One of the interesting contrasts at the recent Layer123 Zero Touch Automation conference in Madrid was the juxtaposition of the work being undertaken by the ETSI Open Source MANO (OSM) project and the MEF Lifecycle Service Orchestration (LSO) project; presented on the preliminary day of the conference.
With its Madrid location OSM, and its key sponsor Telefonica, were bound to feature heavily. OSM is now 2 years old, has over 90 members and is approaching its 4th release of code. Throughout the morning there were many presentations on what OSM could achieve and on various proof of concepts that it was being used in. For me, what was missing was any real use of OSM in actual products within Telcos or externally with others. The Proof of Concepts still felt like early stage internal, engineering views of what an NFV product should be, rather than something a customer would buy and use.
“We will require 3 times as many software developers in the future from what we have today.” – Telefonica
I’m sure the intention of this quote, given at the conference, was to highlight how Telefonica was transforming into a software-oriented organisation. However, it also felt like an indication of how software projects, like OSM, may be becoming too large and complex. Rather than a statement of strength it felt like an admission that trying to solve problems like generic orchestration, with no immediate product drivers, had lead to projects that had too wide a scope and limited prioritisation driven by actual external need.
The MEF LSO project , by contrast, seemed to be rapidly moving to a point where it would be successfully used to support inter-carrier ordering of live products between Telcos. Verizon and Colt had already demonstrated two way ordering between them for carrier ethernet products. Importantly this was not on a test network but on the real live network of each carrier and based on real products that those Telcos sell. It was clear that the MEF inter-carrier project would be being used in live products by the end of the year.
The scope of what MEF is attempting is less than OSM. However, by solving a real and immediate problem, with a tighter scope, for known and current Telco products, it appears to be far more likely to succeed and be used. Most importantly by being used it will be exercised and the prioritised needs for additional functionality will be rapidly determined and delivered.
One of the challenges with initiatives like OSM, is that it can be seen as solving problems that Telcos aspire to have, rather than real problems they actually have today. OSM and other MANO initiatives are based on the idea that a Telco wants and needs to rapidly innovate services with new VNFs being constantly onboarded. The reality is that after nearly 10 years of NFV Telcos are still creating products that are largely evolutions of their existing portfolio with VNFs as static replacements of physical network functions.
The power of rapid and live use of software in real products can be seen with Amazon Snowball Edge (a ruggedized edge compute node) product. This has been rapidly released with a developer API and then evolved with application developers and external customers. In the same time that OSM has been being developed, Amazon have deployed and iteratively developed a consumer configurable, API driven, edge compute “CPE like” product.
Appledore will be publishing a research paper on the competitive open source Telco orchestration project ONAP in the near future.