that classify the animal kingdom with a taxonomy of species. The problem with that is that deep hierarchies are hard to change when we discover new facts about something that might be low down in the hierarchy. Hierarchies, with their simple branching structures also constrain the nature of relationships. You end up having to re-do the whole chart, which is expensive. Software, virtual by nature, becomes rigid when designed in static hierarchies. It’s better to think about relationships in terms of graphs, which is a mathematical concept more closely aligned to computing.

As discussed in the session, hierarchies represent a conceptualization of a designer. In the case of a process, its context is fixed at the time of design. Today, if we want composable telecom, we are going to need dynamic processes. I proposed that it would be better to de-compose those hierarchies into a set of rules and relationships, which can be evaluated as part of the run-time time context for a process. In this way, we can “inherit” logical industry policies as part of generating an implementation graph. Voila! As we say in the States, “You can bake your cake and eat it too!”

This thinking can also be powerfully applied to the Information Framework, helping organizations implement the information model in a more modular manner. There were some early signals that this, too, is being recognized in Information Framework circles.

Taken together, these are significant movements that bode well for the industry. We are happy to work alongside our colleagues in the TM Forum contributing to this transformational effort. I look forward to TM Forum Live! 2016 in Nice and the next Team Action Week in Vancouver.