View these slides, presented by Prakash Siva, VP, Technology & Strategy, hosted by Intel Network Builders, around the subject of Mobile Edge Computing.
Presentation on how to chat with PDF using ChatGPT code interpreter
Getting to the Edge – Exploring 4G/5G Cloud-RAN Deployable Solutions
1. Getting to the Edge – Exploring 4G/5G
Cloud-RAN Deployable Solutions
Jul-31 ‘18
Prakash Siva
VP, Technology & Strategy
psiva@radisys.com
2. 2
The Webinar Today
At a time when the telecommunications industry is in significant
transition, Communication Service Providers (CSPs) and equipment manufacturers
are turning to open source and multi-access edge computing (MEC) solutions. These
solutions are required to meet key infrastructure constraints such as cost efficiencies,
deployment flexibility and deliver mission- critical services with high quality. Intel®
with their Network Edge Virtualization (NEV) and FlexRAN initiatives is at the
forefront of providing cutting edge platform and framework to build these solutions.
These enable CSPs to accelerate and innovate their 4G/5G cloud RAN initiatives and
ensure better user experience by moving compute intensive, latency-sensitive
applications to the network edge.
Radisys is committed to ensuring that open telecom initiatives are not ‘science
projects’ – but actual, deployable solutions that can be commercialized with faster
time-to-market. This webinar will focus on:
1. Key considerations required for rapid integration and deployment of mobile edge
services using Intel®-based Mobile Edge Compute (MEC) infrastructure
2. Key considerations required for 4G and 5G Cloud-RAN implementations using
Intel®’s FlexRAN development framework
3. The role and purpose of multiple open initiatives such as: xRAN, TIP, ONAP,
ONF in evolving the technology stack of service providers along with associated
integration requirements and challenges
Getting to the Edge – Exploring 4G/5G Cloud-RAN Deployable
Solutions
Prakash Siva
VP, Technology & Strategy
Radisys
psiva@radisys.com
3. 3
The Next 45 Minutes
Why the Edge?
• The importance and challenges of Mobile Edge
• Technology & Business Drivers
Challenges of Building an Edge Focused Mobile Solution
• Pulling together a Mobile Edge focused architecture – solution domains and choices
• MEC Placement & System Integration
Solving the Integration Challenges
• Intel® Architecture and Radisys Solutions
• Integrated solutions for Use Cases – RAN aware Media Breakout, Edge Media Analytics, Firmware Over the Air (FOTA)
1
2
3
4. 4
Demonstrating Leading Edge of Open Mobile Innovation
Real Time Video Analytics, Content
Overlay (AR) & Event TriggersRAN Aware Media Optimization
at the Edge
Over the Air upgrade of Vehicle
Firmware
1 2 3
Real-time Media
Breakout
Analytics Firmware OTA
X-RAN/ ORAN – Policy Driven Control & Management of eNB using RAN Controller
ONAP Orchestration - Policy Driven Automation of Virtual Network Functions
Edge Applications
5
4
6. 6
The 4th Industrial Revolution
Connected Everything: People, Places and Things
Connected Hardware
Cost
Connected People
Innovation
Connected Everything
Flexibility
50B
Connected
Devices
1000x
Mobile Data
Volumes
10x-100x
End-user Data
Rates
5x
Lower Latency
7. 7
RAN Re-Architecture - A MUST to Evolve 4G and 5G Networks
Proprietary
Vendor lock-ins, slow time to market and high
innovation cost
Embedded
Integrated control and data, proprietary
interfaces
Fixed
Pre-programmed control logic and
fixed network resources
Integrated
Integrated RRU & BBU with
high cost last mile transport
Software-Defined
Open white box hardware and software
(NFV, SDN) efficiencies
Unbundled
Disaggregated control / data,
OpenAPIs
Programmable
Programmable based on real
world network conditions
Flexible
Deployment architecture
and front haul agnostics
TraditionalRAN
NextGenerationOpenRAN
Software
Modularity
Application and
NetworkAdaptability
Deployability and
Manageability
Platform
8. 8
RAN Evolution: From Distributed & Integrated to Virtual & Centralized
Synch
Control
Transport
Broadband
Antenna
cell
vBBU
S1/X2
RF RRH
Synch
Control
Transport
Broadband
RF
PA
Antenna
cell
BS
Antenna
RF RRH
Synch
Control
Transport
Broadband
Virtual BBU Pool
S1/X2
Antenna
RF RRH
Antenna
RF RRH
Fiber – Digital BaseBand Coax cable – RF
A) Traditional macro base station B) Base station with RRH C) C-RAN with RRHs
Evolution
Distributed
RAN
Distributed
vRAN
Centralized
vRAN
• Control plane (CP) and user plane (UP) using standard interfaces
• Open APIs between L1/ L2/ L3 to support Multi-Vendor RAN solution
• Modular Radio Access Unit (RAU) supporting standard interfaces
9. 9
Multiple Forces Driving the Need For an Edge Cloud
LAW OF PHYSICS LAW OF ECONOMICS LAW OF THE LAND
• Customers want to build
applications that make the
most interactive and critical
decisions locally, such as safety-
critical control.
