This presentation discusses the challenges of cloud computing at the edge. From the exploding number of nodes, the need for integrated monitoring and zero touch discovery. We introduce Lenovo Open Cloud Automation, an automated framework built in collaboration with Red Hat to help address these challenges.

1. Automated Deployment and
Management of Edge Clouds
Jay S. Bryant – Cloud Architect – Lenovo
Chadie Ghadie – Director of Telco Business Development – Lenovo
Gareth Jenkins – Senior Solutions Architect – Red Hat

2. 2
Agenda:
• Challenges of Edge Computing
• Red Hat Vision for Open Hybrid Edge
• Introduction to Lenovo Open Cloud Automation
• Lenovo Open Cloud Automation – VCO 3.0 Demonstration

3. Challenges of Edge Computing

4. 4
Key Points about Edge Computing
▪ Edge has multiple tiers and often requires different HW and Software
solutions for each.
▪ Edge platforms must be connected end to end.
▪ The closer to the end user, each of the following become more challenging:
▪ Performance per Watt
▪ Form Factor
▪ Thermals
▪ and Security (less visibility, monitoring)
▪ Different Edge locations run different services.
▪ The need for the right KPIs (data privacy, BW, latency, type of backhaul
connectivity, security, etc.)

5. 5
Multi-Tier Requirements
< 1ms
Small Cell
On-premi
se
Equipmen
t or
Cloudlets
Multiple Points of
Aggregation
Cloud/OTTCell Tower
Core
TIERS
(~Size / country)
NETWORK
LATENCIES
(Wire Round trip)
< 1ms
1-5ms
5ms + 1-2
ms (every
100kms)
5ms + 5
ms (every
400kms)
> 60 ms
RESTRICTIONS
Use cases
Smart Cities
V2V
Retail
Video
Analytics
WHERE, WHAT & WHY
Use case
AR/VR
Retail
RT Streaming
Healthcare
Use case
Smart Cities,
V2V
Video Analytics
Drone/IoT
Rural
Use case
Smart Cities,
Video Analytics
Drone/IoT
Healthcare
CDN & Storage
FaaS
AR/VR/MR
Use case
CDN
5GC
Storage GW
HOW
(per Territory = country
for a big operator)
1^6x
units
1Kx units
100Kx
units 1Kx units 10x units
Challenge: Optimal platform and CPU system architecture is required to provide the right edge solution while optimizing TCO at scale
and preserving security and scalability
The close to the user the more restrictive the power consumption, thermals, form factor, security, management

6. 6
Infrastructure
▪ More and more hardware is moving outside the
datacenter:
▪ Needs to be managed, monitored, scaled-out, etc.
▪ Ability to Troubleshoot and auto-heal
▪ Networking isn’t just switches anymore
▪ As compute footprint grows, so do the network complexities
▪ Growing software defined complexities
▪ Multiple clouds must be managed across multiple sites
▪ SLAs must be met for multiple users
▪ Consistent performance must be achieved
▪ Problems must be quickly discovered and isolated
▪ Multiple standards with many different requirements

7. 7
7
Workloads
▪ Diverse workloads must be accommodated
▪ Database
▪ NFV
▪ Edge
▪ Multiple environments
▪ Baremetal
▪ Virtual Machines
▪ Containers
▪ Storage demands also grow with the increased workloads

8. 8
Operators Are Challenged
▪ Initial deployment is difficult
▪ Careful planning is required
▪ Many different tools must be employed
▪ Future needs must be considered
▪ Monitoring is difficult
▪ Many different tools with different jobs
▪ Different hardware vendors provide different support
▪ Expansion is difficult
▪ Adding new hardware requires many different steps

9. 9
Integration is Key
▪ Integration between hardware and …
▪ Inventory
▪ Deployment Automation
▪ Software
▪ Management
▪ Monitoring
▪ Security
▪ Workloads
▪ Integration must come to the user out of the box

10. Enabling the Open Hybrid Edge

11. TECHNOLOGY
SCALE
PEOPLE & EXPERTIZE
PROCESSES & DATA
ENVIRONMENTAL (DIL)
SECURITY
COST
LATENCY
BANDWIDTH
RESILIENCE
SECURITY
SOVEREIGNTY
INTEROPERABILITY
COST
WHAT IS EDGE COMPUTING
11
...
MOTIVATION CHALLENGES
AGGREGATE DISTRIBUTE
Edge Computing

12. Field/Remote
Edge
Regional
Data Center
Infrastructure
Edge
Provider
Device
Edge
Provider
Access
Edge
Provider
Aggregation
Edge
Core
Data Center
EDGE COMPUTING
12
Edge Tiers
End-User Premises Edge
(i.e. Far Edge / Fog Edge)
Provider Edge / Regional Edge
(i.e. Near Edge)
Provider/Enterprise Core
“last
mile”
...
Device or
Sensor
WHAT IS THE EDGE
Centralize where you can, distribute where you must.

