O-RAN Architecture and RAN

Intelligent Controller (RIC)

Marcin Dryjański, Ph.D.

O-RAN Webinar
Marcin Dryjanski

Co-Founder of Rimedo Labs

Working within 5G area since 2012

5G RAN Architect at Huawei R&D

Ph.D. on RRM, Senior IEEE Member

Consultant & Technical Trainer

Book ”From LTE to LTE-A Pro and 5G”

Outline

O-RAN definitions and entities

O-RAN architecture and implementation options

O-RAN RAN Intelligent Controller (RIC)

Architecture
and RIC Non-RT RIC and A1 interface

Near-RT RIC and E2 interface

Summary

rimedolabs.com
O-RAN definitions and entities

rimedolabs.com
Open RAN
Core
Network CN
(CN)

Open RAN

CU

RIC
Traditional RAN • Disaggregation
DU
• Open / enlarged vendor ecosystem
• Open Interfaces
Open FH
• Decoupling HW from SW
• Intelligent management RU/RRH
• Lower CAPEX
• Minimization of proprietary solutions

rimedolabs.com
 O-RAN Alliance Defined Entities
Application designed to run on Non-RT RIC:
SMO • independent of the Non-RT RIC
System supporting orchestration of O-RAN • may be provided by 3rd party.

rApp

rApp

rApp
components

Logical node enabling Non-RT control/optimization of RAN

elements and resources, AI/ML workflow, and policy-based
Non-RT RIC guidance of applications/features in nRT RIC

Application designed to run on near-RT RIC:

• independent of the near-RT RIC

xApp

xApp

xApp
• may be provided by 3rd party.
Logical node enabling near-RT control/optimization of RAN
elements and resources via fine-grained data collection
and actions over E2. May include AI/ML workflow.
near-RT RIC
Logical node hosting SDAP and UP part of PDCP

Logical node hosting RLC/MAC/High-PHY layers based on

Logical node hosting RRC and CP part of PDCP O-CU-CP O-CU-UP O-DU LLS

Cloud Computing platform comprising: O-RU Logical node hosting Low-PHY layer (FFT/iFFT, PRACH
• PHY infra nodes to host O-RAN fcns (e.g. near- extraction) and RF based on LLS
RT RIC, O-DU),
• Supporting SW components (e.g. OS, VM O-Cloud
Monitor, Container Runtime)
• MANO functions

rimedolabs.com
Based on defs from: O-RAN Alliance Specifications
O-RAN Alliance WGs
WG1 • UC, arch, slicing, security Focus: identify/study use cases and
overall architecture

WG2 • Non-RT RIC, A1

Focus: optimization and automation on
RRM using RIC, definition of RIC
WG3 • Near-RT RIC, E2 architecture

WG4 • Open FH (IEEE1914, eCPRI, CPRI)

WG5 • Open F1, W1, E1, X2, Xn

WG6 • Cloudification and Orchestration

WG7 • Whitebox HW reference design

WG8 • Stack reference design (CU/DU)

WG9 • Open X-Haul (FH, MH, BH)

WG10 • OAM

rimedolabs.com
 Open RAN Technical Priority Document

https://www.o-ran.org/ecosystem

Sections:
rimedolabs.com
Source: „OPEN RAN TECHNICAL PRIORITIES, UNDER THE OPEN RAN MOU”, by Deutsche Telekom, Orange, Telefónica, TIM and Vodafone
 Open RAN Technical Priorities – RIC

rimedolabs.com
Source: „OPEN RAN TECHNICAL PRIORITIES, UNDER THE OPEN RAN MOU”, by Deutsche Telekom, Orange, Telefónica, TIM and Vodafone
O-RAN architecture and
implementation options

rimedolabs.com
O-RAN Architecture SMO

Non-RT RIC

policy, enrichment
Open FH
O1 info, ML model mgmt A1 O2
M-Plane
(O1 FFS)
policy feedback

xAPP

xAPP

xAPP

xAPP
platform resources
FCAPS for O-RU: and workload
SW mgmt, CM, PM, FM, to Near-RT RIC management
File mgmt, installation (scaling, FCAPS, SW
near-RT RIC mgmt, resource
create/delete)
E2 E2 E2 E2

to O-eNB NG-c to 5GC (AMF)