• This is determined by basic
laws of physics: it takes time to
send data to the cloud, and
networks don't have 100%
availability.
• In many industries, data
production has grown more
quickly than bandwidth, and
much of this data is low value.
• Local aggregation and filtering
of data allows customers to
send only high-value data to
the cloud for storage and
analysis.
• In some industries, customers
have regulatory or compliance
requirements to isolate or
duplicate data in particular
locations.
• Some governments impose
data sovereignty restrictions on
where data may be stored and
processed.
Source: L. Peterson, CTO, ONF & Wikipedia
LAW OF BIOLOGY
• Vestibulo-Ocular Reflex (VOR) is
a stabilizing function of the
human eye to compensate for
head movements.
• Immersive experiences rely
upon VOR - eye movements lag
head movements by ~10ms
• Longer visual feedback times
cause motion sickness
10. 10
The New Mobile Edge
Source: “Future X Network” & GSMA 5G Intelligence
Increasing Mobile BW Needs
4G/5G Caters to new enhanced mobile
broadband, mission critical and ultra
low-latency application demands
Reduced Latency Requirements
Computing Edge is Shifting
Closer to the Customers
New
Mobile
Edge
11. 11
Open Virtualized RAN & Edge Architecture for 4G and 5G Networks
Virtualized mobile core for 4G and 5G
Networks with end to end Network Slicing
Programmable RAN Control plane using
open reference architectures
Centralization and Virtualization of RAN
eNodeB functions + Mobile Edge Apps
Unbundled RAN (CU/DU) using standard APIs
Modular RRU software functions
supporting multiple RAT co-existence
Distributed Remote Radio Unit (RRU) — Data Unit (DU)
3G / 4G
Small Cells
4G+ NR
Capable
5G NR
Front-haul
eNB/gNB
Back-haul
EPC 5G Core
Control Unit (CU)
Multi-access Edge
Centralized vBBU
12. 12
Question #1:
Which of the following use cases or applications are highest priority for you
to drive Multi-access Edge Computing (MEC) adoption?
A. Video & Edge Content Caching
B. AI & Analytics (including AR/VR)
C. Autonomous Vehicles & Transportation
D. Industrial IoT
E. Emergency/Disaster Communication Services
Poll: Audience to select highest priority item
Result: Present % choice for each letter selection
Poll Questions #1
14. 14
What does it take to pull together a Mobile Edge Solution?
3 major problems to be solved
= A Domain
• ETSI MEC Architecture is used as an example, but
it’s use is not mandatory
• Each domain represents a specific aspect of the
solution with a number of potential choices
• Each choice introduces tradeoffs that affect the
overall solution architecture
• We’ve not explored all domains (OSS, UE apps etc.
