How Should Equipment Vendors Approach the Open RAN and vRAN Trend?

In June, Vodafone announced that it will deploy virtualized Radio Access Network (vRAN) in the United Kingdom (UK), utilizing Samsung’s vRAN and open radio access network (RAN) solutions. This will not be a full-scale deployment and will only cover the low and mid-band spectrum. A month before the announcement, Mavenir announced an open RAN deployment for a smart city project in Thailand, and meanwhile, Rakuten further expanded its open RAN partnership to NEC and Fujitsu to develop open RAN compliant radio solutions for Rakuten Communications Platform (RCP).

In response to the increase in commercial open RAN and vRAN activities, ABI Research sees major incumbent Network Equipment Vendors (NEV), such as Nokia, stepping up their development for these technologies. For example, Nokia announced in June 2021 the opening of an open RAN Collaboration and Testing Centre in Dallas to test open fronthaul interfaces and xApps for their new RAN Intelligent Controller (RIC) platform—the Service Enablement Platform (SEP), which was announced in March 2021. Ericsson, while more averse to open RAN compared to Nokia, has also launched Open Lab in March 2021 to test vRAN solutions. Although NEVs’ business model has traditionally been selling full end-to-end RAN networks, the demand by Communication Service Providers (CSPs) for disaggregated and open RAN solutions is too large to ignore.

vRAN refers to the disaggregation between hardware and software of a baseband unit (BBU). A cell site would normally include a remote radio unit (RRU) and a BBU which is responsible for baseband processing of signals to and from the RRU to the core network. The BBU has traditionally been a fixed proprietary hardware with fixed capacity. With vRAN, the BBU’s processes are disaggregated from the hardware into virtualized central units (vCU) and virtualized distributed units (vDU). The vCU and vDU run as virtual network functions (VNFs) on top of commercial off-the-shelf (COTs) hardware, which is less expensive than proprietary hardware. As a result of this disaggregation, the vDU can be scaled up or down to meet changes in traffic demand and introduce new opportunities for automation, such as zero-touch provisioning and artificial intelligence (AI) to optimize network operational efficiency. New services can also be introduced on-demand through VNFs because it is now software. Introducing NFV concepts to the RAN can also break apart the vertical siloes which the RAN is traditionally built upon, fostering collaboration between multiple vendors and increasing innovation.

Open RAN is a different concept and refers to the opening of transport interfaces between the BBU and RRU, and within the BBU itself. These include interfaces such as the fronthaul interface, which connects the BBU to the RRU, and the F1 interface, which connects the CU to the DU. Opening these interfaces allows for CSPs to mix and match vendors to build their RAN network, potentially reducing the total cost of ownership (TCO) due to competitive pricing, and allowing CSPs to take a “best of breed” approach versus traditional single vendor deployments where the vendor deploys a pre-integrated siloed end to end system which might create a single vendor “lock in”. Open RAN is a domain where software-centric players such as Mavenir, Altiostar, and Parallel Wireless are active in. However, while NEVs are generally more averse to open RAN compared to vRAN, Nokia has taken the most steps towards an open RAN environment, having trialed the deployment of an open fronthaul interface with some of their customers, which will utilize Nokia’s baseband and a third-party RRU.

Nokia has taken very bold steps towards developing open vRAN solutions. These include Nokia AirScale Cloud RAN (which splits vDU and vCU), Nokia open Fronthaul Gateway (trialing the opening of Lower Layer (LL) split to integrate third-party radios with Nokia’s baseband), Nokia AirFrame Open Edge Server (for vRAN edge deployments), and Nokia SEP (a near-RT RIC and MEC platform). Nokia was also the first major NEV to join the open RAN Alliance and Open RAN Policy Coalition (Ericsson is not part of the Open RAN Policy Coalition), and is the first major NEV to develop a near-RT RIC platform.

However, while Nokia is developing open RAN, their strategy remains largely to deliver the multi-vendor environment through an end to end pre-integrated system, leveraging on existing Nokia assets such as EdenNet (Self Organizing Network for non-RT), CloudBand Infrastructure Software (CBIS, a virtual infrastructure), Nokia Container Services (NCS, to manage containers), and radios. But with the development of open interfaces, Nokia could partner with CSPs and stakeholders to swap out certain components, limited to certain technical specifications, a move which is not something that any other NEVs is making.

It is prudent to note that open RAN and vRAN are not perfect technologies and have many challenges in implementing them. For instance, the RAN has stringent bandwidth and latency requirements which usually require specialized hardware. As such, even if vRAN enables generic hardware, this hardware should also be specialized to meet certain requirements such as hardware acceleration. Most NEVs take this stance, and while they offer vCU solutions, they often keep their vDUs on proprietary or highly specialized generic hardware to enable use cases such as massive MIMO, beamforming, etc. For open RAN, it is simpler to open the A1 and O1 interfaces, allowing for multi-vendor implementation of the Service Management and Orchestration Framework, which includes the non-RT RIC. These are the interfaces that most NEVs are willing to open. However, the lower layer split for open fronthaul and E2 interfaces is something that is much more difficult to open. For the LL split, Ericsson believes that opening the LL split decreases performance and is opting for maintaining it as a proprietary interface, whereas Nokia is actively testing the conversion of their Fronthaul Gateway’s CPRI radios to eCPRI 7-2X specified by the open RAN alliance. As such, it could very well be that opening the LL is more of a business strategy issue than a technological one.

ABI Research recommends NEVs to be adaptive to open RAN trends and evolve their radios to new technology. This will help them remain competitive and adapt to market pressures for a more open and multi-vendor RAN environment. While it makes sense for NEVs to focus more on vRAN than open RAN, they must also anticipate a future where CSPs might opt for a more multi-vendor environment and adapt their products to be interoperable with these environments. As such, incumbent NEVs should take small steps towards trialing open RAN technology with CSPs by deploying open RAN in smaller footprints with CSPs. This will help CSPs, NEVs, and their ecosystem partners understand challenges early before open RAN enter mass market commercialization. Through this process, NEVs would also be able to better identify CSPs needs and use cases and thus value add to them. For example, certain open interfaces such as the fronthaul split might cause performance drops, which might be better serviced with a proprietary interface, and NEVs can consider only opening up certain interfaces. This way, NEVs will better be able to position themselves and adapt to the future needs of their customers. CSPs will also be able to benefit from an open RAN environment by being able to choose “best of breed” components, increase innovation and reduce the effects of vendor “lock in”.