ARC-OTA 1.5
v1.5

Product Brief

  • A wideband, real-time platform to replace existing narrow-band, non-real-time systems

  • A full-featured platform for NG wireless evolution

  • C/C++ programmable from the physical layer through to the Core Node (CN)

  • Quick network onboarding and algorithm development in real-time networks

  • Accelerated AI experimentation in wireless RAN workloads

  • A pipeline for data collection, storage, and parsing using 3GPP schema for wireless communication.

The configuration and capabilities of ARC-OTA 1.5 are outlined in the following sections.

Number Antennas

4T4R

Number of Component Carriers 1x 100MHz carrier
Subcarrier Spacing (PDxCH; PUxCH, SSB) 30kHz
FFT Size 4096
MIMO layers DL: 4 layers; UL: 1 layers
Duplex Mode Release 15 SA TDD
Number of RRC connected UEs 16
Number of UEs/TTI 2
Frame structure and slot format DDDDDDSUUU
DDDSU
User plane latency (RRC connected mode) < 10ms one way for DL and UL
Synchronization and Timing IEEE 1588v2 PTP; SyncE; LLS-C3
Frequency Band n78
Max Transmit Power 22dBM at RF connector
Peak throughput SMC-GH DL: ~1.03Gbps; UL: ~125Mbps
Dell R750 / Gigabyte DL: ~800Mbps; UL: ~110Mbps
Bi-directional UDP Traffic > 10+ hours exercised (SMC-GH)
> 4.0 hours exercised (Dell R750 + A100X)
> 4.0 hours exercised (Gigabyte + A100 + CX6-DX)
Note

OTA test was performed with the following configuration: Samsung S22 + Gigabyte + DDDDDDSUUU.

Tip

To learn how KPIs have changed from last release, refer to the Release Notes.

Feature

Description

Full stack software A 3GPP Release 15 compliant and O-RAN 7.2 split 5G SA 4T4R wireless stack, with all network elements from Radio Access Network and 5G Core. Aerial CUDA-Accelerated RAN Layer 1 is integrated with Open Air Alliance (OAI) (https://openairinterface.org/) Distributed Unit (DU), Centralized Unit(CU), or a 5G NR gNB and 5G Core Node(CN) network elements.
Radio network hardware components The COTS hardware Bill of Materials used for NVIDIA qualification is available in the OTA-qualified HW BOM manifest
Source code access Complete access to source code in C/C++, from Layer 1 through 5GC to jump start customizations and next-generation algorithm research. Review the Licensing section for more details.
Developer extensions and plugins To accelerate innovation, developer extensions and contributions are welcome. For example, The Kubernetes Service Management optional developer extension from Sterling provides two capabilities: Kubernetes Service Orchestration and Service Monitoring.
AI frameworks AI frameworks and tools are integrated to ease the AI/ML developer journey for advanced wireless research. For example, pyAerial and Aerial DataLakes has been integrated with ARC-OTA.

5G NR gNB Features

Component

Capabilities

gNB PHY Aerial CUDA-Accelerated RAN Layer 1 PHY (cuPHY) adheres to 3GPP Release 15 standard specifications to deliver the following capabilities. PHY capabilities include the following:
  • Error detection on the transport channel and indication to higher layers
  • FEC encoding/decoding of the transport channel
  • Hybrid ARQ soft combining
  • Rate matching of the coded transport channel to physical channels
  • Mapping of the coded transport channel onto physical channels
  • Power weighting of physical channels
  • Modulation and demodulation of physical channels including
  • Frequency and time synchronization
  • Radio characteristics measurements and indication to higher layers
  • Multiple Input Multiple Output (MIMO) antenna processing
  • Transmit Diversity (TX diversity)
  • Digital and Analog Beamforming
  • RF processing

3GPP standards specifications that define the Layer 1 compliance are:

