CSPs are embracing Open RAN, and engineers must prepare for this new approach to network infrastructure and operations.
The theory of capitalism is that increased competition within an open market drives innovation. But when it comes to building the infrastructure needed for the creation of public and private cellular networks, communication service providers (CSPs) and engineering organizations have historically had few vendor options for radio access network (RAN) components.
This led to the concept of an Open RAN system. By “opening up” RAN, companies can better virtualize and automate their network operations while saving money and energy in the long run through the transition from legacy, proprietary hardware to software-driven, interoperable network deployments. Most CSPs seem excited and eager to embrace Open RAN to realize the full potential of 5G capabilities. With Open RAN, companies can essentially pick and choose the components and vendors they use in their networks, increasing customization and improving performance for mobile users.
Recently, Open RAN was the subject of many announcements by CSPs at the Mobile World Congress. As CSPs and organizations creating private cloud networks move to embrace Open RAN, there are several things that engineers should know about this shift in network architecture and operations.
So What is Open RAN?
Open RAN is sometimes used as a catch-all term to refer to the evolution of RAN technology. Generally speaking, a RAN is used for wireless telecommunications and uses base stations to connect phones, computers or other devices in a network through a radio link. Usually, these stations consist of (1) antennas that convert electrical signals into radio waves, (2) radio units that turn digital information, such as voice or video, into wireless signals and (3) controllers that process these signals and connect the base station with a CSP’s core network. The decomposition of RAN functions breaks this down into the Radio Unit (RU), Distributed Unit (DU) and Centralized Unit (CU).
In recent years, many efforts have focused on developments to standardize telecommunications globally and create more stable networks. In 2018, AT&T, Orange, China Mobile, Deutsche Telekom and NTT DOCOMO founded the O-RAN alliance to develop open, interoperable, virtualized mobile networks. The model includes an open RU and DU interface that can support a multi-vendor RAN. Overall, the movement helps CSPs choose RU, DU and CU components from different vendors, a flexibility that will be critical for the increased complexity of RANs required for 5G.
At the end of the day, the benefits of Open RAN come down to increased performance and flexibility within mobile networks. With Open RAN, CSPs will be able to offer extra services for certain applications. For example, guaranteed reliability, enhanced privacy or speedy responses might be useful for enhanced network capabilities at a hospital or autonomous vehicle manufacturing facility. Plus, private networks can use Open RAN to decrease latency and improve performance on smaller scales to support AI-driven data analytics and real-time decision-making.
What Engineers Should Know About Open RAN
With this added knowledge of RAN technology, here are six things that engineers should keep in mind about the technology.
- Open RAN Is not Just for CSPs
- vRAN and Cloud RAN Need Open RAN
- Disaggregation Means Improved Network Resiliency
- Accelerate and Automate Upgrades
- The Benefits of Intelligent Controllers
- Integration Won’t Necessarily be Simple
Open RAN will also be important for private LTE and 5G networks. For example, manufacturers looking for a network to enable IoT and real-time data analytics will need more flexibility than a wired network and lower latency than Wi-Fi. Currently, Wi-Fi can’t handle the type of flexibility required for 5G connectivity in a warehouse or production-type setting. Open RAN is the best option to reduce latency and provide optimized edge computing, especially in a manufacturing setting. This is likely where Open RAN will emerge as a leading architecture first as companies use its flexibility for private networks before its full-scale adoption by CSPs.
Virtualized RAN (vRAN) involves the disaggregation of RAN software and hardware to allow some of the RAN elements to run as software. Cloud RAN is simply vRAN built using cloud-native solutions, including running the DU and CU workloads in the cloud. However, vRAN and cloud RAN don’t necessarily comply with O-RAN standards and, therefore, are not necessarily capable of incorporating multiple vendors. Plus, open RAN doesn’t necessarily need to be virtualized or cloud-native. This is incredibly important as investing in vRAN or cloud RAN that is not open may result in a network that is not flexible enough to handle the future of 5G network evolution. With multi-vendor options, companies can optimize solutions for every aspect of the network, leaving plenty of space to grow as consumer demands shift and networks evolve.
A relatively straightforward benefit of RAN disaggregation is the ability to use multiple vendors across global networks. By diversifying and virtualizing RAN components, companies can develop more resilient networks. For cloud engineers, this also means that different vendors and cloud-native solutions can be developed or deployed to optimize each element of a RAN. Plus, any upgrades for management and orchestration don’t necessarily require hardware changes and can be handled by cloud-native solutions in a vendor-agnostic manner.
Whereas legacy systems often require hardware updates and maintenance for upgrades, Open RAN will be able to facilitate automatic updates to the network. With open cloud RAN, cloud engineers can automate upgrades like with any other cloud-native application. Other cloud benefits, such as canary deployment, can be taken advantage of to allow a certain percentage of users to test a new service or solution before it’s deployed throughout the entire network. With the ability to integrate AIOps, cloud engineers can optimize and automate provisioning, updates and other operational tasks, something that was not previously possible for network engineers.
One of the key features of Open RAN is the RAN intelligent controller (RIC), which provides an open platform for AI-driven applications that control and optimize network functions. Cloud engineers can take advantage of the RIC to enable cloud-native solutions to drive real-time decision-making within a network and modify network operations to meet local traffic.
Despite all the benefits of Open RAN, for engineers, integration of many different software and even hardware components from any number of vendors will not necessarily be straightforward at the beginning. With increased flexibility comes increased complexity, and cloud engineers will need to integrate cloud-native solutions, at scale, for network operations and maintenance. Virtualized RAN is also still somewhat in its infancy, so integration of hardware and cloud software solutions remains complex and time-consuming as the industry shifts to this new normal. Thankfully, as vendors adopt the O-RAN standard, integration and crosstalk should become easier and more efficient over the next few years.