Akraino, EdgeX, CORD, OpenEdge, ioFog…

What’s the best solution for your edge?

Ruoyu Ying, [email protected]
Kailun Qin, [email protected]
Agenda
• Edge computing definition
• Typical edge computing implementations, use cases and their features
• Available edge computing projects in the market
• Summary
Definition of Edge Computing

EDGE COMPUTING IS THE PLACEMENT OF ENDPOINT

DATA CENTER-GRADE
Closer to DEVICES
NETWORK,
COMPUTE & STORAGE

TO IMPROVE Optimize Comply with Data And reduce application

SERVICE TCO Locality Latency
CAPABILITI
ES
Outmost BEFORE TRANSITION
layers of TO THE ENDPOINT
The Edge
Is the Processing OR
or Network Another Network

3
Edge Characteristics
Low Latency, Real Time, Optimized Infrastructure and Rapid Response

Massive Various Data Storage and Movement, Data Sovereignty

Enhanced Security and Data Privacy

Context or Location Awareness, Localization

Multi-Access Networking across Large-Scale and Small-Size Sites: Unreliable, Limited, High-Bandwidth

Intelligence, Smartness, Autonomy, Zero-Touch, Self-X

4
Typical implementations of edge computing
Definition of the implementation
• Fog computing(FC): “A decentralized Computing infrastructure based on Fog
Computing nodes (FCNs) placed at any point of the architecture between the end devices and the
cloud. The FCNs are heterogeneous in nature and thus can be based on different kinds of elements
including but not limited to routers, switches, access points, IoT gateways as well as set-top
boxes.”

• Mobile/Multi-access Edge Computing(MEC): “To bring computational

and storage capacities to the edge of the network within the Radio Access Network to
reduce latency and improve context awareness. The MEC nodes or servers are usually
co-located with the Radio Network Controller or a macro base-station. The servers run
multiple instances of MEC host which has the capabilities to perform computation and
storage on a virtualized interface.”

6
Definition of the implementations – Cont’d
• Cloudlet(CC): “Treated as ”data center in a box” running a virtual machine capable
of provisioning resources to end devices and users in real time over a WLAN network.
The services are Cloudlets are provided over a one-hop access with high bandwidth, thus
offering low latency for applications.”

Reference:
[1] Koustabh Dolui and Soumya Kanti Datta, “Comparison of Edge Computing Implementations:
Fog Computing, Cloudlet and Mobile Edge Computing”. 1-6. 10.1109/GIOTS.2017.8016213.

7
Characteristics of the implementations
Type of Implementation FC MEC CC
Location Near end device, dense Radio Access Network Local/Outdoor
and distributed Controller/Base station Installation in one place
Device Routers, Switches, Servers running in base Compact-size data
Access points, station or CO centers
gateways…
Access Mediums(mostly) WiFi, LTE, ZigBee, MQTT, WiFi, LTE… WiFi…
Bluetooth…
Logical Proximity One/multiple hops One hop One hop
Ability for near-real-time High Medium Medium
Interaction
Multi-tenancy Supported Supported Supported
Computation power Medium High High

8
Characteristics of the implementations – Cont’d
Type of Implementation FC MEC CC
Power Consumption Low High Medium
Context Awareness Medium High Low
Coverage Low High Low
Server Density Medium Low High
Cost/CAPEX Low High Medium
Traffic Continuity High Medium High
Active users High Medium Medium

Reference:
[1] Koustabh Dolui and Soumya Kanti Datta, “Comparison of Edge Computing Implementations:
Fog Computing, Cloudlet and Mobile Edge Computing”. 1-6. 10.1109/GIOTS.2017.8016213.
[2] Eugen Borcoci, “Fog Computing, Mobile Edge Computing,
Cloudlets - which one?”, 2016
[3] Baktir, Ahmet Cihat & Ozgovde, Atay & Ersoy, Cem. (2017). How Can Edge Computing Benefit from
Software-Defined Networking: A Survey, Use Cases & Future Directions. IEEE Communications Surveys &
Tutorials. PP. 1-1. 10.1109/COMST.2017.2717482. 9
 Edge use case overview

