-
Couldn't load subscription status.
- Fork 2.3k
Description
Is your feature request related to a problem? Please describe
Protobuf documents used in gRPC today are not fully capitalizing on gains from binary encoding. For both Search APIs and Bulk APIs, the gRPC/Protobuf APIs accepts/returns a JSON document in a protobuf bytes field, which does not benefit from Protobuf serialization.
Describe the solution you'd like
Support SMILE as a supported document format in the gRPC Bulk API and gRPC Search API.
Related component
Search:Performance
Describe alternatives you've considered
No response
Additional context
What is SMILE and why use it?
SMILE is a binary serialization of the generic JSON data model, which is already supported in the OpenSearch server as an XContent format.
Per SMILE gRPC benchmarking results (slides taken from OpenSearch Con NA 2025 - Advancing OpenSearch With gRPC and Protobuf at Uber), SMILE can reduce both gRPC search and gRPC indexing latency, reduce the payload size across the wire, as well as reduce client-side decoding latency.
- Search:
- Bulk:
- Client-side decoding and payload size reduction:

Metadata
Metadata
Assignees
Labels
Type
Projects
Status
Status