  • TS 38.211 (38.211 v15.8.0) numerologies, physical resources, modulation, sequence, signal generation
  • TS 38.212 (38.212 v15.8.0) Multiplexing and channel coding
  • TS 38.213 (38.213v15.8.0) Physical layer procedures for control
  • TS 38.214 (38.214v15.8.0) Physical layer procedures for data
  • TS 38.215 (38.215v15.8.0) Physical layer measurements
  • TS 38.104 (base station radio Tx and Rx) Base Station (BS) radio transmission and reception

Aerial CUDA-Accelerated RAN complies with ORAN FH CUS specification version 3 (version 4 for power scaling) Aerial CUDA-Accelerated RAN complies with northbound interfaces adopted by industry based on Small Cells Forum for Layer 1 and Layer 2 (SCF FAPI).

gNB MAC
  • MAC -> PHY configuration using NR FAPI P5 interface
  • MAC <-> PHY data interface using FAPI P7 interface for BCH PDU, DCI PDU, PDSCH PDU
  • Scheduler procedures for SIB1
  • Scheduler procedures for RA
    • Contention Free RA procedure
    • Contention Based RA procedure
      • Msg3 can transfer uplink CCCH, DTCH or DCCH messages
      • CBRA can be performed using MAC CE or C-RNTI
  • Scheduler procedures for CSI-RS
  • MAC downlink scheduler
    • phy-test scheduler (fixed allocation and usable also without UE)
    • regular scheduler with dynamic allocation
    • MCS adaptation from HARQ BLER
  • MAC header generation (including timing advance)
  • ACK / NACK handling and HARQ procedures for downlink
  • MAC uplink scheduler
    • phy-test scheduler (fixed allocation)
    • regular scheduler with dynamic allocation
    • HARQ procedures for uplink
  • Scheduler procedures for SRS reception
    • Periodic SRS reception
    • Channel rank computation up to 2x2 scenario
    • TPMI computation based on SRS up 4 antenna ports and 2 layers
  • MAC procedures to handle CSI measurement report
    • evaluation of RSRP report
    • evaluation of CQI report
  • MAC scheduling of SR reception
  • Bandwidth part (BWP) operation
    • Handle multiple dedicated BWPs
    • BWP switching through RRCReconfiguration method
gNB RLC
  • Segmentation and reassembly procedures
  • RLC Acknowledged mode supporting PDU retransmissions
  • RLC Unacknowledged mode
  • DRBs and SRBs establishment/handling and association with RLC entities
  • Timers implementation
  • Interfaces with PDCP, MAC
  • Interfaces with gtp-u (data Tx/Rx over F1-U at the DU)
  • Send/Receive operations according to 38.322 Rel.16
gNB PDCP
  • Integrity protection and ciphering procedures
  • Sequence number management, SDU discard and in-order delivery
  • Radio bearer establishment/handling and association with PDCP entities
  • Interfaces with RRC, RLC
  • Interfaces with gtp-u (data Tx/Rx over N3 and F1-U interfaces)
  • Send/Receive operations according to 38.323 Rel.16
gNB SDAP
  • Establishment/Handling of SDAP entities.
  • Transfer of User Plane Data
  • Mapping between a QoS flow and a DRB for both DL and UL
  • Marking QoS flow ID in both DL and UL packets
  • Reflective QoS flow to DRB mapping for UL SDAP data PDUs
  • Send/Receive operations according to 37.324 Rel.15
gNB X2AP
  • Integration of X2AP messages and procedures for the exchanges with the eNB over X2 interface according to 36.423 Rel. 15
gNB NGAP
  • Integration of NGAP messages and procedures for the exchanges with the AMF over N2 interface according to 38.413 Rel. 15
    • NGAP Setup request/response
    • NGAP Initial UE message
    • NGAP Initial context setup request/response
    • NGAP Downlink/Uplink NAS transfer
    • NGAP UE context release request/complete
    • NGAP UE radio capability info indication
    • NGAP PDU session resource setup request/response
  • Interface with RRC
gNB F1AP
  • Integration of F1AP messages and procedures for the control plane exchanges between the CU and DU entities according to 38.473 Rel. 16
    • F1 Setup request/response
    • F1 DL/UL RRC message transfer
    • F1 Initial UL RRC message transfer
    • F1 UE Context setup request/response
    • F1 gNB CU configuration update
  • Interface with RRC
  • Interface with gtp-u (tunnel creation/handling for F1-U interface)
gNB GTP-U
  • New gtp-u implementation supporting both N3 and F1-U interfaces according to 29.281 Rel.15
    • Interfaces with RRC, F1AP for tunnel creation
    • Interfaces with PDCP and RLC for data send/receive at the CU and DU respectively (F1-U interface)
    • Interface with SDAP for data send/receive, capture of GTP-U Optional Header, GTP-U Extension Header and PDU Session Container.