are left out)
#1: Identifying choices
that map to each
domain & location
#2: Software and
Hardware Integration of
those choices
#3: Functional &
performance testing,
including scaling
15. 15
Domain Choices for Complete Solution
Independent Choices in each Domain
• #1: 3GPP Network Choices
• RAN: Intel® FlexRAN or 3rd party; multiple split choices
• Packet Core: 4G, 4G with CUPS support, 5G Core
• #2: MEC Host Location Choices
• Cell Site, Aggregation site (Tier1, 2, 3), Edge site
• #3: Data plane Choices
• Local Breakout, Inline (uni vs. bi-directional), Tap
• #4: NFVi Choices
• X86 CPU, GPU, FPGAs (NIC offload, Inline Offload), Multiple Network
Switch Types
• #5: VIM & Compute environment Choices
• Bare-metal, VM/Containers; OpenStack, Container Based Micro-services
• #6: MEC Mobility Choices
• Intra vs. Inter host mobility, Stateful vs. stateless application mobility
• #7: Mobile Edge Platform (MEP) Choices
• Intel® NEV (now part of Akraino), 3rd party
• #8: Edge Orchestrator & MEP Manager
• ONAP, Cloudify, 3rd party; Extent of Global vs local orchestration
= A Domain
16. 16
3GPP 4G/5G RAN Deployment Choices
Source: 3GPP, NGMN
RAN Split for
use cases
Centralized Unit (CU)
Distributed Unit (DU)
Radio Unit (RU)
• Use cases support Option 2, Option 7-1 & Option 8 splits
• CU & DU execute as VMs on x86 servers using BBU pooling
framework
• Intel® Hi-Phy, RRC/PDCP/RLC/MAC from Radisys
• Data path in CU/DU uses DPDK & fd.io for high performance with
SR-IOV
• Mapping of cells to VMs
• Each CU can control up to #n DUs (depending on system load)
• DU can register dynamically with CU
• Certain Low-PHY functions could be implemented with HW
acceleration (FPGA)
Option 2: High
Level Split (HLS)
Option 7-1: Low
Level Split (LLS)
eNB/gNB
CU DU RU
17. 17
Overall Architecture & Site Choices for RAN, MEC Deployment
Optimal Choice
Source: NGMN
• Function placement depends on specific CSP requirements
• Physical site constraints + Transport network constraints – topology,
latency, capacity
• System integrator can propose overall design/architecture and help
test/validate optimization tradeoffs
3 major site choices
for MEC:CU
DU
RU
Edge Site
Aggr. site
Cell site
18. 18
Latency for Multiple Tiers inside Aggregation Block
Source: 3GPP, NGMN
Indicative
Numbers
UPF Location Tier 1 Tier 2 Tier 3
Relative # of sites 1000 100 10
1-way Transport Latency 0.6ms 1.2ms 4.2ms
Est. 5G Latency (RTT)
for UPF presence
9.2ms [eMBB] 10.4ms [eMBB] 16.4ms [eMBB]
2.2ms [URLLC] 3.4ms [URLLC] 9.4ms [URLLC]
Aggregation Sites • Transport latency figures based on 95th-
percentile of transmission delay (i.e. 95% of cell
sites are within this)
• Assumptions:
• Maximum Latency figures based on 95th-
percentile of transmission delay (i.e. 95% of cell
sites are within this)
• 5G RTT assumes 8ms overhead for 5G New
Radio & Next-Gen Core (eMBB case), 1ms for
URLLC (as per 3GPP release 15)
Cell Sites
19. 19
RAN & MEC Placement to meet Latency Requirements
MEC workload placement for ~1ms RTT (ultra-low latency) CU/DU placement for ~10ms RTT (low latency)
MEC placement for ~5ms RTT (very-low latency)
MEC
MEC
MEC
MEC MEC
MEC
$$$ cost
$$ cost - optimal
$ cost
20. 20
MEC Edge Placement for Different Application Workloads
Cell sites
The New Mobile Edge
Reduces
latency &
improves QoE
Enable new
services
based on
location
Dynamic
service
creation
Reduce load
on back haul &
core network
MEC
MEC
MEC
MEC
Aggr. sites
21. 21
Question #2:
What is the most critical challenge you see in deploying Multi-
access Edge Computing (MEC) applications?