O-eNB
O-CU-CP to eNB / en-gNB,
to O-CU-CP gNB / ng-eNB

X2-c, Xn-c
E1
monitor, suspend /
stop, override or
FCAPS control via policies
NG-u to 5GC (UPF)
config, reconfig, to O-CU-UP
discovery, registration,
O-CU-UP to eNB / en-gNB, gNB
data collection / ng-eNB
versioning, monitoring
and feedback X2-u, Xn-u
F1-c F1-u
performance &
resource updates
O-DU

Open FH I/F

O-RU

O-Cloud rimedolabs.com
 Selected O-RAN Implementation Options (1/2)
Disaggregated Network Functions Aggregated O-CU-CP, O-CU-UP, O-DU

SMO SMO
Non-RT RIC Non-RT RIC
O1

A1 O2 O1 A1 O2

near-RT RIC near-RT RIC

E2 E2 E2 E2

O-CU-CP O-CU-CP

E1

O-CU-UP O-CU-UP

F1-c F1-u

O-DU O-DU

Open FH Open FH Open FH Open FH

M-Plane M-Plane
O-RU O-RU

O-Cloud O-Cloud

rimedolabs.com
Based on: O-RAN.WG1.O-RAN-Architecture-Description-v02.00
 Selected O-RAN Implementation Options (2/2)
Aggregated near-RT RIC, O-CU-CP, O-CU-UP All Nodes Aggregated

SMO SMO
Non-RT RIC Non-RT RIC
O1
O2 O1 O2
A1 A1

near-RT RIC near-RT RIC

O-CU-CP O-CU-CP

O-CU-UP O-CU-UP

F1-c F1-u E2

O-DU O-DU

Open FH Open FH
M-Plane
O-RU O-RU

O-Cloud O-Cloud

rimedolabs.com
Based on: O-RAN.WG1.O-RAN-Architecture-Description-v02.00
RAN Intelligent Controller (RIC)

rimedolabs.com
 O-RAN RIC Overview
RIC is defined to enable eNB/gNB functionalities as xApps. Applications like mobility management, admission control, and
interference management are available as apps on controller, which enforces network policies towards RAN elements.

SMO
rApp1 rApp2 rAppM Non-Real-Time RIC:
• provides configuration, management and analytics
R1 Interface (visibility into network, AI-based feeds, recommendations
to near-RT RIC)
Non-RT RIC Framework • supports non-RT intelligent radio resource management,
higher layer procedure optimization, policy optimization in
RAN, and provides AI/ML models to near-RT RIC
A1 interface: Non-RT RIC
• intent based interface
• policy feedback to non-RT RIC O1 A1
• policy, enrichment info and ML
model mgmt for near-RT RIC Near-Real-Time RIC:
xApp1 xApp2 xAppN • a software platform for a set of xApps for the RAN
(e.g. QoS (e.g. mobility (e.g. slicing
optimization) optimization) optimization) • enables near-RT control and optimization of RAN elements
and resources via fine-grained data collection and actions
O1 interface: Open API for xApps over E2 interface
• data collection and control
• example use cases: network intelligence (policy
• FCAPS for RIC and RAN (data O1 Near-RT RIC Framework enforcement, HO optimization), resource assurance (radio-
collection, provisioning,
link monitoring, advanced SON), resource control (load
performance, fault mgmt)
balancing, slicing policy)
near-RT RIC
E2 interface:
• RAN closed loop E2
• RIC control and policy towards RAN
• data collection and feedback to
near-RT RIC E2 Node rimedolabs.com
O-RAN RIC: Functional Split & Control Loops
O-RAN Control Loops

from NW mgmt, NW
fcns, other sources
enrichment data
AI / ML model Service and
rApps
training policy mgmt

latency
Non-RT
• Service and policy mgmt

>> 1s
Non-RT RIC • RAN analytics
• AI/ML model training
policy, enrichment
policy feedback A1 info, ML model
management