Reference:
[1]
https://www.mckinsey.com/industries/high-tech/our-insights/new-demand-new-markets-what-edge-computing-means-fo 10
r-hardware-companies
Relationship between edge implementation and use case
Use case features
Band- Latency Extensi Context Power Scalability Privacy &
width bility Awareness Consumption Security
Implementation characteristics

Access Medium
Ability for near-real-
time Interaction
Computation power
Context Awareness
Multi-tenancy
Logical proximity
Coverage
Power consumption

Highly related Somehow related

11
Some typical use cases
Features Use cases
Smart RAN-aware Context Augmented E-Health Autonomous Smart Video Caching
Cities Optimization Reality Vehicles Grid & Analysis
Bandwidth
Latency
Extensibility
Context
Awareness
Power
Consumption
Scalability
Privacy &
Security

Critical Depends
12
Work Flow

13
Recommendation for the use cases
Use Cases Recommendation
Smart Cities FC + MEC
RAN-aware Context Optimization MEC
Augmented Reality MEC/CC + FC
E-Health FC
Autonomous Vehicles FC + MEC
Smart Grid FC
Video Caching & Analysis MEC/CC

14
Open source projects available in the market
Regarding the projects…

16
Akraino Edge Stack
“ Fully integrated edge
infrastructure ”

“ Intend to develop solutions

and support of carrier, provider
and the IoT networks ”

17
BPs within the project
• Connected Vehicle Blueprint
• Edge Video Processing
• Edge Lightweight and IoT Blueprint
• Integrated Edge Cloud Blueprint
• Kubernetes-Native Infrastructure for Edge
• Micro-MEC
• Radio Edge Cloud
• StarlingX Far Edge Distributed Cloud
• Time-Critical Edge Compute
18
Airship

“Declarative, YAML-driven
deployment”

“The implementation of
Openstack on Kubernetes
(OOK)”

19
StarlingX
“A deployment-ready,
scalable and highly
reliable edge
infrastructure software
platform”

Ability for near- real-time interaction

Power Consumption Multi-tenancy

Coverage Computation Power

Logical Proximity Access

Context Awareness

FC MEC CC StarlingX

20
Evaluation of StarlingX
• China Unicom, together with Intel, 99Cloud build a new ME-IaaS (Mobile
Edge-Infrastructure as a Service) based on the StarlingX. [1]
• The approved Akraino blueprint that submitted by Tencent on connected
vehicle has StarlingX proposed with TARS.[2] StarlingX is also proposed to be
used in another blueprint submitted by WR on Far Edge Distributed Cloud. [3]
• China Mobile Suzhou Software has evaluated StarlingX for its edge and
cloud plan, and China Mobile Research Institute and Intel experimented
vCPE onboarding on top of ONAP with StarlingX.
• China Telecom Research Institute Guangzhou has evaluated StarlingX as a
candidate for its edge solution
[1] Chinese ver: https://mp.weixin.qq.com/s/dlOpeo1Le5HEYCiSt3yUxg
[2] https://wiki.akraino.org/display/AK/StarlingX+Far+Edge+Distributed+Cloud
[3] https://wiki.akraino.org/display/AK/Connected+Vehicle+Blueprint

21
CORD
“Manage their Central
Offices using
declarative modeling
languages for agile,
real-time
configuration of new
customer services”
Multi-tenancy

Power Consumption Ability for near- rea

Coverage Computation Power

Logical Proximity Access

Context Awareness

FC MEC CC CORD
22
VCO
“Successfully completed two
demos on residential,
enterprise and mobile services
”

Multi-tenancy

Power Consumption Ability for near- rea

Coverage Computation Power

Logical Proximity Access

Context Awareness

FC MEC CC VCO

23
EdgeX Foundry
“Loosely coupled
microservice framework
with device management
and various protocols
supported”

Multi-tenancy

Power Consumption Ability for near-

Coverage Computation Po

Logical Proximity Access

Context Awareness

FC MEC CC EdgeX

24
Open Edge “Open edge computing framework that provide
temporary offline, low-latency services, and include
remote synchronization, function computing, video
access pre-processing, AI inference, etc.”