5G Core Features

AMF Features NGAP AMF status indication (3GPP TS 38.413)
Add UE Retention Information support (3GPP TS 38.413)
Support of Location services with LMF and AMF (3GPP TS 29.518, 3GPP TS 38.413, 3GPP TS 23.502)
Update NAS with Rel 16.14.0 IEs: Refactor code for Encode/Decode functions; cleanup NAS library (3GPP TS 24.501)
Fixes Fix typo for N1N2MessageSubscribe (3GPP TS 29.518)
Fix issue when receiving PDU session reject from SMF (3GPP TS 29.518, 3GPP TS 23.502)
Technical Debt Reformatting of the SCTP code
Refactor promise handling
Removing dependencies to libconfig++ (Only YAML file can be read as configuration)
SMF Features Add N1/N2 info in the message response to AMF if available (3GPP TS 29.502)
Fixes Add connection handling mechanism between NRF and SMF
Technical Debt Refactor SMF PFCP associations to use UPF profile
UDM Fixes Add connection handling mechanism between NRF and UDM
UDR Technical Debt Fixed builds
Add connection handling mechanism between NRF and UDR
Improve MongoDB support
Common New HTTP Client library (CPR) for all the NFs
Support mobility registration update procedure (3GPP TS 23.502)

5G Fronthaul Features

RU Category

Category A

FH Split Compliance 7.2x with DL low-PHY to include Precoding, Digital BF, iFFT+CP and UL low-PHY to include FFT-CP, Digital BF
FH Ethernet Link 25Gbps x 1 lane
Transport encapsulation Ethernet
Transport header eCPRI
C Plane Conformant to O-RAN-WG4.CUS.0-v02.00 7.2x split
U Plane Conformant to O-RAN-WG4.CUS.0-v02.00 7.2x split
S Plane Conformant to O-RAN-WG4.CUS.0-v02.00 7.2x split
M Plane Conformant to O-RAN-WG4.CUS.0-v02.00 7.2x split
RU Beamforming Type Code book based

To ease developer onboarding, this section provides reference blueprints with key ingredients that were combined to create a tested product prototype: A full-stack innovation sandbox to accelerate innovation in wireless networks and provide new insights on experiments and research. Many assumptions made for analytical and simulation studies may likely improve benchmarks or prove to be invalid in a real network. Our experience in innovation labs as we design, setup, deploy, and add tools and frameworks is available to all developers. We have deliberated on the hardware components, software configurations, and deployment strategies. We have also run into difficulties and pitfalls, and want to ensure others that the hardware components and software configurations have undergone a rigorous qualification process. Each developer delta for the lab experimental networks is expected to be limited to the environment variability, the transmission power, attenuation, and limited set of variables.

As we look to leverage, extend, and innovate, here are the key guiding attributes for the blueprints:

Attribute

Description

Prototype Serve as a reference implementation
Re-Use is important to avoid having to “re-invent” the frameworks, tools, platform
Automation to ease setup and minimize deployment pitfalls for software environments and configurations
Uniformity Access to exact set of combined ingredients or recipe used in the reference
Configuration customization is minimal and needed when developer wants to change behavior
Availability Service health monitoring of the network necessary
Extendibility Easy to extend based on O-RAN modular, flexible, open architecture
Performant OTA Network Develop innovation prototypes and validate benchmarks versus expected on-paper or simulation
Extensions and Plugins Developer blueprints for others to leverage as building blocks