13. 13
Who is doing edge computing?
Telco 5G EdgeEnterprise Automotive / Healthcare
(AI/ML)
Industrial / Factory
(IoT)
Public Sector
Radio Access Network
(vRAN)
(Retail) Remote Office
Connected Sensors
and Controllers
Video Surveillance Trading Sensitive Apps
In-field Operations Content Delivery Networks
Universal Customer Premise
Equipment
Connected Car, Train, Plane
Healthcare Monitoring and
Data Processing
and more...
EDGE COMPUTING WITH RED HAT

14. ENTERPRISE
EDGE
PROVIDER
EDGE
REGIONAL
EDGE
NATIONAL
CORE
REMOTE
EDGE
14
APP APP APP APP
PHYSICAL VIRTUAL
PRIVATE
CLOUD
PUBLIC
CLOUD
Extending the Open Hybrid Cloud Vision with Edge Computing
Any workload
THE OPEN EDGE COMPUTING VISION
, any footprint, any location.

15. 15
Edge computing
Telco
One focus
Enterprise Public sector
IoT Healthcare
Market segments
Various use cases and industries
Artificial intelligence
/Machine learning
Edge-enabled Product Portfolio
Red Hat portfolio approach
Single, edge-enabled Portfolio
delivering consistency for both
Development and Operations
EDGE COMPUTING WITH RED HAT
● Delivering Edge capabilities into current portfolio.
● Enabling broad portfolio to run natively at the edge.
OS - Virtualization - IaaS - Cloud - x86 - ARM - Middleware - Containers
Automation - Security - 5G - Hyperconverged - Storage - Dev/Ops - etc.

16. Common requirements
Zero-touch provisioning
Multi-cluster management
Small footprint
Scale
Security (physical, SW)
Distributed storage
Middleware enablement (IoT)
Network disruption tolerance
Automated operations
16
Telco 5G
Remote Office
Industrial IoTmore ...
more ...
PARTNER
ECOSYSTEM
EDGE COMPUTING WITH RED HAT

17. Red Hat’s Vision for Open Hybrid Edge
● Delivering a single, edge-enabled platform for any workload,
any footprint and any location by extending our vision of
Open Hybrid Cloud all the way out to the Edge.

18. LOC Automation

19. 19
Challenges with cloud implementation
• Takes several weeks to
onboard new customers
• Takes hours to set up a new
VM
• Requires expertise or need to
learn as you go
• Expensive and lengthy
professional services
engagement
• Each vendor only automates
their process, creating islands
• Lack overall visibility impacting
other vendors
• If one vendor fails, there are no
remedy actions across the
platform
Time Expertise Efficiency
Delays time to revenue
Shift to cloud first
technology – 30% growth
More than $1.3 trillion in IT
spending will be directly or
indirectly affected by the
shift to cloud
28% of cloud budgets:
consulting, implementation,
migration and managed
services
Industry
trends
by 2022
Cloud
challenges
Source: https://www.gartner.com/en/newsroom/press-releases/2019-04-02-gartner-forecasts-worldwide-public-cloud-revenue-to-g
https://www.gartner.com/en/newsroom/press-releases/2018-09-18-gartner-says-28-percent-of-spending-in-key-IT-segments-will-shift-to-the-cloud-by-2022
Impact

20. 20
Customized
cloud
Zero touch
deployment
Lifecycle
management
LOC Automation capabilities
20
Workload Component
Service Component
Run-Time Component
OpenShift Container VM
Platform Inventory
Discovery
Automation
Config & OS Deploy
Code Repository
SW Repos Mgmt
Validation/Benchmark
Single Pane of Glass
Workload Installer
Red Hat RHHI
RHV / RHVM / Gluster
VMWare
ESXi / vCenter
Bootstrap Tool Kit
Low touch
deployment
Bare
Metal

21. 21
CoSP use case
NB
AWX
Git
XL
ELK
LOC Automation
Data Center
Central Office
Admin focused with Edge management
Designed to manage lots of hardware
Cloud 1
vBNG
Central Office
Cloud 2
vBRAS
Tower
Edge Cloud
vRAN, vFW
Enterprise edge
Edge Cloud
SD-WAN
Centralized &
automated
deployment
Single pane of
glass management
and monitoring
Admin
benefits:
LOC Automated deployment
Telemetry

22. 22
Enterprise use case User focused
Self service: Easily deploy multiple clouds
LOC Automation
Tenant Instance 1
Tenant Instance 2
Tenant Instance 3
User 1
User 2
User 3
Customer
benefits:
Centralized &
automated
deployment
Single pane of glass
management and
monitoring
LOC Automated deployment
End user interaction
Telemetry
NB
AWX
Git
XL
ELK