Mobility Mgmt / Radio

Interference
Traffic Steering QoS Mgmt xApp Connection
Mgmt xApp
xApp Mgmt xApp

Trained ML
Other xApps
Model

near-RT
latency
• RAN Control and optimization

> 10ms,
xApps for use cases

<1s
•
xApp management
RAN & UE Database • UE and cell specific metrics
and security
near-RT RIC

data collection, control and

feedback E2 policy

RT latency
< 10 ms
E2 Node • Real time actions
• Resource management
• Radio scheduling, HARQ, Beamforming

rimedolabs.com
 O-RAN RIC: rApp vs xApp

xApp rApp

xApp1 xApp2 xAppN rApp1 rApp2 rAppM

Open API for xApps R1 Interface

Near-RT RIC Framework Non-RT RIC Framework

near-RT RIC Non-RT RIC

• An application designed to run on Near-RT RIC. • Application designed to run on Non-RT RIC.
• Runs inside RAN domain. • Runs inside the management plane.
• Likely to consist of one or more microservices. • Provide value added services (VAS) relative to intelligent RAN
• At point of on-boarding identifies which data it optimization and operation. Example VASs:
consumes and which data it provides. • Providing policy-based guidance and enrichment
• Is independent of Near-RT RIC and may be provided by information across A1 interface.
third party. • Performing data analytics, AI/ML training, and inference
• E2 enables a direct association between xApp and RAN for RAN optimization or for use of other rApps.
functionality. • Recommending configuration management actions
over O1 interface.

rimedolabs.com
Based on: O-RAN.WG2.Non-RT-RIC-ARCH-TR-v01.00
Non-RT RIC and A1 interface

rimedolabs.com
Non-RT RIC

• Goal of Non-RT RIC is to support intelligent RAN optimization by providing

policy-based guidance, model management and enrichment
information to near-RT RIC function so that RAN can be optimized
Non-RT RIC
• Functionality includes: configuration management, device management,
fault management, performance management, and lifecycle
management for all NW elements
rApp
• Similar to Element Management System (EMS) in legacy NWs
• All new radio units are self-configured by Non-RT RIC, reducing the need
R1 for manual intervention (important e.g. for MMIMO 5G deployments and
small cells)
• By providing timely insights into NW operations allows to optimize NW by
applying pre-determined service and policy parameters
Non-RT RIC Framework
• Can use data analytics and AI/ML training/inference to determine RAN
optimization actions for which it can leverage SMO services such as data
collection and provisioning services of O-RAN nodes
• Trained models and RT control functions produced in Non-RT RIC are
distributed to Near-RT RIC for runtime execution

rimedolabs.com
 Non-RT RIC Detailed Functional Architecture
Used for termination of an
interaction when
associated functionality
can optionally be
considered part of Non-RT
RIC framework or not.

Used for termination of

interaction when
associated
functionality is inherent
to Non-RT RIC
framework
Used for termination of
interaction when
associated functionality is
not part of Non-RT RIC
functional space (neither
part of Non-RT RIC
framework nor part of rApp)

rimedolabs.com
Source: O-RAN.WG2.Non-RT-RIC-ARCH-TR-v01.00
 Non-RT RIC – Example rApps
• UE QoE predictior
Cell utilization predictor • Consumes measurements on UE RF signal (actual RAN measurement or prediction),
RF signal predictor: • Consumes cell utilization measurement of cell site capacity utilization (actual or prediction)
• Consumes O1 measurements of RF measurements regarding actual • Calculates QoE experience by particular UE:
signal experienced by UE for capacity utiliation for a cell site over • Estimates actual QoE based on actual RF signal and actual cell util
serving/neighbor cells, time • Estimates QoE if in neighbor cell based on actual RF signal relative to neighbor cell and
measurements for UE location • Outputs future time prediction of the actual neighbor cell util
• Outputs future time prediction of cell site utilization based on the trend • Estimates future QoE if connected to serving/neighbor cell based predicted signal and
location of UE, prediction of RF signal predicted cell util
at that location for serving/neighbor

rApp R rApp U rApp Q

Consumed

Outputs
inputs

R1

Non-RT RIC Framework

Non-RT RIC

SMO Inherent
Functionality (O1) rimedolabs.com
Source: O-RAN.WG2.Non-RT-RIC-ARCH-TR-v01.00
 Non-RT RIC: A1

