Your browser is not supported. Please update it.

16 February 2021

‘Open’ is an ill-defined term in 5G, and it rarely means ‘open source’

For years, organisations that wanted to disrupt the closed market for mobile networks looked wistfully at the upheaval in the enterprise IT industry in the 1990s, when companies like Red Hat turned the ecosystem on its head. Open source platforms enabled a wave of innovation in applications, services and software delivery models.

In the 2020s, there is a credible prospect that open platforms will take a strong position in mobile networks, and operators are keen to see the supply chain opened up to a wider range of suppliers and a broader innovation base. However, there is a significant difference between ‘open’ and ‘open source’, and there are many barriers to the latter becoming a dominant model in telecoms networks.

Of course, open source is hugely influential on the economics of mobile operators. Open source software underpins the Android community in handsets, and many development in computing – such as in machine learning and edge computing – which will affect how MNOs build their networks and deliver their services.

But at the heart of their own networks, the influence of true open source is limited. This is especially true in the RAN because it retains a significant reliance on advanced processors and on other hardware elements such as antennas, and despite the rising profile of open source processors such as RISC-V, the impact of open source on specialized hardware has been limited to date.

Most of the open initiatives focused on the mobile RAN and core are not actually open source, or make limited use of open source processes and technologies (see item below about O-RAN).

The dream that a vendor or operator could download open source, free software to an open source baseband or radio unit is too challenging to the established business models, even for Rakuten’s vision of an app store-like platform for building a RAN. According to Daryl Schoolar, practice leader at research firm Omdia, the total contribution of open source to RAN in the first nine months of 2020 “would be at a maximum $600,000 against a market we estimate to be around $29.5bn”.

Many vendors and component makers will find no incentive to contribute to emerging platforms for free, unless this places them in a position to dominate the resulting commercial products and services. Open source platforms need to drive scale like any others, and it often requires a small number of powerful players to do that. Operators may dream that, in O-RAN, those kingmakers will be from their number, rather than large suppliers, but that merely moves the stranglehold of power from one group to another (or perhaps more probably to the large chip providers).

So we must remember how, in big commercial markets, even where a core platform is open source, this has rarely led to the massively diverse ecosystem and low barriers to entry that is the vision of the idealists. Who was arguably the company that leveraged the open source movement most successfully in the large enterprise? IBM, the company that had dominated the proprietary era too, and which is now looking at a new wave of expansion based on its acquisition of one of the open source pioneers, Red Hat.

In handsets, Android is Linux-based and has a fully open source variant, but all significant commercial success in Android handsets, and therefore applications, has been based on the Open Handset Platform (OHP) implementation, which is tightly controlled by Google, to the extent that users have to support certain Google services. Just ask Huawei about how open that platform really is.

There is a good case that open source will never be ideal in markets that require very high levels of investment in complex platforms, since the return on that investment may be hard to achieve if core technologies are licensed free of charge. It worked for companies like Red Hat because there were layers of value added tools and services to go on top of the open, free foundations, but the same may not be true for a processor or base station developer.

And while one of the key arguments for open source is that it drives massive scale, because so many companies can participate in the ecosystem, it can also lead to fragmentation for the same reason. The industry needs to focus on lowering barriers and supporting common standards rather than being too religious about the licensing processes, and most importantly, it needs to arrive at something that has been elusive in cellular – a way to reward innovation while guarding against destructive rounds of legal actions or closed-door deals around patents, which always favor the giants over the newcomers.