liveBook · Manning

gRPC

As explained in Designing Deep Learning Systems: A guide for software engineers

An open-source remote procedure call framework that enables communication between distributed services.

gRPC is a remote procedure call framework for communicating between distributed services. Its concise service definitions make it practical for exposing APIs without introducing unnecessary framework detail. section 2.2.3 section 3.1.1

A service declares callable methods with request and response types. A dataset management API uses CreateDataset, UpdateDataset, GetDatasetSummary, and ListDatasets, with the last method returning a stream of summaries. section 2.2.3

A production-oriented gRPC workflow can separate file transfer from service communication:

  • Upload data to cloud object storage and obtain a downloadable URL. section 2.2.3
  • Send the URL and dataset metadata in a CreateDataset request. section 2.2.3
  • Let the service download, transform, store, and summarize the data before returning a response. section 2.2.3

Dataset ingestion API

rpc CreateDataset (CreateDatasetRequest) returns (DatasetSummary); rpc UpdateDataset (CreateCommitRequest) returns (DatasetSummary); rpc GetDatasetSummary (DatasetPointer) returns (DatasetSummary); rpc ListDatasets (ListQueryOptions) returns (stream DatasetSummary).

How Designing Deep Learning Systems: A guide for software engineers covers it

How deep it goes
Defines RPC methods and traces dataset creation, but stops short of explaining gRPC internals or framework configuration.
Examples
Uses dataset creation, updates, summaries, listing streams, MinIO storage, and downloadable data URLs.
What it assumes
Assumes basic distributed-service and network communication concepts while demonstrating API design through dataset operations.
What you can do after
Apply gRPC by defining typed service methods for distributed APIs and returning structured responses or streams.
How it connects
Connects remote procedure calls with RESTful interfaces, HTTP services, cloud object storage, and shared-resource model training.

Also explained in

Compare how 10 books explain gRPC →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Designing Deep Learning Systems: A guide for software engineers. The section links point to the parts of the book it draws on.