A1 Interface A1 Protocol Stack

• enables non-RT RIC function to provide policy-based

guidance, ML model management and enrichment
Non-RT RIC information to near-RT RIC function so that the RAN
can optimize RRM under certain conditions JSON JSON

• provides inter-connection of non-RT RIC functionality HTTPS HTTPS

in SMO domain with near-RT RIC functionality in RAN
supplied by different manufacturers TCP TCP
A1 • provides policies for individual UEs or groups of UEs
IP IP
• provides basic feedback on policy state from near-RT
RIC that enables non-RT RIC to monitor use of policies L2 L2

• provides enrichment information as required by near- L1 L1

near-RT RIC RT RIC
A1
Non-RT RIC Near-RT RIC

rimedolabs.com
Source: O-RAN.WG2.A1AP-v02.00
 Non-RT RIC: A1 Policies
RAN operation can be optimized using A1 policies that:
- are not critical to traffic;
- have temporary validity;
- handle individual UE or dynamically defined groups of UEs;
- act within and takes precedence over the configuration;
- is non-persistent, i.e. does not survive a restart of the near-RT RIC.

Non-RT RIC
A1 policy - identified by a policy identifier (PolicyId) assigned by non-RT RIC.

Scope ID - represents policy statements to be applied on (e.g. UEs, QoS flows, or cells). Identifier
for a single UE (ueId), group of UEs (groupId), slice (sliceId), QoS (qosId), cell (cellId)
A1

represents the directives to near-RT RIC and covers policy objectives and policy resources.

near-RT RIC

rimedolabs.com
Source: O-RAN.WG2.A1AP-v02.00
 Non-RT RIC: A1 Policy Examples in JSON (1/2)

rimedolabs.com
Source: O-RAN.WG2.A1AP-v02.00
 Non-RT RIC: A1 Policy Examples in JSON (2/2)

rimedolabs.com
Source: O-RAN.WG2.A1AP-v02.00
Near-RT RIC and E2 interface

rimedolabs.com
Near-RT RIC
• Near-RT RIC uses A1 and O1 for management and optimization of RAN and is
Non-RT RIC responsible for necessary optimization-related tasks across different
RANs, utilizing available RAN data from all RAN types (e.g. macro/small cells,
A1 Massive MIMO nodes)
• xApps use E2 interface to collect near RT information (on a UE or cell basis)
Near-RT RIC
• Controls over E2 nodes is steered via policies and data provided via A1 from
Non-RT RIC

xApp xApp xApp • RRM functional allocation between Near-RT RIC and E2 node is subject to
the capability of E2 node and is controlled by Near-RT RIC
• Near-RT RIC may monitor, suspend/stop, override or control node via Non-
Open API RT RIC enabled policies
• In the event of a Near-RT RIC failure, E2 Node shall be able to provide
services, but there may be an outage for certain VAS provided using Near-
RT RIC
Near-RT RIC Framework
• Leverages embedded intelligence and is responsible for per-UE controlled
load-balancing, RB management, interference detection and mitigation
• Radio-Network Information Base (R-NIB) captures near RT state of
underlying network and feeds RAN data to train AI/ML models, to be fed to
E2 Near-RT RIC to facilitate RRM
• Interacts with Non-RT RIC via A1 interface to receive trained models and
E2 Node execute them to improve network conditions

rimedolabs.com
 O-RAN Near RT RIC – Internal Architecture
SMO

Non-RT RIC
Terminates O1 interface from SMO layer and
forwards management messages to Near-RT RIC Terminates A1 interface from Non-RT RIC and forwards
management function O1 A1 A1 messages

near-RT RIC
O1 Termination A1 Termination
• allow RRM control functionalities to be
executed at Near-RT RIC and enforced in E2
Nodes;
Open API • initiates xApp-related transactions over E2;

xApp 1 xApp 1 … xApp N

• handles xApp-related responses from the E2;

merges subscriptions from different xApps and

provides unified data distribution to xApps message interaction amongst Near-RT RIC internal
Messaging Infrastructure functions

xApp
Conflict Mgmt
resolves potentially overlapping or conflicting Subscription Security security scheme for xApps
Mitigation services
requests from multiple xApps Mgmt