Blueprint Table

Blueprint

Description

Downloadable Version

ARC-OTA Disaggregated Mobile Network As a Research Sandbox PDF
Full Stack Programmable Launchpad for Advanced Wireless Developers – Gateway to Developer Innovations and Extensions PDF
Multi-Vendor Integration NVIDIA Qualified COTS Multi Vendor Hardware and Software Blueprint PDF
Multi-Vendor Disaggregation NVIDIA ARC-OTA developer plugin Multi Vendor Interop with a modular element change blueprint PDF
On-Prem Data Center Deployment NVIDIA Qualified On Prem Data Center Deployment Blueprint PDF
K8 Service Management NVIDIA Qualified Service Management Blueprint PDF
O-RAN 7.2 Split NVIDIA GPU inline accelerate high PHY PDF
CSI Dataset NVIDIA ARC-OTA Multi-UE Channel State Information dataset blueprint PDF
OpenRAN Gym Developer extension integrated with OSC RIC PDF

ARC-OTA

blueprint_arc_ota.png

Component

Feature

Hardware Stack Leverages COTS (Common Off The Shelf) vendors
Software Stack Fully programmable in C/C++
Network uses CUDA Accelerated RAN Layer 1 and OAI software.
Extensible through network services
Integration Developer community is encouraged to extend the stack by contributions across all layers of the stack. Early examples include Sterling SkyWave Service Management and O-RAN OSC RIC.
Deployment NVIDIA SDK Manager offers automation to easily deploy this NVIDIA qualified blueprint.

Full Stack Innovation

blueprint_full_stack_innovation.png

Compontent

Feature

COTS hardware COTS infrastructure composed of compute, virtualization, radios, fronthaul networking, precision timing, accelerators.
Virtualization Virtualized RAN workloads from NVIDIA and Open Air Alliance
AI/ML Frameworks Data Lake + pyAerial for AI/ML frameworks : RF / IQ data + FAPI
Standards 3GPP Release 15+ O-RAN 7.2 split P5G on-prem lab network
Developer Tools Reference OAI CI leverage for developer integration workflow
Developer Extension – Sterling NVIDIA NGC APIs for Sterling K8 service orchestration of network functions
Developer Extension – OpenRAN Gym Developers can deploy AI/ML models for and on the radio access network using OSC radio intelligent controller by following Northeastern OpenRAN Gym tutorial.
Developer Tools – Netowrk as a service NVIDIA SDK Manager development environment setup automation
Note

Developer contributions through extensions and plugins - for community benefit and to accelerate pace of innovations welcome!


Multi-Vendor Integration

blueprint_multi-vendor_integration.png

Leveraging O-RAN, and 3GPP specifications and interfaces enables multi-vendor interop towards the full stack building blocks and developer extensions and plugins. A multi-vendor reference blueprint is extended with the SCF FAPI interface between the O-DU low and O-DU high.

Organization

Features

Northeastern E2 interface plugin leveraging O-RAN OSC RIC and template xApps
OpenAirAlliance O-DU-High (Layer 2), O-CU and 5GC
NVIDIA O-DU Low / Phy High
Foxconn O-RU Phy Low
Others Handsets (Apple iPhone 14, Samsung S23), Viavi Qualsar Grandmaster, Dell FH switch, Supermicro server.

Multi-Vendor Disaggregation

blueprint_multi-vendor_disaggregation.png

Leveraging O-RAN, and 3GPP specifications and interfaces enables multi-vendor interop towards the full stack building blocks and developer extensions and plugins. A multi-vendor reference blueprint is extended with the SCF FAPI interface between the O-DU low and O-DU high.

