Preetham Uthaiah, EVP – Marketing and Strategy, Saankhya Labs

In 5G networks, the major disruption is “disaggregation” and “virtualisation.” These two concepts are increasingly making the telco networks look like IT cloud networks. All the scaling benefits accrued in an IT cloud network are being brought to the telco cloud simplifying deployment, provisioning, and automation with the application of AI/ML.

The Open RAN (Radio Access Network) and the ORAN initiatives have been advocating open interfaces between the disaggregated components of the RAN. Disaggregation has been achieved by splitting the Layer 1 (L1) signal chain into the Remote Radio Unit (RU) and the Distributed Unit (DU) “horizontally” and is bringing about limited pooling gains.

Saankhya Labs is working on several innovative solutions to bring the true potential of Open RAN. Some of the innovative solutions which Saankhya is bringing to the Open RAN ecosystem include:

  • Vertical Disaggregation – API and a methodology-based framework to bring openness and portability between hardware and software.
  • Hyper Observability – Create more data sources at Layer1 to deploy a more effective RAN Intelligent Controller (RIC), which uses AI/ML on this data to bring more automation and efficiency to the networks.

Vertical Disaggregation: The disaggregation in the DU landscape has propelled the separation of the DU hardware and software. DU High PHY has adopted a software-defined approach executing on COTS hardware. This has resulted in a diverse ecosystem of DU Independent Software Vendors (ISV) vendors. But the Independent hardware vendor (IHV) ecosystem is still very niche and results in vendor lock-in.

The COTS DU hardware methodology is not open, making it very difficult for operators to multi-tenant DU hardware platforms. Interchange of a DU ISV with a DU IHV can be possible only when the L1 DU software is truly “portable” across multiple hardwares.

Saankhya Lab’s RANWiser platform attempts to solve this problem by taking the approach of vertical disaggregation. RANWiser is a standardised software framework for RAN workloads which is integrated into a Telco cloud infrastructure software. The framework provides a uniform model for the hardware underneath and lowers the barrier to build portable RAN software.

The RAN-Wiser consists of a

  1. A “static” methodology part that requires DU Cloud-native Network Function (CNF), which is the L1 software, to be developed using the APIs provided by the RANWiser SDK.
  2. A dynamic runtime system that comprises the Super-Scheduler and the Hardware Architecture Description.

Apart from portability, other benefits of RANWiser include

  • Virtualisation of the RAN resources: Fine grain control allows efficient implementation of RAN slicing functionality
  • RAN Monitor provides hyper Observability, enabling the development of AI/ML xApps for power, performance management, and network tuning
  • Domain level thinking and problem abstraction
  • Security: Additional level of indirection allows security checks of the CNF so that there is no spoofing
  • DU CNF functionality audit for Power, Performance, Security before deployment

Hyper Observability: RAN Intelligent controller (RIC) was included in the Open RAN architecture to provide advanced control functionality, which delivers increased efficiency and better radio resource management.  These control functionalities leverage analytics and data-driven approaches, including advanced AI/ML tools to improve resource management capabilities. One of the key drivers that determines effectiveness of AI/ML techniques, is the amount of data that is available for efficient learning. Another one is the speed with which this data can be made available to the RIC so that it can be applied in near Real-Time use cases. Currently, there are no easy answers to address this Hyper Observability or data acquisition issue for near real-time applications.

Saankhya is bringing new techniques to gather more data by ‘digitising’ the Radio environment. Some of the use cases that are being targeted from this technology include:

  • Smart Scheduling
  • Coverage and Power Optimisation
  • Interference Management
  • Cross-Layer Optimisation in Layer-2/3/4
  • MIMO Beamforming optimisation