Dipesh Modi Communications Network Engineering, Principal, MITRE
Javed Khan Sr. Director, 5G RAN Product Management, Rakuten Symphony
Realizing the promise of 5G necessitates a revolutionary architecture to support it. No single organization can create it alone.
Previous generations of wireless networks focused on human-to-human and human-to-device communications. Now, 5G promises to accelerate life-changing advancements through device-to-device (D2D) communications that will affect global and local economies, and even how human beings behave with and perceive the world around them.
However, the revolutionary impacts of D2D require a revolutionary network – not simply an extension of LTE. Accelerating the architecture underlying this 5G network is critically important, and it can’t be done without radical collaboration.
Unique Characteristics of 5G Architecture
The disruptive characteristics of 5G require a different approach to building its constituent networks. The monolithic architectures of previous generations must be re-imagined to leverage the potential of cloud-based, D2D communications. Fundamental differences in 5G architecture include:
Service-based modeling
Network disaggregation has introduced a service-based architecture that will help deploy new services faster. Separation of the centralized unit (CU), distributed unit (DU), and radio unit (RU) functions of the Radio Access Network, along with open interfaces defined by 3GPP and the O-RAN Alliance, enables multi-vendor RAN deployments.
Additional disaggregation, like separation of the control plane (CP) and user plane (UP), adds further flexibility in deploying different architectures enabled by the freedom to scale and locate the CP and UP independently. When this disaggregation and distribution are coupled with virtualized implementations and orchestration offered by cloud platforms, groundbreaking new use cases are unlocked.
Native Edge Computing
Multi-access edge computing allows intelligent computation needed to process user data at the edge of the network, where it can keep latency low and minimize the volume of data that needs to move to centralized network components.
Orchestration
In the past, for telcos to deploy a service in a new area, you could expect months – maybe years – of planning to determine and acquire the necessary hardware, software, and human resources for installation.
Orchestration has become an integral part of 3GPP specifications, allowing third-party orchestration software to communicate with your infrastructure and fire up container-based functions in the operator’s production network. Now, we’re talking about minutes instead of weeks or months for rapidly deploying new services utilizing mature IT infrastructure deployment models which can set up a network slice in minutes, whereas previously, you were looking at weeks to months of planning before you could even discuss turning up a slice.
Network Slicing
Network slicing, another significant enabler in 5G architecture, allows flexibility to fully support and reserve resources for different services with different quality-of-service (QoS) levels for an end-user. If you have a service that requires ultra-low latency versus massive IOT devices versus mobile broadband, you don’t want them interleaved and using the same resources. Slicing allows for deployment and control of use cases with various qualities of service.
Speed to market
Vitally, the openness and cloud-native infrastructure of 5G architecture mean speed to market. The delivery of software and capabilities and the DevOps environment the ecosystem generates are significant enablers: end-users get these capabilities much faster than they would have in the past, when waiting for a major release was required.
Despite all its advantages, 5G architecture is a major shift for any operator. Getting these features implemented and moving in this rapidly developing direction presents several challenges to realizing a robust and efficient 5G infrastructure.
Unique Challenges of a Revolutionary 5G Architecture
Some of the advantages to this new architecture play into the challenges of building it.
Whereas 4G baseband isn’t marked by open interfaces, node B is completely broken apart in 5G. You have the RU, DU, and CU, and even the CU is broken up into the control plane and user plane, resulting in additional open interfaces even within the radio baseband.
Previously thought of as a single box, the core in 5G is now composed of services. The entire ecosystem is far more complex because it has been disaggregated both horizontally and vertically.
Not only are these functions broken up into services and components, but the hardware, the platform, and the applications are all decomposed and enable a multi-vendor ecosystem. We’ve disaggregated them, and now we’re trying to make them all work together.
There are clear hurdles to this openness. It requires a significant integration effort to get the COTS hardware and virtualization in the telco space and to address the real-time nature of the applications. The Open RAN community has resolved some of these issues. They have achieved integration with numerous third-party radio vendors and have accomplished integration with different cloud and server hardware vendors to create blueprints of different combinations for cloud and server vendors. As time progresses, additional work will provide standard APIs, and this maturing ecosystem will become easier to mix and match.
These challenges to a new architecture can be overcome but require evolution. The industry is still trying to solve issues and solutions have not been standardized, so this lack of standardization must also be addressed.
Collaboration to Reach the Future Faster
There is substantial public and national interest in advancing 5G architecture, though it shouldn’t come at the expense of an open free market.
The fastest way to our 5G future is radical collaboration that will deliver on the potential of 5G. This cooperation cannot be limited to discussions around standards but must include pooling the expertise and resources of organizations you might not normally consider allies and letting them build off each other. It means developing opportunities to collectively test use cases in realistic environments.
MITRE Engenuity is opening the doors to this unprecedented collaboration, which includes more than enterprise telco players. The Open Generation consortium incorporates universities and startups — proven forerunners in technologies and applications testing — and provides members unique opportunities to engage with government leaders through a liaison program. Together, Open Generation members are establishing testbeds and learning how to effectively build out a 5G architecture that will support the life-changing concepts we’re dreaming up.
Accelerating Architecture through the Lens of Uncrewed Aerial Systems (UAS)
Of all the compelling use cases for 5G technologies, uncrewed aerial systems (i.e., drones) present an opportunity to spur significant progress, in part because of the significant challenges the use case poses. Addressing standards that support 5G UAS operations advances the business case for several key industry sectors, including telecommunications and service providers.
Open Generation has selected UAS as one of its initial focus areas. By tackling capabilities like massive MIMO and beamforming antennas, network slicing, xApps over RIC, and native cloud edge computing, consortium working groups are laying the foundation for the transformational architecture needed to keep the United States a leader in 5G technologies.
Learn More About Collaborative 5G R&D
Learn more about Open Generation’s work to develop the revolutionary 5G architecture that no organization can accomplish on its own:
READ THE OPEN GENERATION POSITION PAPER