Organization

Features

Northeastern Developer extension with E2 interface plugin leveraging O-RAN OSC RIC and template xApps
Plugin to leverage https://open5gs.org/ for 5GC instead of NVIDIA reference Open Air Interface 5GC. This highlights the use of modular, open, and interoperable components within disaggregated ORAN architecture
OpenAirAlliance O-DU-High (Layer 2), O-CU and 5GC
NVIDIA O-DU Low / Phy High
Foxconn O-RU Phy Low
Other Handsets (Apple iPhone 14, Samsung S23)
Viavi Qualsar Grandmaster
Dell FH switch
Supermicro server

On-Prem Data Center Deployment

blueprint_on-prem_data_center_deployment.png

Component

Feature

Deployment Private data center can be housed and maintained by developers and researchers in their own facilities.
Third-party ISV Managed Developer Service can be leveraged to procure, install, configure and monitor the on-prem data center.
Virtualization On-prem infrastructure can be used to run a private cloud.
ARC-OTA compute resources are virtualized for gNB and 5GC.

K8 Service Management

blueprint_k8_service_mgmt.png

ARC-OTA provides Kubernetes for container orchestration. Both single node and multi-node deployment topologies are supported. ARC-OTA uses helm to manage applications.

Deployment

Description

ARC gNB The helm chart uses K8 deployment to create pod gNB. The gNB pod contains the containers nv-cubb and oai-gnb-aerial. Both containers are installed in the same pod to allow the use of shared memory between Layer 1 and Layer 2+. This helm chart is available for download from NGC.
Deployment ARC 5GC The helm-chart is installed on the CN5G server or same physical server as the gNB. This helm chart creates multiple K8 deployments depicted on left. This helm chart is available for download from the OAI GitLab repository

ARC-OTA Service Monitoring feature uses a combination of Grafana, Loki, Promtail and Prometheus. The feature was developed using open-source industry standard tools and it can be extended to specific developer needs

Reference the Sterling developer extension for additional details.

O-RAN 7.2 Split

blueprint_o-ran_7.2_split.png

O-RAN’s split-RAN concept disaggregates the RAN into multiple functional components. These components can be deployed on different hardware and software platforms and can be interconnected using open interfaces.

ARC-OTA leverages the 7.2x split, which divides the protocol stack into the following:

Component

Description

O-RU (O-Radio Unit) The O-RU is responsible for the physical layer (PHY) processing, including RF signal processing and analog-to-digital conversion.
O-DU (O-Distributed Unit) The O-DU is responsible for the higher-layer processing, including MAC, RLC, and PDCP
Fronthaul interface O-RAN alliance specified fronthaul interface between the O-DU and O-RU based on the 7.2x split. This interface supports control, user, synchronization (CUS), and management(M) planes

CSI Dataset

blueprint_csi_dataset.png

  • ARC-OTA integrated Aerial Data Lake provides the ability to capture OTA radio frequency (RF) data from the base station (BS). Raw IQ samples from the 7.2x split fronthaul (FH) interface are collected in a database file

  • Using the Aerial Data Lake database APIs, pyAerial can access RF samples in the database and transform these samples into training data or data sets for signal processing functions.

  • A sample Jupyter notebook has been provided that can create a sample multi-UE CSI dataset

  • SDKManager automation will help developers install AI/ML frameworks to easily generate a dataset

OpenRAN Gym

blueprint_openran_gym.png

  • Streaming of relevant key performance metrics (KPMs) and the enforcement of control actions to reflect decisions taken by the xApps on a Near-Real-Time (Near-RT) RAN Intelligent Controller (RIC) blueprint is shared through the OpenRAN Gym integration and an example monitoring xApp.

  • Potential xApps can be developed for network intelligence like handover optimization, policy enforcement, resource assurance or radio link management or resource control applications like load balancing or network slicing.

  • Northeastern has integrated the E2 interface with O-RAN OSC RIC and a template monitoring xApp using a custom E2 Agent to E2 Service Model. xApp Python bindings and xApp Connector collectively provide the RAN close loop monitor and control functions. This tooling helps developers incorporate network adaptability functions with AI/ML based xApps

Previous Getting Started
Next Installation Guide
© Copyright 2024, NVIDIA.. Last updated on Oct 16, 2024.