Amazon Ditches Microservices for Monolith: Decoding Prime Video's Architectural Shift

1

Apache Kafka

When it comes to the limitations of AWS Step Functions, let us look at what it was doing. Step Functions handled communication between the different steps of their stream quality architecture and error handling. When it comes to communication between services, tools like Kafka exist and can be used to transfer data (or state) between services. Kafka uses a pub/sub (publish and subscribe) messaging model that allows producers to publish topics to consumers, that can then act on the topics they are subscribed to. Kafka's pub/sub model allows for efficient and reliable data streaming, making it perfect for building event-driven systems, such as one that handles monitoring video quality.

#Stream Processing #Data Integration #ETL 120 social mentions

2

AWS Step Functions

Amazon's tool was broken into three main components: the media converter, defect detectors, and orchestration. The media converter ran as an AWS Lambda function, converted audio and video streams, and stored the data in an S3 bucket. The defect detectors, also running as an AWS Lambda function, would pull the parsed data from the S3 bucket and analyze the frames and audio for any issues. Finally, the orchestration controlled the system's flow using Amazon's AWS Step Functions.

#Project Management #Workflow Automation #Web Service Automation 57 social mentions

3

Amazon S3

Amazon's tool was broken into three main components: the media converter, defect detectors, and orchestration. The media converter ran as an AWS Lambda function, converted audio and video streams, and stored the data in an S3 bucket. The defect detectors, also running as an AWS Lambda function, would pull the parsed data from the S3 bucket and analyze the frames and audio for any issues. Finally, the orchestration controlled the system's flow using Amazon's AWS Step Functions.

#Cloud Hosting #Object Storage #Cloud Storage 170 social mentions

4

AWS Lambda

Amazon's tool was broken into three main components: the media converter, defect detectors, and orchestration. The media converter ran as an AWS Lambda function, converted audio and video streams, and stored the data in an S3 bucket. The defect detectors, also running as an AWS Lambda function, would pull the parsed data from the S3 bucket and analyze the frames and audio for any issues. Finally, the orchestration controlled the system's flow using Amazon's AWS Step Functions.

#Cloud Computing #Cloud Hosting #Backend As A Service 243 social mentions

5

Amazon EC2

But here, that's a wasted benefit as these functions are executed multiple times per second, every second, managing a live video stream. The media converter and defect detectors could have likely been run on their own EC2 instances. They would have had their always-running, scalable servers, like AWS Lambda. Though the issue of storage between services would still remain.

#Cloud Computing #Cloud Infrastructure #VPS 62 social mentions

6

Amplication

Whether you build your next project as a monolith or with microservices, check out Amplication. Amplication helps developers build backend services quickly using the highest standards that are consistent and scalable.

#APIs #Open Source #Developer Tools 70 social mentions