Not sure what Telecom vendors are thinking lately. In the last couple of weeks I sat in a couple of webinars of major industry vendors, where they pitched some great slide-ware only to admit at the end they haven’t quite figured how to do what they described.
Today, a presenter said, and I quote “Technology for this has not yet built, just the beginnings in academia – the goal is tying orchestration engines to policy engines, policy drives orchestration automatically/dynamically.” The objective is spot-on, but by that account, sounds like all current Network Architects will be retired long before Carrier transformation can take place.
Here are my two gripes: A) they could have labelled the webinar accordingly and noted that in the description, or at least the outset of the presentation….; and B) just because you can’t do it doesn’t mean it doesn’t exist!!!
We’ve been demonstrating Metadata-driven Policy-Controlled Orchestration with CloudNFV for well over a year. CloudNFV was the 1st ETSI PoC and we even took the cake at TMF Live for “Most Innovative” catalyst project in Nice last June.
On a related note, their is yet another SDO initiative to work on “intent-based” systems, but from the threads and the admissions of participants themselves, it seems to be floundering. The self-assessed problem, SDOs focus on interfaces because they are an explicit deliverable, but the bottom up approach focused on syntactic integration always leads to static, siloed implementations – we’ve seen this movie before.
What the world needs now are new architectures for managing diverse and distributed end-points dynamically. This is a Computer Science problem, not just a Network Engineering issue. This requires high-level abstractions that present a unified view of heterogeneous application, process, data, network and UI resources so they can be reasoned over as if they were the same enabling virtually centralized policy-management across silos and layers – that’s what we’ve been done.
Carriers – you can start testing advanced SDN and NFV use-cases today at low-cost, low-risk and roll-out solutions incrementally, based on success. We have sites already testing C-RAN and the boundaries of edge computing.
As William Gibson famously said, “The future is already here — it’s just not very evenly distributed.”
Originally posted on LinkedIn