Industry
Commercial Automotive
Offering
Big Data Analytics on Cloud
Workload
Track and trace application
Cloud
AWS
Project scope
— Design and deployment of cost effective analytics solution on cloud
About the client
The client is $4 billion automotive giant who has pioneered the manufacture of commercial vehicles — buses and trucks — in India. They have manufacturing bases in India, the Middle-East and the UK, and a footprint across 50 nations.
They have stayed ahead of the game by leveraging technology to power their innovations. including the use of statistical performance data to improve product design and deploying intelligent logistics solutions to aid sales managers, fleet managers and dealers across the country.
Business Challenge
Sales Admins, Fleet managers and dealer across the country needed up to date information about the vehicles they managed, and the enterprise was using bleeding edge technologies in Big Data to handle this problem. But the sheer volume (substantiated with these number 44 X 6 etc.) and the expedited growth meant that their existing IaaS solution hosted on current infrastructure was getting prohibitively expensive.
Each truck was fitted with an IoT device that would collect and stream data on 44 metrics such as speed, RPM, heat generated, location, gear, etc. every 6 seconds, resulting in the creation of 633,600 records per day. With the help of their partner, the client had deployed this solution on a fleet of virtual machines (VMs) and a data warehouse hosted on Google Cloud (GCP).
The problem arose when the client wanted to scale up to 200,000 vehicles by the end of 2019. With additional load the client had to invest in additional licenses, VMs, storage, manpower and other operational expenses (monitoring, backups …). With all these overheads, it meant that the project was becoming prohibitively expensive even on cloud.
The client was desperately seeking a partner who could help them find a solution that would help them leverage Big Data while keeping costs and management overheads to a minimum.
Solution
As a multi-cloud transformation enabler, 1CloudHub understood the client’s unique dilemma.
In order to help the client make the most of their data, we built for them a solution optimised for cost efficiency. We designed a serverless solution comprising of the following Platform-as-a-Service (PaaS) offerings:
- Elastic MapReduce (EMR), a managed Hadoop cluster capable of running both Spark jobs and HBase.
- S3 for massive data storage at low cost.
- Elastic Beanstalk managed Tomcat environment to run Java app.
- Redshift Columnar Data Warehouse to store data for reports.
- We leveraged machine learning capabilities on SageMaker for predictive analytics in the future.
Our approach
Design
Pilot
Scale
01. Design
- In collaboration with AWS, we designed and architected the solution.
- We provided the necessary technical expertise on AWS to help the application team adapt PaaS and follow best practices.
02. Pilot
- We recommended a low-risk 3-month pilot program for a few selected vehicles.
- We effectively ran and managed the service on AWS powered by our 24/7 monitoring and managed services.
03. Scale
- Once the pilot was successful, we recommended that our client add new vehicles each month, in batches of a few thousand, to the AWS platform.
- In addition, we recommended that the client switch existing workloads also in small batches.
Outcomes
A guaranteed a 99.99% uptime with a commercially backed SLA from AWS.
Reduced cost of the solution by nearly 20%.
Reduced management overheads significantly.
No licensing costs.
Scale on demand.
Looking forward
We are delighted to have helped the client deploy their futuristic, intelligent logistics system at optimal costs. We look forward to working with them to scale these up across their production line.
