Caylent Team
Caylent Catalysts™
IoT
Connect, understand, and act on data from industrial devices at scale to improve uptime, efficiency, and reliability across manufacturing, energy, and utilities.
Discover how we helped a global conveyor belt manufacturer build a real-time IoT data platform for Conveyor Belt Cleaners using Amazon Kinesis.
Client Overview
A global conveyor belt manufacturer specializing in solutions that maximize efficiency, safety, and uptime for belt conveyor systems across industries like mining, logistics, and food processing. With over a century of experience, the company provides a comprehensive range of mechanical fasteners, belt cleaners, maintenance tools, and digital technologies that help customers reduce downtime and improve productivity.
Challenge
A Global Conveyor Belt Manufacturer sought to move beyond simple equipment maintenance and develop a comprehensive Conveyor Intelligence Solution.
- Data Ingestion Requirement: The core problem involved developing a new IoT platform, built around the Software AG Cumulocity Platform, that could ingest several metrics from approximately 500 deployed Conveyor Belt Cleaner devices.
- Holistic Data View: Once ingested, this streaming data needed to be merged with stateful ontological data within Cumulocity to produce a holistic view of each cleaner's current operational state.
- Machine Learning Inference: The data then had to be pushed through an AWS Account Environment to run against deployed inference points. These points utilize machine learning models developed by the Manufacturer's Data Science team to provide critical insights.
- Core Business Goal: The ultimate goal of these insights was to support the Manufacturer’s business objective of achieving a '1 Blade per Cleaner per Year' metric, optimizing usage and maintenance schedules.
Solution
Our team assisted the client in developing a new IoT device management solution, with the critical data pipeline powered by Amazon Kinesis.
- Real-Time Streaming Platform: The solution involved implementing a real-time streaming platform backed by the Amazon Kinesis Family. This platform was specifically designed to process unbounded data sets in real time.
- Processing Capabilities: The platform was architected to handle machine learning model inference, execute branch data operations, and store data as it exists in every possible variation.
Results
Our Amazon Kinesis-backed solution provided a robust, scalable foundation for the International Conveyor Belt Manufacturer’s future data and intelligence efforts.
- Accurate Real-Time Results: The platform was capable of flawlessly returning real-time processing and inference results back to the device management solution.
- Foundation for Discovery: The long-term data storage capability ensures the data is available for future use, supporting new model development, validation of current processing results, and general discovery of new patterns.
- Conveyor Intelligence: The overall IoT platform successfully enabled the International Conveyor Belt Manufacturer to transition its simple Belt Cleaning devices into a full conveyor intelligence solution.
Key Takeaways
- Amazon Kinesis Data Analytics as a Micro-Batcher: A major learning opportunity was discovered when leveraging the SQL API for Pyflink within Amazon Kinesis Data Analytics (KDA).
- Optimal Processing Pattern: The team found a highly effective pattern in using Amazon KDA to window data points and then pass the bundled data to AWS Lambda functions for the actual processing.
- Scalable Compute: This paradigm was particularly effective when significant compute was needed, allowing Amazon KDA to act as a micro-batcher for Lambda and providing an additional level of scalability for compute separate from the stream windowing process.
IoT
Data Modernization & Analytics
Caylent Team
View Caylent's articles
Related Blog Posts
Amazon Kinesis Data Streams On-demand vs. Provisioned Billing Mode Cost Comparison
Compare Amazon Kinesis Data Streams on-demand vs. provisioned pricing. See at what utilization levels provisioned mode saves money based on payload size and throughput.
Data Modernization & Analytics
How and When to Use Amazon EventBridge Pipes
Learn when Amazon EventBridge Pipes can replace simple AWS Lambda connector functions and when they fall short. Includes practical guidance on InputTemplates and data transformation.
Managed Services
IoT
Building a Simple AWS Data Warehouse Solution with Data Streaming
Build a serverless data warehouse on AWS by streaming Amazon DynamoDB data to Amazon S3 with AWS Lambda. A cost-effective architecture for historical analytics and business reporting.
Data Modernization & Analytics
Managed Services
IoT