Your browser is not supported. Please update it.

9 March 2021

Is Facebook a support or a distraction for MNOs’ open 5G efforts?

The Telecom Infra Project (TIP) is having an increasingly heavy influence on the evolution of open cellular networks. It provides the pragmatic face of an open movement, especially on the RAN side, that can often seem over-simplistic and sometimes politicized in its goals.

While bodies such as the O-RAN Alliance and Open Networking Foundation work on specs; operators run trials and pressurize incumbent vendors; and governments seek to influence the 5G supply chain, TIP is mainly concerned with making deployments more practical. It focuses on operator requirements, on design and roll-out blueprints, and on filling gaps in the platform where others have not devoted resources.

However, TIP does have something of a split personality. Founded in early 2016 by Facebook, as a telco version of the Open Compute Project, the aim was extend the scalable cloud-based economics of OCP to the telecoms networks.

This was mainly about bringing affordable connectivity – and Facebook services – to ‘the next billion’ by encouraging open, simple reference designs that a wide ecosystem of vendors could embrace to enable low cost equipment. Companies that have been active for years in bringing affordable connectivity to unserved populations, such as Intel, piled in.

But the scope has broadened considerably since then. TIP, and platforms it supports such as O-RAN, have been seized upon by large operators too. Supporters like Vodafone and Telefónica quickly realized that a significantly funded initiative to develop a fully open, disaggregated network on cloud infrastructure would benefit them too. Early trials may often have been in rural areas, and there has been plenty of talk about bridging digital divides, but the central agenda has been to drive new competition and innovation into the mainstream RAN supply chain. Operators want new choices of suppliers, and bigger sticks with which to beat the existing ones, and they want to be able to direct specifications to meet their own requirements in full.

While an open, disaggregated network lies at the base of both sets of objectives, the implications for equipment, ecosystem partners and deployments are very different. The Facebook-driven initiatives, like the Terragraph platform, last year’s Supercell launch, and a new alliance with Marvell all remain focused on super-affordable cellular networks for rural and underserved communities.

Such developments may help large MNOs to reach those remote customer bases more affordably than before, and encourage them – often for the first time – to target rural users. But rural locations will never be significant in the classic cellular operator business case. These operators have very different goals for their TIP and O-RAN involvements, and are mainly focused on accelerating progress and innovation in architectures that will support urban densification (a key element of 5G enterprise-centric strategies), and eventually address sophisticated 5G macrocells.

These targets require very different approaches from vendors, and developing 5G macro-capable systems will need different funding levels, partnerships and go-to-market strategies from those in the simplified rural network market – hence the involvement of big names, such as Nokia and Samsung, in O-RAN.

The two sides of TIP are not in conflict, and both reflect important industry goals, but some work may be needed on reconciling priorities. This may become more challenging as the group expands its remit. In the past few weeks alone, it has announced two major new initiatives, which highlight the contrast between TIP’s easy wins and its more complicated activities.

On the one hand, it has announced a private networks project (see below), which could address some clear gaps in the route to mass deployment of enterprise cellular. This is an immature market in which a unifying force with a dose of realism will be very welcome.

On the other hand, its new WiFi project group will entangle it in a whole new ecosystem and set of industry groupings. It is sure to do interesting and valuable work, and may help to form a much-needed bridge to the cellular community at operator deployment level, but it will also inevitably consume energy as TIP will be engaging with groups, commercial interests and political agendas that have been brewing for almost two decades. The requirement gaps will be harder to define, and there may be some interesting dilemmas – in the debate over licensed 6 GHz spectrum usage, for instance.

Meanwhile, the Facebook-centric side of the TIP agenda has perhaps been overshadowed, in the past year, by the intensifying activities of the large MNOs and the high profile of open network initiatives in the context of restrictions on Huawei in Europe and elsewhere. But activity here has been bubbling away regardless, and TIP now has a significant portfolio of designs that could add up to an end-to-end cellular network that could deliver good quality connectivity to remote communities in an affordable way for smaller vendors and service providers. It won’t support a tier 1 5G macro network, but platforms like OpenCellular and Supercell could be very important in addressing the underserved.

Any affordable, open platform, whether for high end Massive MIMO 5G or for rural 4G small cells, can only flourish if it has unified silicon foundations. Small Cell Forum has recognized this by focusing many efforts on its FAPI and 5G FAPI chip-level standard interfaces, which are adopted by large providers like Qualcomm and by start-ups.

Sometimes, the universal silicon foundation has been built by one supplier, as Intel showed when its architecture turned a fragmented microcomputer industry into a mass-scale PC platform on which hundreds of vendors could innovate and compete. But that, of course, put a huge amount of power into the hands of one company.

In open 5G, operators and vendors will be hoping for the scale economics of a unified silicon architecture, but one that can support more than one or two companies. Intel and Qualcomm are both active in TIP and also in developing their own approaches to 5G RAN semiconductors, but this is not a two-horse race, as Marvell’s latest open network announcements show. Marvell is spanning both sides of the TIP agenda, with new partnerships with Facebook and with Fujitsu (the latter targeting tier 1 5G roll-outs).

This kind of strategy by the major chip vendors, and by innovative start-ups, are the best way to square the circle between high performance and ubiquitous coverage, and overcome conflicting agendas to arrive at an ecosystem that could support both.