What is O-RAN SC?

O-RAN Software Community (O-RAN SC) was formed in April 2019 under Linux Foundation, sponsored by the O-RAN Alliance. Focus of O-RAN SC is to deliver open source enabling modular, open, intelligent, efficient, and agile disaggregated Radio Access Network (RAN).

“The O-RAN Software Community is a collaboration between the O-RAN Alliance and Linux Foundation with the mission to support the creation of software for the Radio Access Network (RAN). The RAN is the next challenge for the open source community. The O-RAN SC plans to leverage other LF network projects, while addressing the challenges in performance, scale, and 3GPP alignment.” [1]

RAN Intelligent Controller Applications (RICAPP), Near-Realtime RAN Intelligent Controller (RIC), Non-RealTime RAN Intelligent Controller (NONRTRIC), Operations and Maintenance (OAM), O-RAN Central Unit (OCU), O-RAN Distributed Unit High Layers (ODUHIGH), O-RAN Distributed Unit Low Layers (ODULOW), Infrastructure (INF), Simulations (SIM), Integration and Testing (INT) are key sub-projects that constitute the O-RAN SC.

In August 2021, O-RAN Alliance announced the fourth release of O-RAN SC : “Dawn”. 60+ developers from about 10 companies worked together tirelessly to meet the release deadline for Dawn release with significant contributions to all the sub-projects [2]. Dawn release includes major enhancement to SMO (Service Management and Orchestration), ODUHIGH, SIM, and OAM projects to support “closed-loop automation” usecase. This usecase demonstrates the end-to-end automation of configuration and monitoring for any network events. Also there was a strong focus to harden the platform by complying with the Linux Foundation Core Infrastructure Initiative (CII) Best Practices Badging Program, which has emerged as one of popular and authorized barometers to guarantee the best practice “Open Source” in terms of its code review process, governance, security management, DevOps and so on. Most of the sub-projects completed the CII badging criteria to meet basic “passing” grade in this release. Overall, this release has been a stepping-stone to bring RAN innovation and openness on a secure platform and has also laid the framework for rApps (running on top of non-RT RIC platform) and xApps (running on top of near-RT RIC platform) development. Details for the D release can be found at O-RAN SC site [3].

What is Samsung’s contribution in O-RAN SC?

Samsung has been actively contributing into the O-RAN SC since 2019 with significant contributions to the RIC and RICAPP projects. Samsung contributors also hold a committer position in seven repositories (e2mgr[4], e2[5], a1[6], nodeb-rnib[7], hw-go[8], hw-python[9] and libe2ap [10]). Subhash Kumar Singh from Samsung Research India Bangalore is one of the top contributors (overall 12th position) in O-RAN SC and is holding the committer role for few key repositories related to RIC platform.

Dr. Sunghyun Choi, Corporate SVP and Head of Advanced Communications Research Center, Samsung Research, Samsung Electronics, quoted in the official O-RAN SC press release about the focus and vision of Samsung’s contribution [11].

"Making the RAN intelligent and context-aware is critically important for network optimization and automation. The focus of Samsung's O-RAN SC contribution has been towards enhancing the A1-Enrichment Information (A1-EI) & xApp Framework. Overall, we believe that strengthening the O-RAN AI/ML Framework will play a big role in making operators achieve proactive closed-loop optimization and automation of RAN operations."

The near-RT RIC platform is a very critical component in O-RAN architecture for bringing intelligence to RAN. On top of the near-RT RIC platform, developers can build their own applications xApp to optimize network parameters by utilizing the data and the machine-learning (ML) framework provided by the platform. The accuracy of an ML algorithm is largely dependent on the number of data points that are available. Based on the use case, xApps might need information that is outside the purview of the RAN but has significant impact on the overall network performance, e.g., weather conditions. In this release, we focussed to implement A1-EI interface at near-RT RIC as per the specification provided by O-RAN Alliance. This feature will enable xApps to consume enrichment information from external source, which when used in conjunction with the data from the network should improve the accuracy of the prediction algorithm and in turn promote the usage of predictive closed-loop automation for RAN optimization. Additionally, the python libraries related to the xApp framework were improved to support an ML framework that can be used by xApp developers to apply ML Algorithms to RAN control.

Further in this release, O2 interface implemented as per the WG6 use case “Deploy xApp in near-RT RIC” in O-RAN Orchestration Use Cases v2.0. Now, xApp initiates registration on its startup. We improved the python framework to support the registration on startup. To make the encoding and decoding of E2AP v1.0 messages more simplified by introducing message structure and wrappers in libe2ap repository. As our focus we have also improved KPIMON xApp to implement E2SM KPM 2.0.3, which is recently approved by O-RAN Alliance. New E2SM model will include new KPI parameter and expose those details to RIC to bring new innovation and ideas. We are also focussed to promote O-RAN SC as developer friendly by bringing reference xApps (Hw-go and Hw-python). xApp developers can refer to it in order to get familiarized with the features of xApp framework and start building new apps for RAN-related decision.

Our developers are paving path towards enabling RAN intelligence with consistent contributions to the O-RAN SC RIC & RICAPP projects . Following contributors from Samsung have been very active in the O-RAN SC:

While the O-RAN SC near-RT RIC platform has AI/ML capabilities, the realisation of a complete AI/ML workflow (as defined in the O-RAN Alliance WG2) would require additional components for the model training such as ML designer and ML training host. It also requires ML model serving frameworks to deploy inference service and inference optimization to reduce latency with efficient resource utilization. The Advance Solution Team at Samsung Research in collaboration with Samsung Research India Bangalore is working on building this end-to-end workflow by utilizing various open source projects

What are the next step?

The E Release of O-RAN SC is currently in the planning phase and scheduled for December 2021. In this release, we can expect to see new xApps, usecases, and more mature platform to lead industry towards open, intelligent, secure, and fully interoperable RAN. On the end-to-end AI/ML framework front, we plan to demonstrate the capabilities of the framework with practical usecases that can help operators realise the true potential of predictive closed-loop automation.