A. Vendor Readiness
B. Multi-vendor Solution Integration
C. Maturity of Standards
D. Skills of internal teams
E. Cost & RoI Justification
Poll: Audience to select highest priority item
Result: Present % choice for each letter selection
Poll Question #2
24. 24
FlowEngine™MediaEngine™MobilityEngine™ ServiceEngine™
Conceptualization
Software
Customization
Network Design
and Integration
Testing and
Automation
Field Deployment
Global Support
Strategic
Focus
Markets
Value to
Customers
Traffic
Distribution &
Programmable
Data Plane
SDN, NFV, Security,
Service Function
Chaining, TAP, Load
Balancing
> 80% Lower Cost per
40G Port
> 50 Reduction in
Recurring OPEX
> 40% CAPAX
Savings
> 50% Reduction in
CAPEX and TCO
Virtual RAN, LTE-A,
CBRS, 5G, MEC and
IoT
VoLTE, VoWiFi,
Conferencing, Video,
Transcoding
Virtualized Media
Server Processing for
Rich Media Services
Software Defined,
Unbundled and Open
Cloud RAN Solutions,
3GPP, xRAN, CORD,
TIP Compliant
Strategic Product and Service Portfolio
26. 26
Management Layer
• SDN control
• VNF management
• Orchestration management
• Orchestration tool
Platform Agnostic Solution
Cloud Native Platforms
Cloud Edge & 3rd Party Mobile Edge Applications
CDN Node Video Apps
AI Inference &
Analytics
Cloud Native Platforms – Data Center
Compute (x86) Networking Storage
NFVI ( VIM – OPENSTACK or Container Mgmt/Kubernetes)
VIRTUALIZED
OLT
(Software defined
broadband access)
Radio NW
Services
VIRTUALIZED
RAN
(Baseband Unit –
BBU)
Location
Services
Networking
Services
(x86 Data Plane,
Packet Routing &
Filtering)
EdgeComputingManagementServices
SDN-C
(3GPPCUPS)
VNFManager
Orchestrator
(ONAP)
RNIS APIs USER INFO
APIs
NW APIs
ETSI MEC APIS –
mp1 & mp2
Radisys Edge Computing Platform Software
VNF/Container Onboarding Layer
Software services around onboarding
3rd party MEC applications as
containers/VNFs
Multi-Access
• Wireless (RAN
solution)
• Wireline (software
defined broadband
access)
• P4 based Data Plane
Software
Cloud
Applications
Internet
Traditional
CDN
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE NETWORK
(MNO EPC/BNG)
CLOUD APPLICATIONS
INTERNET
TRADITIONAL CDN
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE NETWORK
(MNO EPC/BNG)
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE NETWORK
(MNO EPC/BNG)
CLOUD APPLICATIONS
INTERNET
TRADITIONAL CDN
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
Management Layer
• SDN control
• VNF management
• Orchestration management
• Orchestration tool
Platform Agnostic Solution
Cloud Native Platforms
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE NETWORK
(MNO EPC/BNG)
CLOUD APPLICATIONS
INTERNET
TRADITIONAL CDN
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE
(MNO
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
User Info Services
Networking Services
(x86 DATA PLANE, PACKET
ROUTING & FILTERING)
Radio NW
Services
EDGE COMPUTING PLATFORM
SERVICES
RNIS APIs USER INFO
APIs
NW APIs
CDN Node Enterprise App Analytics Probe
CLOUD EDGE 3rd
PARTY MOBILE EDGE APPLICATIONS
COMPUTE (x86) NETWORKING STORAGE
NFVI ( VIM - OPENSTACK/MIRANTIS/WINDRIVER/.... )
CLOUD NATIVE PLATFORMS – DATA CENTER
EDGECOMPUTINGMANAGEMENTSERVICES
SDN-C
(3GPPCUPS)
VNFMANAGER
ORCHESTRATOR
(ONAP)
ETSI MEC APIS – mp1
&mp2
VIRTUALIZED RAN
(BASEBAND UNIT –
BBU)
CORE NETWORK
(MNO EPC/BNG)
CLOUD APPLICATIONS
INTERNET
TRADITIONAL CDN
VIRTUALIZED OLT (sw
defined wireline
access )
PON
WHITEBOX
OLT
ONU/
ONT
SPLITTER
ONU / ONT Whitebox OLT
Splitter
PON
Radisys
Software
Edge Computing Platform Services
27. 27
Intel® HW/Software Based Solution
Management Switch
Fabric Switch
Head Node
OpenStack, XOS, ONOS, XRAN
CORD Compute Nodes
FlexRAN+ EPC: MME, SGW, PGW
MEC Hosted vApps (vMRF, AR/VR)
MEC Host SW Platform (with RNIS)
ONAP Orchestration Node
P4 Data Plane x86 DPDK
PSUs
ONAP
MEC vApps
MEC Host
Mobile
-CORD
POD
P4 Data
Plane
RRU
Open Compute
Use Case
Config (HW)
Potential
Intel® Content
X86 Servers
NEV SDK
+ OpenVINO
NICs w/DPDK supp.
