Architecture
Building upon the architecture discussed in Chapter 5, Integration Pattern: Batch Metadata Extraction, for the first batch processing example, we will adopt a similar cloud-native, serverless approach on Google Cloud to handle the batch processing of client applications for summarization. This scalable setup leverages various services, allowing us to seamlessly integrate the AI model and store the generated summaries.
The architecture will consist of the following components:
- Object store (Google Cloud Storage): This highly scalable and durable object store will be used to store client applications, which can be in various formats, such as PDFs, Word documents, or structured data files.
- Messaging queue (Google Cloud Pub/Sub): A messaging queue will be employed to coordinate the data flow and manage the processing of client applications.
- Processing component (Google Cloud Functions): Cloud Functions will serve as the processing component, executing...