Does Open RAN Stimulate Innovation?

Proponents of open RAN argue that opening the RAN interfaces will stimulate innovation. They claim that non-interoperable interfaces allow incumbent vendors to lock out new innovative players and stifle the market. With a small number of incumbent vendors owning much of the market share, the pressure to innovate is low.

Incumbents counter by stressing the optimized performance across the RAN functions; the result of investing billions of dollars and years of field experience. They say that open interfaces between tightly coupled functions reduces performance and increases cost because someone has to integrate functions from different sources.

With this background, I have been looking to answer two questions:

1. Does open RAN stimulate innovation?
2. What areas of innovation will open RAN stimulate?

Areas of Innovation

The nature of wireless standards introduce boundary conditions for vendors to comply with. Vendors will support the same features according to the 3GPP standards for the air interface protocol. One vendor may get ahead of others in supporting a certain feature, but overall, vendors will support the same features. This leaves innovation centering on matters of implementation, including:

1. Software, in particular algorithms used in the operation of critical functions. For example, the layer 2 MAC scheduling or the implementation of massive MIMO and beamforming algorithms. This falls into the overall category of firmware that runs the RAN.
2. Hardware, especially that of the radio unit. Some may discount the importance of hardware in a world where software commands most of the value. However, hardware has direct impact on performance and cost. Modern networks support multiple frequency bands, multiple carriers, multiple antennas and wide channels. Designing high-power, multicarrier radios is a specialized skill in the domain of the incumbents. Moreover, economies of scale plays critical role in hardware. For this reason, new entrants stress the need for incumbents to open the radio interface.
3. Implementation of the RAN baseband related to hardware and software disaggregation with network function virtualization. Cloud computing concepts also leads to how products are integrated, tested and deployed in the field to meet various service and application requirements. [See here and here for additional info]

New vendors vs. Incumbents

The incumbent vendors are strong in #1 and #2. Overtime, they worked out corner cases and optimized their algorithms to maximize throughput, coverage and capacity under different interference scenarios, poor coverage and other situations. They invested into building hardware – high-performance radios and SoCs for baseband engines – that would be take new entrants time and high investment to replicate.

This leaves point #3 as a way for new entrants to differentiate and compete – essentially by “changing the game.” Disaggregation and virtualization / containerization of RAN functions is what the new entrants stress more than incumbents. It is also a reason why the definition of open RAN diverges between the incumbents and the new entrants. Everyone agrees that open RAN is about open interfaces. However, new entrants would emphasize virtualized infrastructure and COTS hardware. Incumbents try to balance the use of virtualized functions as they stress optimization of the implementation for performance. This subtle difference leads to divergent definition of open RAN that we encounter in the market between new entrants and incumbents.

We cannot discard the approach of the new entrants seeking flexibility of implementation. Neither can we discard the claims of the incumbent related to performance. Along the continuum of flexibility and performance there is a happy medium for the right mix. This is where I would expect cloud computing technologies will inspire RAN innovations. We see cloud providers like AWS, Google and Microsoft design their own acceleration engines, smart NICs and compute cards to lower service costs and improve performance. In the RAN, vendors approach architecture from two extremes while seeking an optimal architecture that balances among competing requirements. Technologies that enable such an architecture is where the value lies.

Concluding Thoughts

In 2008/2009 timeframe, Huawei introduced its SingleRAN base station that supported multiple technologies. It saved operators substantial costs by solving practical problems rising from multiple RANs (power, space, upgrades and frequency refarming, etc.). This innovative design propelled Huawei to greater market share. I think this cannot be repeated today; more is required.

The nature of innovation in open RAN alters well-used and tested processes for product evaluation, testing, deployment and operation. This where value could be found provided we are realistic of the timeframe and investment needed to change service providers’ processes and culture.

Finally, I and the rest of our team have come across many successes and failures having been involved in different “Open-X” initiatives for around a couple of decades (networking, software and hardware…). Ultimately, the markets will decide on the basis of the best compromise in price, performance, and flexibility with a mix of models running in parallel. I expect the same will happen in Open RAN.

---

Note: I particularly welcome comments and suggestions on how you think open RAN will stimulate innovation. I look forward to hearing from you by posting a comment or contacting me.