1 Answer
- Newest
- Most votes
- Most comments
1
Hi,
there can be multiple options and patterns, one possibility would be to use an architecture based on the following components:
- Amazon Kinesis Stream: you stream your data to Kinesis, it can also be done using a API Gateway in front of it to capture your POST event
- Amazon Kinesis Firehose + Lambda to transform the data and make it available in S3 (the lambda after the transformation could also call a Sagemaker inference)
- If the analysis transformation to be done need to be more complex you could consider to use Flink on Kineses Data Analytics
- Amazon S3 to store the transformed data, you could also store the payload and have your CDN point to it
- Lambda to send back the payload , maybe with API Gateway to publish the APIs.
Two examples of these patterns can be seen in this reference architecture for Monitoring Streaming Data with Machine Learning and in this blog post on Real-Time In-Stream Inference with AWS Kinesis, SageMaker, & Apache Flink
hope this helps
Relevant content
- asked a year ago
- asked a month ago
- Accepted Answerasked 9 months ago
- asked 7 months ago
- AWS OFFICIALUpdated 8 months ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated 9 months ago