Your browser is not supported. Please update it.

7 December 2020

Open RAN is looking more integrated as it rises to 5G challenges

Nokia is moving away from ‘end-to-end’ as a mantra, claiming operators do not want to buy all their wired and wireline, access and transport equipment and software from one provider. In reality, the change of direction is likely to be designed to make Nokia easier to break up and parcel up for potential acquirers, from industry or private equity, that might find the whole firm hard to digest.

Elsewhere, challengers to Nokia’s RAN and core crown are becoming more likely to stress ‘end-to-end’, albeit in a narrower context. In the early days of open RAN, supporters talked about the benefits of being able to assemble a network from many different suppliers with endless flexibility. RAN virtual network functions (VNFs) from one or more vendors, core VNFs from others, all deployed on cloud infrastructure from yet another, and with radio/antennas from multiple further partners. Rakuten showed that this could be done in a commercial network, and is even offering to package it all up for other operators in its Rakuten Cloud Platform (RCP).

However, many operators are getting less excited about deploying a network from large numbers of vendors, because of the difficulties of implementing that to the same standard of performance as an integrated solution. A pre-integrated solution like RCP (and others are sure to emerge from Microsoft and others) would remove many headaches but could tie operators into certain vendors or architectural assumptions. A product-neutral system integrator is another route, but adds cost and lacks deep knowledge of the individual radio technologies – acceptable for a small cell or enterprise network, possibly not for a national 5G macro RAN.

That can leave operators doing the heavy lifting themselves, which some lack the time, funds or skills to do; or reverting to a more integrated network, even if open interfaces allow them to procure this from new suppliers; to mix and match two or three vendors perhaps; and to future-proof their deployment, so that new suppliers can be swapped in one day, more easily than is possible with current RANs.

Among the open RAN challengers, some already offer a full stack from radios to VNFs and management tools – Parallel Wireless is an example, and it also boasts support for legacy 2G/3G networks, which larger entrants such as Samsung lack, and which is important to some operators, especially in Europe.

Mavenir, which could already boast VNFs for both RAN and core, as well as its long-established IP voice and messaging platforms, positioned itself purely as a software provider which could work with any radio vendor. It is providing RAN software to Dish’s forthcoming 5G network, for instance, and this will use radios from multiple sources, Fujitsu being the first one announced.

But Mavenir is building a more complete solution too. It recently acquired UK-based small cell pioneer IP.access, which brought it that valuable 2G/3G support, as well as small cell hardware. Now it plans to develop two outdoor 5G radio units, one for sub-6 GHz spectrum and one for millimeter wave frequencies. Next year, according to a post by GlobalData analyst Ed Gubbins, a Massive MIMO radio is also in the works.

Mavenir “wants to be known as something of an end-to-end supplier,” Gubbins wrote in his post. “In some deployments, Mavenir has responsibility for end-to-end network performance, which gives it extra incentive to tailor its own RUs, rather than a partner’s, to ensure performance requirements are met.” He added that the whole open RAN idea will be sold on the basis of the performance of early roll-outs, so this gives Mavenir “even more reason to take its fate into its own hands rather than rely too much on partners”.

“We are … producing various radio products that will break the incumbent’s monopoly in the global market,” a Mavenir spokesperson told Light Reading.

Of course, such moves strengthen the argument, made by many operators, that while it is important to break open the RAN supply chain to more providers, the multivendor RAN remains very challenging in the macro network, if there are not to be significant performance trade-offs compared to integrated architectures. End-to-end, top-to-bottom, integrated networks may not be a thing of the past after all, which does not mean open RAN will not fly, but does introduce some necessary compromises into the fully disaggregated multivendor vision.