IoT Scale Event-Stream Processing for Connected Fleet at Penske

IoT Scale Event-Stream Processing at
Penske
Shruti Modi - Sr. Manager, Penske
Krishna Gogineni - Sr. Platform Architect, VMware
September, 2020

Copyright © 3/24/2020 Penske. All rights Reserved. Page 2
We're Here For You

How to Connect with your entire ﬂeet of trucks in real time to React and Respond to data
being collected and ﬁx problems before they occur.
Also, how do you monetize the capability?
Connected Fleet at Penske

• Business outcomes for Connected Fleet
• Connected Fleet Solution Architecture
• Implementation - Application & Platform concerns
• Challenges we faced & How we addressed them?
• Platform Highlights
• Summary
Agenda

Connected Fleet Business Outcomes
🌎 Run Remote Diagnostics on any truck anywhere in the country
🚦 Get greater insight into fleet performance
👩🏽‍💻 Make data-driven fleet management decisions
🛠 Access tools that put data at your fingertips
📋Help drivers manage required paperwork
🏛 Ensure compliance with all laws and government regulation

Solution Architecture tenets
Application
• Spring Boot as De-Facto standard
• Microservices patterns support
• Ability to Auto-scale the Application
instances
• Automated Path to Prod
• Integrated Conﬁguration & Credential
management
Platform
• Developer Self-service capability
• Zero downtime upgrades during business
hours
• Maximize Memory and Compute utilization
• Increased efﬁciency & consistency through
automation
• Comply with Penske Security and Audit needs

Copyright © 3/24/2020 Penske. All rights Reserved. Page 7 7
Connected Fleet Architecture
Tanzu Application Service
Poller Apps
Spring Cloud Data Flow
Streams
Microservices
Business Users
Real Time
Analytics
VMware Tanzu
Postgres
Observability as a Service
VMware Tanzu
Gemﬁre
Full Fleet of
Trucks
2 Billion Pings/Day
Tanzu MySQL
850 App instances with 50 Streams ingesting 700GB Data per day

Copyright © 3/24/2020 Penske. All rights Reserved. Page 8 8
Sample flow
Streaming Event Processing
Provider Topic
Truck
Sources
(API)
Event Mapper
Provider agnostic
Topic
Process
&
Normalize
Data Routes
Domain
Topic
Normalized
Data Topic
Domain
Topic
Domain
Topic
Domain
Topic

Applications - Challenges & How we addressed?
Problem
• Implement an end to end Data ingestion pipeline
leveraging the common Developer frameworks
without having to learn altogether new framework
Solution
• With Spring Boot and Data Flow, were able to
create the Data pipelines in a short time. Without
Spring, it would have elongated the Dev cycles.
• Poller & Streams Apps need to be HA to avoid data
processing backlogs
• Externalizing the configuration across many
applications and securing Credentials for Apps
backing Services
• Configured apps to auto-scale on Tanzu Platform
with 4 levels of HA
• Used Tanzu Platform:
• CredHub for Credential management
• Spring Cloud Config Service for
Configuration management

Platform Highlights
Stability
● Zero Downtime Upgrades
during business hours
● Upgrade, Patch and
Repave are handled by
Platform Automation
● Automation of Platform
rollout ensures no
snowﬂakes
Security
● All credentials are managed
using Platform Vault.
● 500+ vulnerabilities have
been patched so far this
year
● OS, Runtime dependencies
are patched to 100%
compliance
Scalability
● Multiple teams on boarded
to the Platform in minutes.
● On-demand data ingestion
scaling within minutes.
● Reduced cost for supporting
multiple types of workloads

Platform & Data Services
• Docker containers hosting for both Stateful & Stateless workloads
• Automated deployment of Kafka components
• Automated lifecycle of MySQL for Microservices and Domain based designs
• Automate Backup, Restore, Upgrade and HA
Provide users with platform add-on services to Enhance, Secure, and Manage applications/services.
• Update of Control Plane with Terraform
• Support CI/CD for DevOps teams
• Postgres Implementation for Geo-Spatial transactional data
• Postgres + Timescale for Geo temporal use cases

Platform Foundation
Stability &
Resiliency
Core building blocks to ensure Applications, Services and Data are on a Secure & Reliable Platform
Platform, App Health & Performance Monitoring:
• Metrics ingestion, Dashboards & AlertsObservability
Platform, App Logging:
• ELK Stack
Applications Distributed Tracing:
• Spring Cloud Sleuth, Open Tracing
Security
Security and Access Control:
• Identify and bridge security gaps
• Educate current and existing members
• Evaluation of App for Stability and Resiliency
• Explore HA options (multi-layer HA, platform and App), DR

Community of Practice
Application
Organize & share knowledge to Resolve issues, Improve skills and Learn from each others experience
• Establish Platform SLA and SLO
• Create Architectural/Code standards checklist
• Use Slack Channel to answer questions
• Educate customers on SLA and SLO
• Foster Adoption and Education of platform usage
• Identify key themes for brown bag sessions
Education
& Adoption
Platform

Summary
• Spring Boot/Cloud + Tanzu Platform = IoT scale Stream processing
• Automation is the key ingredient of our Success
• Automated Application Path to Prod(CI/CD).
• Automated Platform Rollout, Upgrade and Patch
• Automated Scalability
• Security has been a Key tenet in the design and implementation of Apps and
Platform
• Setting up paved ways through common patterns has accelerated outcomes

We are Hiring!!
https://penske.jobs/information-technology-jobs/new-jobs/

IoT Scale Event-Stream Processing for Connected Fleet at Penske

IoT Scale Event-Stream Processing for Connected Fleet at Penske

Recommended

More Related Content

What's hot (20)

Similar to IoT Scale Event-Stream Processing for Connected Fleet at Penske (20)

More from VMware Tanzu (20)

Recently uploaded (20)

IoT Scale Event-Stream Processing for Connected Fleet at Penske