4. 4
Brown to Green
A Open Platform for Evolutionary
Architecture - Connecting a
Secure Edge to Intelligent Apps
REDHATOPENENERGY
John Archer
Principal Cloud Platform Solution Architect
Red Hat - Energy
archer@redhat.com
5. Where Innovation is Happening
5
Source: Red Hat success story. “BP modernizes infrastructure, introduces DevOps with self-service platform” accessed May 2019.
Source: ExxonMobil - Red Hat Summit keynote video from Summit 2019
The combination of microservices, containers, and a fully automated
CI/CD platform provides what developers have been asking for for
years. They now have full self-service to deliver change from the
initial idea, through the innovation, right through to production, as
quickly as humanly possible.
Paul Costall
Head of Application Engineering Services, BP
8. SECURE EDGE CHALLENGES
8
● Cultural
● Insecure
○ Most deployed edge solutions today are at risk
● Data ownership is mixed
● Sunset of 2.5G/3G moving to LTE-M and 5G
● Costly
○ Are metered by tags by traditional SCADA vendors
● Intelligent backhaul still difficult
9. DATA SCIENCE CHALLENGES
9
● Cloud GPUs at needed scale still expensive
● Many efforts stuck on desktops
● Data
○ Can not find curated and authoritative data
○ What they can find is random quality, latency, complete
○ Lacking metadata standards and pedigree
● Most teams still maturing their ML/AL/DL/RL
● Very crowded and fast moving space
● No one wants to be called a Citizen Data Scientist
10. 10
KEY FUNCTIONALITY
FOR AN END-TO-END ARCHITECTURE
Securely connect, authenticate
and manage disparate
connected devices that speak
different protocols
Apply analytics at the edge with
machine learning and business rules
to enable local, low-latency decision
making
Centralize IoT data processing,
analytics and machine learning
to enable deep business insights and
actionable intelligence
Enable integration with enterprise and
business applications to bridge the
gap between OT and IT
and reduce complexity
Tools to enable end-to-end data security, compliance, authorization and authentication
Device Management
& Connectivity
Intelligent Edge Processing
& Analytics
Advanced Analytics
& Machine Learning
Business & Application
Integration
End-to-End Security & Compliance
11. RED HAT
OPEN INNOVATION LABS
1
1
BRIDGING THE SILOS
Accelerating innovation
CONTEMPORARY
NON-LINEAR
TRADITIONAL
LINEAR
12. CAPABILITIES ENGINEERED TO WORK TOGETHER
CODE
CONTAINER NATIVE PLATFORM
MESSAGING
API
RULES
BUSINESS
AUTOMATION
FUNCTION
AS A SVC
13. ● Measured Boot
● Secure Boot
● Full End-to-End Disk Encryption
● Signed Containers
● Bashless OS
● OS and Firmware Over-the-Air Style updates
WHAT MAKES UP A NEXTGEN SECURE EDGE OS?
13
14. SECURE EDGE / INTEGRATION HUB
MQTT
AMQP
1.0
Edge
Private or
Public Cloud
HTTP
LoRaWAN
CoAP
Enterprise &
AI/ ML
Applications
Device Registry
Telemetry
Commands
16. YOUR DIFFERENTIATION DEPENDS ON YOUR
ABILITY TO DELIVER INTELLIGENT APPS FASTER
CONTAINERS, KUBERNETES, DEVOPS & DATAOPS ARE KEY INGREDIENTS
Innovation
Culture
Cloud-native
Applications
AI & Machine
Learning
Internet of
Things
Virtual GPU
17. CONNECTING THE EDGE TO DATA SCIENTISTS
Highly Scalable,
flexible, elastic,
microservice based
architecture
Fully Portable – On
Premise to any
public cloud vendor
Leverages the
power and agility
of open source
software without
lock-in
Architecture
Tenets
Data
Scientist
Data
Manager
s
Citizen
Data
Scientist
Cognitive AI
Vision
Speech
Face
Audio
Video
Text
Data
Models
Curation
Prep
Quality
Publishing
SecurityPython, R, Jupyter.org, Tensorflow, Keras, Pandas, Bokeh, Dash, Prometheus,
Grafana, SciPy, NumPy, SumPy, Julia , Spark, PySpark, Theano, Scikit, FaceDetect
Packages:
AI/ML/Data Science Pods
MongoDB, MariaDB, mySQL, Postgres, Couchbase, Redis, MS-SQL, OraclePersistence
:
SSOandAuthentication
OIDC
SAML
OAuth
JWT
Kerberos
DevOps
Node.js, .Net Core, Java, Python, PHP, Ruby, Rails, Javascript, PerlApp Dev:
AppDev & App Services and Persistence Pods
REST
ODBC
JDBC
WS
Predictive
Maintenance
Autonomous
Operations
Supply Chain
Improvements
Downstream
Reliability
Use Cases
Multitenant – CPU
and GPU powered
workloads
REST
IoT “Things”
MQTT
Integration, BPM, Rules, Messaging, API, IoT, Microservices, IstioApp Services:
OnPremise Public Cloud
WSS
Kafka
18. DATA SCIENTIST DEVELOPERS NEEDS
All Developers need
● Choice of architectures
● Choice of programming languages
● Choice of databases and persistence
● Choice of application services
● Choice of development tools
● Choice of build and deploy workflows
Data Science Additional Needs
● Access to GPUs and varied storage
● Access to Curated Data
● Automated ScienceOps pipelines
● Collaboration with the Business
● Access to specific data science
languages and toolsets
They don’t want to have to deal with the infrastructure.
