Special Report: 5G supply chains
We often argue that 5G has, so far, failed to deliver the radical changes in architecture and user experience that will be needed, if it really is to amount to a brand new generation of the mobile business model. Another area where there have been high hopes of radical change is in the value chain, and here too, the signs of near term transformation look poor.
The visions for the 5G era focused on a cloud-native and software-defined architecture that would, in turn, open the door to new suppliers by turning the whole mobile platform, not just the 3GPP radio standards, into an open framework. Recently, the conversation has been mainly about Open RAN, and it is perhaps not surprising that only a limited impact has yet been felt on the supply chain, considering the scale of investment and risk that operators take in their RANs, and the entrenched nature both of architectures and suppliers. Geopolitics may have sparked some disruption, but as our article examines today, this has not been as dramatic as some had envisaged. Huawei and ZTE have many strongholds, while new entrants, unless they have the scale of Samsung, will have a long battle ahead.
This is not just about the RAN however. Long before Open RAN, operators like AT&T were promising that the implementation of open, cloud-based architectures for software-defined transport and packet core would result in a diverse supply chain that would improve network economics and open the door to innovations from start-ups and the IT/cloud sector.
Despite major initiatives like AT&T’s Domain 2.0, however, the dominant market shares in these domains still belong to Cisco and Juniper in transport, Ericsson and Nokia in packet core. Even large companies like HPE complain that, despite investing heavily in a cloud-native 5G core offering, it is almost impossible to sell this into a tier 1 telco for a mainstream 5G Standalone deployment (enterprise cellular networks are another matter, and the main segment where real disruption will take place in the next few years).
Yes, the new architectures and operator initiatives have put pressure on the major vendors to change their approach and at least to open their interfaces to support, theoretically, multivendor systems. Cisco has clearly made bigger steps towards software-defined networking (SDN) and white box routing than it would have done without the threat of operators defecting to disruptive new players – the same goes for Nokia in the RAN. But while the big vendors may be changing their approach, and providing new flexibility and cost-effectiveness, that is not the same as a new, diverse supply chain. For start-ups targeting the big network domains, the best hope is often to be acquired by the incumbents, as tail-f, a star of Domain 2.0, was by Cisco.
The other entrenched power players in the mobile business are, of course, Google and Apple, via their mobile operating systems, which underpin huge ecosystems of applications, content services and devices. There was a long stream of would-be challengers in the 2010s, from big companies such as Microsoft – first with Windows Mobile and then with its Nokia purchase – to start-ups such as Finland’s Jolla, designer of the Sailfish OS. But none made any dent on the two giants, to the extent that they are now under scrutiny by antitrust authorities in the European Union and North America (see below).
Of course, geopolitics plays a role here too, and the US sanctions on Huawei have led to its being exiled from the Google-based Open Handset Project (the official implementation of open source Android). Huawei has developed its own mobile OS, Harmony, but it remains to be seen whether this will have any success outside China (where Google has little presence anyway).
Open RAN remains, for many, the litmus test for whether a new supply chain, and open multivendor architecture, will play any role in the coming generation of 5G Standalone macro RAN and core deployments. Supporters like Deutsche Telekom are working hard to set out a deployable roadmap for itself and the wider community, to lower some of the barriers (see below), but outside the enterprise and small cell networks, this will inevitably take some years to mature (late 2020s by many operators’ reckoning).
And meanwhile, it will be critical that open innovation does not lead to fragmentation, with different factions emerging around varying interpretations of key specifications such as those for the RAN Intelligent Controller or the fronthaul interfaces. At an event hosted by DT last week in Berlin, Nokia presented its RIC platform, but acknowledged that, even though it was built on the O-RAN Alliance open source specs, it would only currently support Nokia’s own xApps. It would be easy to dismiss this as the latest in a long line of examples of large vendors building walls around so-called open platforms to lock in their customers, but most Open RAN participants point to the same issue with early- stage RIC.
Sufficient operator pressure should iron out such problems, but Open RAN raises another important issue – is it really possible, or even desirable, for a macro RAN to be assembled from many pieces, each sourced from a specialist vendor? This is one of the visions, since it would allow an innovative company to succeed with just one element of the solution, such as a particular virtual network function (VNF), rather than having to supply every ingredient, something that requires huge resource.
But as LightReading pointed out in an insightful piece last week, most early specialists are actually racing to supply full RAN solutions, presumably because their place in the value chain would be limited if they stuck to just one element. Mavenir has added radio unit developments to its RAN software offering; NEC has been developing RAN and core software to build a full platform around its radios; Chinese small cell radio maker Baicells is now developing baseband software in order to provide an integrated Open RAN product.
The revenues from a RAN would need to be split between a lot of players, in the fully-fledged Open RAN vision, which would squeeze the potential for most of the vendors, especially if there were to be any chance of delivering the affordable costs that operators are targeting. And if many of the vendors are small and specialized, there is an obvious scenario in which a few large suppliers – established in the RAN or incomers from the data center – provide the underlying platform, or become primary contractors, and so grab back their power over the value chain by a slightly different route.