Shared Data Layer • Fault, configuration management, and

R/W of RAN/UE information performance management
E2 Termination • Life-cycle management of xApps
captures near RT state of underlying network and Database • Logging, tracing and metrics collection and
feeds RAN data to train AI/ML models transfer to external system for evaluation
• Terminates E2 interface from E2 Node;
• Routes xApp-related messages to target xApp; E2
• Routes non xApp-related messages to E2
Manager;
E2 Node(s) rimedolabs.com
Based on defs from: O-RAN.WG3.RICARCH-v02.00
 Near-RT RIC Implementation Options

Centralized near-RT RIC

(serving 4G and 5G simultaneously) Distributed near-RT RIC

SMO SMO

Non-RT RIC Non-RT RIC

A1 A1 A1 A1 A1 A1 A1

near- near- near- near- near-

near-RT RIC near-RT RIC RT RIC RT RIC RT RIC RT RIC RT RIC

E2 E2 E2 E2 E2 E2 E2 E2 E2 E2 E2 E2

O-CU- O-CU- O-CU- O-CU- O-CU- O-CU-

O-DU O-DU O-eNB O-eNB O-DU O-eNB
CP UP CP UP CP UP

near-RT RIC serving gNB near-RT RIC serving gNB and eNB near-RT RIC serving eNB
One near-RT RIC entity
consisting of one on more
logical near-RT RICs

rimedolabs.com
Based on defs from: O-RAN Alliance Specifications
 Near-RT RIC: E2 Interface
Architecture
E2 interface shall facilitate:

• Connectivity between Near-RT RIC and E2 Node supplied by different vendors; Near-RT RIC
xApp xApp xApp
• Exposure of selected E2 Node data (e.g. configuration information (cell configuration,
supported slices, PLMNs, etc.), network measurements, context information, etc.) towards
the Near-RT RIC
E2 Termination Database
• Enables the Near-RT RIC to control selected functions on E2 Node
E2 Holds data from xApp
and E2 Node, provide
data to xApps
E2 interface functions categories:
Terminate E2 Interface and
• NEAR-RT RIC services: forward/receive E2 msg
E2 Agent
• Near-RT RIC Services (REPORT, INSERT, CONTROL and POLICY).

• NEAR-RT RIC support functions: Func (1) Func (N)

• Interface Management (E2 Setup, E2 Reset, E2 Node Configuration Update, E2 Node

Reporting of General Error Situations) Func (1)
Other fcns E2 Node
E2 Node
• Near-RT RIC Service Update, i.e. a E2 Node initiated procedure to inform Near-RT
RAN functions controlled by
RIC of changes to list of supported Near-RT RIC services and mapping of services
near RT RIC, supporting Near- RAN functions that don’t
to functions. RT RIC Services support Near RT RIC Services

rimedolabs.com
Source: O-RAN.WG3.E2GAP-v01.01
 E2 Service Procedure Example: CONTROL
Near-RT RIC CONTROL Service initiated by NEAR-RT RIC

rimedolabs.com
Source: O-RAN.WG3.E2GAP-v01.01
E2 – Example Data Towards Near RT RIC

SMO

Non-RT RIC

Example E2 Data (per UE data w/ Near-RT granularity)

• UE-specific network interface protocol message IE IDs
• UE-specific PHY signal strength and RRC measurement A1
values
• Per-UE per DRB per flow thrpt / latency / buffer occupancy
• Per-UE PRB utilization, RLC PDU segmentation, QoS flow to
DRB mapping
• Per-UE PDCP duplication indication and data volume
• Per-UE MCS allocation of HARQ PDUs and number of near-RT RIC
scheduled TTIs
• Per-UE numer of HARQ/RLC retx

E2

O-CU-CP, O-
CU-UP, O-DU

rimedolabs.com
 E.g. xApp Operation: E2 CONTROL API Procedure

rimedolabs.com
Source: O-RAN.WG3.RICARCH-v01.00
Summary

rimedolabs.com
O-RAN Architecture and RIC
O-RAN Alliance defines the following entities: O-CU, O-DU, O-RU, Near-RT RIC, Non-RT RIC, SMO, O-Cloud,
xApps, rApps, and interfaces: O1, O2, E2, A1, Open-FH, M-Plane Open FH

