top of page

O-RAN: Compute Automation Is Critical For 5G Success

Updated: Jan 9


Open Radio Access Network - edge compute automation for 5G networks is critical for success.

Historically, the deployment of RAN (Radio Access Network) has been heavily tied to individual vendors while still adhering to the guidelines set by organizations like 3GPP, ITU, IEEE, and others. RAN, being complex due to its necessary adherence to the laws of physics, presents significant challenges in meeting performance KPIs. This complexity has resulted in vendors opting for proprietary implementations involving hardware, software, and interfaces.  

The BBU (Baseband unit), a key component in RAN, used to be a black box unit whose internal implementation varies from vendor to vendor. The BBU is connected to a proprietary RRU (Remote Radio Unit) through vendor-specific implementation of CPRI (Common Public Radio Interface) protocol.  


Common Public Radio Interface based RAN deployment diagram.

Figure 1: CPRI based RAN deployment. 


The present CPRI implementation restricts its ability to fulfill use cases such as massive-MIMO (Multiple Input Multiple Output) due to several limitations: it lacks support for a virtualized BBU architecture because of its fixed RRU-BBU mapping, necessitates continuous message transfer affecting the implementation of Ethernet-based packet technologies, scales according to system bandwidth and antennas rather than data rate, and has strict latency requirements. 


RAN has a big role to play in enabling the three main use cases that 5G promises:   

  • Enhanced Mobile Broadband (eMBB), with peak data speed of 10/20 Gbps (UL/DL)  

  • Ultra-Reliable Low Latency Communications (URLLC), which specifies less than one millisecond of air-interface latency.  

  • Massive Machine Type Communications (mMTC), which must support more than one million devices per square kilometer.  

Every 5G site is intended to cater to a particular use case, whether it's delivering fixed wireless access (eMBB/high BW), ensuring URLLC (traffic priority) for industrial applications, or accommodating a multitude of devices. Hence, it becomes necessary to disassemble the BBU into different functional splits to fulfill site-specific and use-case-specific demands effectively. 


The O-RAN initiative represents a step in the correct direction towards an open ecosystem. It aims to standardize interfaces, thereby facilitating the disaggregation of hardware and software. This approach offers Telcos greater flexibility in deployment and fosters innovation across the entire ecosystem.  


O-RAN alliance is trying to achieve the real potential of 5G through various working groups, including one on the Fronthaul interface between RU & DU (Distributed Unit) as shown below: 


eCPRI based RAN deployment diagram.

Figure 2: eCPRI based RAN deployment. 


The O-RAN Alliance is dedicated to establishing an open, software-centric intelligent RAN. To support this vision, it has introduced newer interfaces. O-RAN has opted for a specific split point called Option-7-2x (with slight room for variation), which facilitates variable data rates and latencies.

  

Telecommunication companies (Telcos) can strategize for a flexible transport bandwidth since 7-2x can scale according to "streams," enabling the utilization of a greater number of antennas. This approach allows them to consolidate their current fiber resources across multiple sites.  


The open interface simplifies the design by employing fewer user-specific parameters within the 7-2x split, leading to enhanced coordination gains due to reduced lower layer splits. 

Open RAN functional splits diagram.

Figure 3: O-RAN functional splits 


Telecom companies have the flexibility to select the appropriate split depending on the bandwidth and latency requirements between the Distributed Unit (DU) and Remote Unit (RU). This architectural approach promotes BBU pooling, enabling multiple RUs to be served by a single DU in a ring-based packet fronthaul structure. This facilitates a cloud-like deployment, ultimately enhancing cost efficiency by minimizing power consumption. 


Now, operators possess the flexibility to configure the 5G network according to user needs and site-specific limitations (such as power, space, connectivity, coverage, and capacity). This is achievable by implementing the suitable functional split—whether they opt for O-RAN's Split Option 7–2, SCF's Option 6, or 3GPP's Option 2—and combining software and hardware components from top vendors, allowing a mix-and-match approach to create the best setup.  


The below figure shows an implementation for realizing URLL/mMTC use case demanding strict latency & jitter requirements. 

Diagram of RAN implementation for URLCC/mMTC use cases.

Figure 4: RAN implementation for URLCC/mMTC use cases. 


Other use cases like eMBB/Macro sites might require a distinct implementation, involving the deployment of DU+RU/CU+DU combinations on a single hardware platform. 


Open RAN facilitates running radio functions on Commercial Off-The-Shelf (COTS) hardware alongside cloud-based software within a multi-vendor setup. It incorporates programmable, open, and interoperable interfaces that enable intelligent radio control using AI and big data. This setup allows the deployment of edge computing resources nearer to the radio components (RU/DU/CU), thereby reducing latency for customers and cutting down long-haul transport expenses. Interference management will be accomplished through RIC. 


Certain initial radio performance challenges are expected and need to be collectively resolved. Efforts are also directed towards other areas such as brownfield deployment, encompassing coexistence with multiple Radio Access Technologies (RATs), various frequency bands, advanced MIMO configurations featuring multiple carriers per sector, Dynamic Spectrum Sharing (DSS), 5G NR Carrier aggregation, synchronization between Remote Radio Heads (RRHs) & Baseband Units (BBUs), antenna weight considerations, among others. 


To overcome above listed challenges, a close collaboration is needed across vendors, operators and telecom standard bodies. 

Partnering with MetalSoft 

Given the site-specific limitations, O-RAN deployments must possess flexibility, necessitating automation and zero-touch provisioning capabilities right from Day-0.  


Visual of the 3 main architectures for O-RAN 5G networks.

The MetalSoft platform enables the following functions to achieve a customizable, adaptable, cloud-based operational model for deploying O-RAN sites.  


Visual diagram of MetalSoft Control plane and Agent architecture, relative to 5G site deployments.

  • Full lifecycle management of infrastructure 

  • Zero-touch provisioning of complete stack including vDU & vCU software. 

  • Hardware discovery and capability exposure: FPGA (wireless protocol offload (L1-Hi)), Hardware Network sync frequency/time methods including O-RAN LLS-C1/C2/C3, PRTC, CPRI/eCPRI, SyncE, PTP, 802.1CM. 

  • Intelligent RAN workload placement 

  • Remote Monitoring & Orchestration 

  • Low footprint & hierarchical topology 

  • API driven and Self-serve interfaces. 

  • Multi-vendor and multi-site deployments 

  • Network Topology Awareness 

  • Security Awareness: hardware root of trust with strong security checks embedded. 

The success of O-RAN hinges on the cooperation of all stakeholders, contributing to interoperability, integration, testing, and various aspects of radio optimization, encompassing interference, intermodulation, beamforming, beam steering, and band utilization. 


Automation, particularly in compute deployment, stands as a crucial facilitator, forming the bedrock for the effective implementation of O-RAN. MetalSoft is dedicated to establishing this foundation for O-RAN through comprehensive automation, empowering telecommunication companies to swiftly deploy networks with minimal on-site interventions.


Our aim is to assist telcos in embracing O-RAN technology extensively, catering to diverse deployment scenarios, including high-density urban areas, to drive its widespread adoption.



bottom of page