NFV Orchestration - ONAP leads, but its not over 'til the fat lady sings

Network Functions Virtualisation (NFV) is a hugely disruptive technology that is set to change the way our networks are architected, services delivered, and how service providers define themselves. While we are seeing increasing numbers of commercial trials and deployments within the industry it can be argued that open source and the lack of standardisation, particularly in "orchestration" is actually slowing down adoption of this technology.

Management and Orchestration (MANO)

The Management and Orchestration (MANO) of Virtual Network Functions (VNFs) is the key to realising the benefits of NFV. NFV is designed to deliver flexible, programmable, dynamic networks that can be switched on, and off, or scaled up and down in response to planned and un-planned events. If you need more capacity in the network; if you need to turn on additional firewalls; if you want to connect a new office, and provide it with cloud services; or even if you want to move services from one data centre to another, all of this is managed through the orchestration system.

The challenge with Orchestration is that as yet there is no defined industry standard, and this lack of standardisation is actually holding back the industry. For every service provider experimenting and trialling NFV services, there are ten waiting for the industry to standardise.

Open Source Projects

In NFV there are two key open source projects - The Linux Foundation's Open Network Automation Platform (ONAP) and ETSI's Open Source MANO (OSM) project that form the likely pre-standards for the industry. However, the emergence of two strong open source contenders for MANO standardization is actually exacerbating the problem. Normally open source projects like Linux and Open Daylight accelerate adoption of technology and bring the industry together, but not for NFV MANO.

ONAP and OSM have split the industry and this article seeks to examine which is likely to end up as the winner in this battle of standards. The industry has been split in this way before with GSM (Global System for Mobile) and ANSI (American National Standards Institute) splitting the cellular world. It was only with the advent of LTE (Long Term Evolution) that we actually had a truly global standard for mobile telecommunications. Its also important to recognise that the industry is moving very rapidly, and its not always the best technology that wins these contests. Remember Betamax, was a much better technical solution for VCRs than VHS...and I realise some of you may not even remember VCRs.

ONAP

ONAP was formed in Feb 2017 through the merger of two Linux Foundation projects - ECOMP (Enhanced Control, Orchestration, Management & Policy) supported by AT&T and Open-O supported by China Mobile and China Telecom. This gives ONAP a very sound footing in North America and Asia, but with less traction in Europe. ONAP has been joined in Europe by Orange Group, who are currently testing ONAP in Poland, but one of the criticisms has been the low number of influential European operators involved in the project.

OSM

OSM was formed in 2016 and is linked to ETSI. OSM has some strong supporters in Europe, notably Telefonica, BT, Telenor, Portugal Telecom and Telkom Austria (owned by Carlos Slim and part of the America Moviles group). This gives OSM a strong European and Latin American bias. Its North American credentials have been shored up by Sprint and Verizon who possibly see OSM as being a more acceptable architecture than that proposed by their arch-rival AT&T.

The multi-billion dollar question is "which of these open source projects will become the defect standard for MANO?" There are a number of factors to consider.

1. Route to standardisation?

The telecoms industry is driven by standards, and many feel that OSM with strong links to ETSI has the best credentials for becoming a true industry standard.

2. Volume of code

With around 8 million lines of code from ECOMP and 2 million lines of code from Open-O, ONAP dwarfs OSM's 350k lines of code. Even through ONAP is going through a process of merging and rationalising the ECOMP and Open-O projects, the 10 million lines of code means ONAP offers significant functionality that goes beyond pure orchestration. ONAP is the clear winner here.

3. Vendor Participants

Both projects have around 30 vendors involved in the project, but in terms of the large Network Equipment Providers (NEPs) OSM is only supported by ZTE and NEC/Netcracker, while ONAP is supported by Huawei, Ericsson, Nokia and Cisco. ONAP is strongest with the vendor community.

4. Service Provider Participants

Both projects have similar numbers of service providers. OSM with Telefonica, BT, Telenor, Sprint, Verizon and SK Telecom are arguably more diverse, while ONAP boasts the two largest operators in the World AT&T and China Mobile and can claim representation of 35-38% of global subscribers. In terms of participants globally ONAP is the winner, but within Europe arguably OSM is ahead.

At the end of "Round 1"

At this stage it appears that ONAP is the strongest contender for forming a defect standard for NFV MANO. The huge code-base includes some very strong features for service creation and real-time inventory that are market-leading. However, it is noticeable that the ONAP architecture has been adopted on a wholesale basis from AT&T's ECOMP architecture, and many observers feel that this architecture is too complex to appeal to any but the largest service providers.

It may also be that the merging of the ECOMP and Open-O takes longer than expected.

OSM is strong in Europe and with some of the key operators such as Deutsche Telekom and Vodafone still not committed we may see OSM continue to be a contender. It appears unlikely that US operators will adopt an ONAP architecture with AT&T's finger-prints all over it any-time-soon, in fact Verizon have publicly announced it has "no plans" to use ECOMP (now ONAP), so global adoption of ONAP is still not assured.

Win/Win

So, although ONAP looks liken the strongest contender there is no clear winner in the NFV MANO standards debate. The industry continues to be split, and this continues to slow the adoption of Cloud Telco. One way for a win-win in the industry would be to see a merger between ONAP and OSM that would create a single open source project that could move from a defacto standard to a global standard in short-timeframe. If this could happen it would only be good for the industry and consumers, but it would mean some of the largest service providers in the world coming together, collaborating and compromising. I hope it happens, but it may take some time.

The comments and observations in this article are the authors own.