x86 Servers,
FPGA (for offload)
NICs w/DPDK supp.
x86 Servers,
4G/5G FlexRAN,
NGIC Packet Core
28. 28
Demonstrating Leading Edge of Open Mobile Innovation
Real Time Video Analytics, Content
Overlay (AR) & Event Triggers
RAN Aware Media Optimization
at the Edge
Over the Air upgrade of Vehicle
Firmware
1 2 3
Real-time Media Breakout Video Analytics Firmware OTA
X-RAN/ ORAN – Policy Driven Control & Management of eNB using RAN Controller
ONAP Orchestration - Policy Driven Automation of Virtual Network Functions
Edge Applications
5
4
29. 29
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
P4 Programmable Data Plane
APIs for
Traffic
Classification
And Control
Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
3rd Party
XRAN/ORAN
Controller
RSYS L2/L3 +
Intel® FlexRAN
The Integration Solution Functional Blocks
MEC Host SW Platform
Radio Network Services
vApps
(FOTA, AR,
Analytics)
Virtual Media Server
Live Video/Content
Application Server
Open Source + Radisys
Location Services
30. 30
Use-case #1: RAN Aware Real-time Media Breakout at the Edge
Multiple
Live High BW
HD Video
Feeds
1
Policy driven
•Time
•Location
•Service
Lower backhaul cost due to
reduced amount of traffic sent to
the core network
Increased revenues via
personalized consumer viewing
experience
Deployment Areas: Stadiums,
Enterprises, Large Events, …
Increase Revenue Reduce Cost & Improve Experience
Reduced latency for user
viewing request
Media
Processing
MEC Host
Server Platform
Content
Recording
Core
Network
RAN
Slice 2
RAN
Slice 1
RAN
Slice 3
Slicing Optional
User
Selected Multi-view
Live Content Viewing
Transrated
&
Transcoded
Video
31. 31
Use Case #1: Integrated Solution for Local Media Breakout at the Edge
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
MEC Host SW Platform
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
P4 Programmable Data Plane
Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
Radio Network Services
Virtual Media Server
vApps
(FOTA, AR,
Analytics)
Live Video/Content
User
Selected
Multi-view
Live Content
Views
Application Server
Key Components
for this Use case
33. 33
Use-case #2: Edge Media Analytics with AI based Inference
Triggered Content
Recording
Audio + Video
Feeds
ML + AI
Video Analytics
Meta Data
Core
Network
Low Bandwidth Event
Notifications
And Alarm
Triggers for Core
Network
2
Policy driven
•Time
•Location
•Service
RAN
Slice 1
RAN
Slice 2
RAN
Slice 3
Media
Analytics
MEC Host
Server Platform
Lower backhaul cost due to
reduced amount of traffic sent to
the core network
Service Opportunities: AR Overlay &
Event-based Triggers of Audio &
Video Feeds (Ex: Safety/Surveillance)
Potential Verticals: Smart
Cities, Retail, Healthcare,
Industrial IoT
Increase Revenue Reduce Cost
Cheaper edge devices and
deployment costs (ex: Cameras,
sensors, etc.)