23. 23
Our automation is based on the following design principles:
End to end automation approach
1. Planning
Network -> IPs, Hostnames,
VLANs, ASNs, Domainnames,
DNS, physical connections
Servers -> UEFI/Adaptor settings,
Performance tuning
Cluster -> number of nodes, tenant
assignment
Tenants -> Tenant name, cluster
assignment
Sites, Rack, physical infrastructure
2. Configuration
Network -> vlans on switches, IPs
in DNS, networks in IPAM
Servers -> initial system
configuration, firmware compliance,
tenant assignment, monitoring
3. Installation
Servers -> OS and platform
installation for dedicated tenant
with already planned parameters
4. Lifecycle
Firmware on hardware, rolling
upgrades
Patches for OS, rolling upgrades
Patches for Red Hat Platforms and
Tools
Meta Database/NetBox
Lenovo integrated meta DB to
provide all required information –
„source of truth“
Ansible / Tower / AWX
Lenovo xClarity/Confluent
connects to open APIs of switches
and servers for device discovery
and management
Lenovo xClarity
provides HW configuration
management and OS deployment
through Ansible playbooks
Red Hat Platforms /
VMware Lifecycle Manager
manages the lifecycle of Red Hat
platforms in combination with
NetBox, Satellite, XClarity and
Ansible playbooks
End to end automation

24. 24
One runtime service deploys other services as needed
LOC Automation: software services view
Platform Management Services
Device and Platform
Inventory
Device Discovery
Configuration and
Automation Repository
Software Repository and
Lifecycle Management
Device Configuration
and OS Deployment
Automation and Task
Execution
Validation and
Benchmarking
Single Pane of Glass
Storage Management Services
Cloud Management Services
Cloud Deployment
Cloud Operations
Storage Management
Runtime service

25. 25
Lenovo Open Cloud Automation edge support
• Edge Compute Deployment Support
• Telco Edge Hardware Deployment
Support
• Future IoT Integration
LOC Enables
• Low Touch Deployments
• LOC is Provided Vital Product Data to
Identify Newly Attached Hardware
• LOC Prepares the Hardware for
Deployment from the Data Center
• Automated
• Inventory Population
• Firmware Update and OS Deployment
• Integration with Cloud Infrastructure
• Seamless Integration with Monitoring
Infrastructure

26. VCO 3.0

27. 27
27
Office becomes a mini data center
The shift to Virtual Central Office (VCO)
• Leverage Lenovo servers to run
VNFs, CNFs
• Gain server economies and
centralized management with
Lenovo Open Cloud
Demanding workloads and mobile usage
require higher speeds and lower latency
VCO Solution for multi-vendor NFV
deployments at the distributed cloud
edge
Designed for use with residential,
enterprise and mobile VCO services
Powerful edge to scale VCO to demand
while offering high performance
CoSP Requirements
Office
Mini Data
Center
NGCO

28. 28
28
Software defined
Open Source
Network Functions
Virtualization
Lenovo working to deliver the Virtual Central Office (VCO 3.0)
● VCO 3.0 is a multi-vendor Open Source Initiative from the Linux Foundation & OPNFV Community for Next-Gen, 5G Cloud Native
Telco Networks and Edge NFV deployments. Solution brief for VCO 3.0 5G Demo in containers.
The Innovation Lab is:
• A Virtual Central Office and PoP
become datacenter @network edge
• Market validation for containerized
microservices and applications
• Proving the readiness of virtualization
for CoSPs around the globe

29. 29
LOC integrates into DevOps and productivity models needed for digital transformation
Benefits of the Lenovo Open Cloud model for CoSPs
Benefits using LOC for CoSPs
• DevOps – increase productivity
• Security – use certified containers and
VMs for scheduled deployments
• Scalability – LOC provides agile
automation
• Efficiency – the ideal resource to
maximize flexibility for multi-type
infrastructures
• Datacenter to edge – seamless
management of HW & infrastructure
across the network
Linux Foundation Innovation Lab VCO 3.0
KALOOM NETWORK FABRIC
LENOVO OPEN CLOUD
THINKSYSTEM SR630/650
RED HAT OPENSHIFT CONTAINER PLATFORM
Containers
SHARED MANAGEMENT TOOLING (VMs)
Red Hat
OpenStack
Platform director
Red Hat Ansible
Tower
RED HAT VIRTUALIZATION x3
Red Hat
CloudForms
Compute
KVM
VMVM
Storage
GLUSTER
Red Hat
Satellite
Red Hat
OpenStack
Controllers x3
Lenovo xClarity
HW Management
RHV-M
cNF cNFcNF UPF-C