O-RAN Alliance allows various implementation options, where individual building blocks can be
combined, which enables flexibility

RIC provides control functionality, aiming at delivering increased efficiency and optimal NW
performance for optimal UX

Non-RT RIC acts as a configuration layer to DU and CU software through policies via near-RT RIC, while
near-RT RIC provides a software platform for xApps for RAN management and optimization

A1 enables the non-RT RIC function to provide policy-based guidance, ML model management and
enrichment information to the near-RT RIC function

E2 enables connectivity between Near-RT RIC and E2 Node supplied by different vendors and enables Near-RT
RIC to control selected functions on the E2 Node

rimedolabs.com
Your trusted partner in: Let’s keep in touch!
LTE, 5G & Beyond, RRM,
Wi-Fi, IoT, O-RAN,
RIMEDO Labs
ul. Polanka 3

Private Mobile Networks.

61-131 Poznan, Poland

info@rimedolabs.com
rimedolabs.com

Marcin Dryjanski
marcin.dryjanski@rimedolabs.com
O-RAN System Training (courses.intelefy.com/courses/oran-system-training)
 Check out our blog! Download our Whitepaper!

https://rimedolabs.com/blog https://bit.ly/3B9k7mm
Who We Are? Dynamic Spectrum Sharing

Radio Resource Management

5G
Wi-Fi Wireless Private Mobile Networks
RIMEDO Labs specializes in providing the best quality
and substantive consulting, implementation and R&D 6G Systems
IoT
services in the field of modern wireless systems. Ultra Dense Networks
LTE
We implement this through an individual and open Novel Wireless Concepts
approach to the client, constantly improving the team
operationally and substantively, updating knowledge
and a unique combination of science and business
applications. #Research
RIMEDO Labs is a spin-off from the Poznan University of #Engineering
#Consulting
Technology, Poland from the Institute of
Radiocommunications.

rimedolabs.com
 Applied Research Consulting Training Technical Content Delivery

The areas of our specialization Having extensive experience in Our training services include We provide technical contents for
cover wireless systems (like LTE, the field of modern wireless online and on-site courses, external training or consultancy
5G, 6G, IoT, Wi-Fi), spectrum systems we offer high quality conferences, meetups or companies delivered as training
sharing and management, radio consulting and advisory services workshops tailored to customer’s materials, technical documents,
resource management, AI for delivered by our seasoned needs and requirements. The dedicated research papers,
wireless systems and private engineers and consultants. topics, which are covered by us books, book chapters, slides,
mobile networks. We offer our RIMEDO Labs Consulting include include: 4G, 5G and beyond, IoT, reports or raw materials for
expertise as part of consortiums cover, among others the following Wi-Fi, spectrum management, further processing. The material
for EU and National funded items: radio planning and site radio resource management, can be developed as insights
projects (like Horizon 2020, Horizon surveys, technology forecasting, private networks, design, planning onto a specific feature or aspect
Europe, NCBR, etc.). We can take preparation of feasibility studies, and troubleshooting of wireless within wireless systems area,
part in those projects as leader, systems architecting, wireless systems, artificial intelligence for including topics like: LTE, 5G and
partner or subcontractor. systems patent analysis, wireless systems. Our top-class beyond, Wi-Fi, IoT, shared
standards tracking, or expert/R&D instructors combine scientific and spectrum, AI, etc. The educational
team outsourcing. educational background with content, can be also delivered in
practical experience. We speak the form of virtual radio labs.
about the systems we design.

rimedolabs.com
The information contained herein is the property of RIMEDO and is provided only if it is not disclosed, directly or indirectly to a
third party, or used for purposes other than those for which it was prepared.

ETSI is the copyright holder of LTE, LTE-Advanced and LTE Advanced Pro and 5G Logos. LTE is a trade mark of ETSI. RIMEDO is
authorized to use the LTE, LTE-Advanced, LTE-Advanced Pro and 5G logos and the acronym LTE.

All information discussed in the document is provided "as is" and RIMEDO makes no warranty that this information is fit for
purpose. Users use this information at their own risk and responsibility.

© 2021 RIMEDO sp. z o.o. All rights reserved.