High Bandwidth Media Content
Edge Cameras & Video Feeds Slicing Optional
34. 34
Implemented Use cases for Video Analytics
Facial recognition on trained images
People Detection & Counting
Detect camera tampering
Vehicle Detection & Counting
35. 35
Use Case #2: Integrated Solution for Edge Media Analytics & AR
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
MEC Host SW Platform
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
VoLTE
Voice
Video
Internet
Data
P4 Programmable Data Plane
APIs for
Traffic
Classification
And Control
Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
vApps
(AR &
Analytics)
Radio Network Services
Virtual Media Server
Live Video/Content
Application Server
Key Components
for this Use case
36. 36
Radisys Media Server Implementation Approach
Media Server
Application
integrated with
Intel® Media SDK
& Intel®
OpenVINO
Offload
compute
intensive media
processing
tasks to GPU
Video RTP
IP Surveillance
cameras encode and
send data to the
Media Server
Media Server
Performs video
analytics and only key
events are encoded
and archived
Surveillance
Camera
Surveillance
Monitor
• Accelerated Media Processing on Gen-9
GPUs using Intel Media SDKs
• Machine Learning (ML) based Video
Analytics using Intel’s OpenVINO SDK
• Integration with Intel’s Network Edge
Virtualization (NEV) MEC Platform SDK
Video RTP
37. 37
Use-case #3: Over The Air (OTA) Upgrade of Vehicle Firmware
MEC
Edge Cloud
MEC
Edge Cloud
• Augments scalability by adding
new features and infrastructure to
products after they are released.
• Continuous Improvement: Bugs
and product behavior can be
continuously improved even after
the device is in the hands of
consumers.
• Agility: Companies can test new
features by sending updates to
one or multiple devices.
• Time to Market: Developers can
deploy frequently and
reliably, knowing that products will
stay functional as updates are
released.CSPs can save on
firmware transport costs
Higher reliability and policy enforcement
to prevent upgrade failures
3
38. 38
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
MEC Host SW Platform
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
P4 Programmable Data Plane
Vehicles that
need to
receive
firmware
upgrade Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
vApps
(FOTA) +
Firmware
Content
Radio Network Services
#A
#B
#C
Virtual Media Server
Live Video/Content
Application Server
Use Case #3: Reliable FOTA from the Edge
39. 39
Solution Orchestration using ONAP
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
MEC Host SW Platform
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
VoLTE
Voice
Video
Internet
Data
P4 Programmable Data Plane
APIs for
Traffic
Classification
And Control
Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
Radio Network Services
vApps
(FOTA, AR
Analytics)
Virtual Media Server
Live Video/Content
Application Server
4
41. 41
M-CORD
(XOS, ONOS, OpenStack)
eNodeB
vRAN
vMME vSGW vPGW
MEC Host SW Platform
IP/Transport
Network
Internet
ONAP
Orchestration &
Automation
VoLTE
Voice
Video
Internet
Data
P4 Programmable Data Plane
APIs for
Traffic
Classification
And Control
Edge
User
Traffic
Service Orchestration
Lifecycle Management
Edge Cloud
Centralized
Cloud
Radio Network Services
eNB Control using XRAN/ ORAN Controller
XRAN/
ORAN
vApps
(FOTA, AR
Analytics)
Virtual Media Server
Live Video/Content
Application Server
5
42. 42
Open Radisys – Taking Pain Out of Open Integration
Broadband. Mobile. Media
Software
Customization
Network Design
and Integration
Testing and
Automation
Global
Support
Proof of
Concept
Field
Deployment
Break Vendor Lock of Incumbent
Proprietary Expensive Solutions
Enable Open Telecom Solutions: Higher scale
and performance at lower cost; Reduce CAPEX
and OPEX
Improving Business Model
Expanding gross margins driven by improved
software revenues; sustainable profitability on
lowered expense structure
DevOps Core Philosophy
Agility enabling Rapid Service Creation:
accelerate integration
and TTM
25 years of experience with System
Integration
With deep enabling core technologies unlike
other integrator - Trusted Partner
45. 45Radisys Corporation - CONFIDENTIAL
Contact Us
Please email Iriana.Rodriguez@radisys.com to set up a meeting with one of our MEC
solution experts.
Thank you for your interest in Radisys! Please contact us for more information.