19. MATURING INTO A DATASCIENCEOPS PIPELINE
Seeing an emerging notion of Data ScienceOps workflows.
Data Prep includes: Ingest, ETL, Pedigree, MetaData, Quality
Polyglot Models for ML Training, Data Streaming and Data at Rest concerns
Dynamic behaviors depending on compute/GPU/Storage/Memory resources
20. MACHINE LEARNING ON OPENSHIFT
Unique performance computing requirements for
Artificial Intelligence, Machine Learning, Neural
Networks and GPUs
Multiple Data Science images:
• TensorFlow
• Pyro/PyTorch
• Scikit-learn
• CNN/GANs
• Keras
• Seldon
• RAPIDS.AI
• Apache Arrow
21.
22. MULTI CLOUD OBJECT GATEWAY
22
App Multi-Cloud Buckets
Multi-site Buckets
S3 APIApp
App
Hybrid Buckets
DEPLOYAND MANAGE DATA SERVICES
23. EFFICIENCY AND SECURITY BY DEFAULT
S3 Write Fragment Dedupe Encrypt StoreCompress
Paris DC
London DC
New York DC
24. HIGH LEVEL ARCHITECTURE
24
HIGH LEVEL ARCHITECTUREApplication
meta-data
NooBaa Core
Optimized
utilization,
resilience,
performance,
locality,
economics,
etc.
Heartbeats
Instructions
Instructions
Scalable NooBaa Storage Node
stores chunks,
runs lambda functions,
monitors host,
sends heartbeats
Scalable NooBaa Endpoint
S3 / Lambda API
auth,
chunking,
dedupe,
compress,
encrypt
Cloud Resource
AWS S3
AWS S3-compatible,
Azure Blob,
Google cloud storage
Data path
25. ● Fedora IoT
● JupyterHub on Openshift
○ Jupyter notebook, JupyterHub, JupyterLab, Openshift Templates
● Kubeflow
○ Kube project for Tensorflow, JupyterHub/Lab, PyTorch, MPI Operator
● Opendatahub.io
○ Ceph, Spark, JupyterHub/Lab, Tensorflow
○ Simplified Multiple Kernel notebook support
○ Simplified GPU Support
○ Resource management and instance culling
● RAPIDS.AI - Open GPU Data Science packages
○ CUDF, DASK, XGBoost, PyTorch
● Seldon.io - Machine Learning Models
OSS EDGE and DATA SCIENCE PROJECTS
26. ● Join Openshift Commons - Energy and ML SIGs https://commons.openshift.org/
● Openshift Self Service Education https://learn.openshift.com
● Install Minishift https://docs.okd.io/latest/minishift/getting-started/installing.html
○ MacOS - brew cask install minishift
○ Manual - https://github.com/minishift/minishift/releases
● Install Jupyter and JupyterHub Openshift templates
○ https://github.com/jupyter-on-openshift/jupyterhub-quickstart
● Review the OpenDataHub.io project
● Email me at archer@redhat.com
HOW CAN I GET STARTED?
30. KNATIVE
SERVERLESS BUILDING BLOCKS
Build
A pluggable model for
building artifacts, like jar
files, zips or containers
from source code.
Serving
An event-driven model
that serves the container
with your application and
can "scale to zero".
Eventing
Common infrastructure for
consuming and producing
events that will stimulate
applications.
"...an extension to Kubernetes exposing building blocks to build modern, source-centric, and
container-based applications that can run anywhere".