Already support functions such as python 27, and

compatible with Baidu CFC

Support both containerized mode and normal process

mode

Multi-tenancy

Power Consumption Ability for near- real-time interaction

Coverage Computation Power

Logical Proximity Access Mediums

Context Awareness

FC MEC CC Open Edge 25

KubeEdge
“First Kubernetes Native Edge
Computing Platform”

“Small footprint(66M and ~30MB

needed for memory).”

“Easy to enable a mini-cloud at

the edge”
Multi-tenancy

Power Consumption Ability for near- real-time inte

Coverage Computation Power

Logical Proximity Access M

Context Awareness

FC MEC CC KubeEdge
26
IoFog
“Deploying, running, and networking
distributed microservices at the edge”

“Construct an Edge Compute Network

(ECN) with Agent, Controller and
Connector”

“Need to write microservices for one’s own

purpose”
Multi-tenancy

Power Consumption Ability for near- real-time inte

Coverage Computation Power

Logical Proximity Access M

Context Awareness

FC MEC CC ioFog
27
Glimpse of Edge Projects
Project Foundatio Key Layer Segment/Focus MANO SDN Latest Infra Code Repo
n Participators version
Akraino Linux AT&T, Intel, ARM, Umbrel All-in-one edge N/A N/A N/A Openstack, http://gerrit.akraino.or
Foundation Nokia, Ericsson, la, Full stack K8S g
Dell, Red Hat, Stack
Juniper, WRS, etc.
StarlingX OpenStack Wind River, Intel, IaaS Industrial IoT ONAP ODL 1.0 OpenStack https://git.starlingx.io/
Foundation Huawei, Ericsson, and MEC cgit
China Unicom,
etc.
Airship OpenStack AT&T, SKT, Intel, Deploy Openstack on ONAP/Tack Calico 0.1 OpenStack/K https://git.airshipit.org
Foundation Mirantis, etc. ment Kubernetes er 8S /cgit

CORD Linux AT&T, SK Telecom, IaaS MEC for XOS ONOS 6.0 OpenStack/K https://github.com/op
Foundation Verizon, China residential, 8S encord
Unicom and NTT, enterprise &
etc. mobile
vCO Linux Red Hat, China IaaS MEC for ONAP/Tack ODL 2.0/3.0 OpenStack No code repo yet. Just
Foundation Mobile, etc. residential, er POC
enterprise &
mobile

28
 Glimpse of Edge Projects
Project Foundation Key Scope Layer Segment Latest Code Repo
Participators /Focus version
EdgeX Linux Dell, Vmware, Common framework for Edge PaaS Industrial 3.0 (4.0 Go:  
Foundry Foundation etc. solutions (SDK). IoT expected https://github.com/edgexfoundry/ed
in April gex-go
2019) Java:
https://github.com/edgexfoundry
OpenEdge N/A Baidu, etc. Open edge computing PaaS 0.1.2 https://github.com/baidu/openedge
framework
KubeEdge CNCF, Linux Huawei, etc Extend native containerized PaaS 0.2 https://github.com/kubeedge/kubee
Foundation application orchestration dge
capabilities at Edge
Azure IoT N/A Microsoft Internet of Things (IoT) service PaaS IoT 1.0.8-dev https://github.com/Azure/iotedge
Edge that offload task to edge
ioFog Eclipse Edgeworx, Edge computing platform PaaS IoT 2.0/3.0 https://github.com/ioFog/iofog.org
Foundation etc. through microservice at edge
Eclipse Eclipse Eurotech, Rad Platform for building IoT PaaS IoT 4.0 https://github.com/eclipse/kura/
Kura Foundation Hat, gateways, enabling remote
Comtrade, management & app
etc. deployment

29
Summary
What’s the best for your edge?

• EC implementations have difference on characteristics

• Use case have their unique features and suitable implementation
• Projects could be categorized into different implementations

Find the features of your USE CASE and choose the

most appropriate implementation for that!

31
Thank You
Q&A