remote::infinispan

Description

Infinispan is a remote vector database provider for Llama Stack. It allows you to store and query vectors in a distributed Infinispan cluster via HTTP REST API. Infinispan provides high-performance, scalable data storage with support for both vector similarity search and full-text search capabilities.

Features

• Vector Similarity Search - Store and query embedding vectors with cosine similarity
• Full-text/Keyword Search - Query documents using Infinispan Query DSL or Ickle queries
• Hybrid Search - Combine vector and keyword search with configurable reranking (RRF or weighted)
• Authentication - Supports both Basic and Digest authentication mechanisms
• HTTPS/TLS Support - Secure connections with SSL certificate verification
• Distributed Storage - Leverage Infinispan's distributed caching for scalability
• HTTP REST API - Simple integration using standard HTTP protocol

Search Modes

Supported:

• Vector Search (mode="vector"): Performs vector similarity search using embeddings
• Keyword Search (mode="keyword"): Full-text search using Infinispan Query DSL/Ickle
• Hybrid Search (mode="hybrid"): Combines vector and keyword search with configurable reranking

Configuration

Basic Configuration (HTTP)

vector_io:
  - provider_id: infinispan
    provider_type: remote::infinispan
    config:
      url: "http://localhost:11222"
      username: "admin"
      password: "password"
      auth_mechanism: "digest"
      persistence:
        backend: "kv_default"
        namespace: "vector_io::infinispan"

HTTPS Configuration with TLS

vector_io:
  - provider_id: infinispan
    provider_type: remote::infinispan
    config:
      url: "https://infinispan.example.com:11222"
      username: "admin"
      password: "password"
      use_https: true
      auth_mechanism: "basic"
      verify_tls: true
      persistence:
        backend: "kv_default"
        namespace: "vector_io::infinispan"

Environment Variables

You can use environment variables for sensitive configuration:

vector_io:
  - provider_id: infinispan
    provider_type: remote::infinispan
    config:
      url: "${env.INFINISPAN_URL}"
      username: "${env.INFINISPAN_USERNAME}"
      password: "${env.INFINISPAN_PASSWORD}"
      persistence:
        backend: "kv_default"
        namespace: "vector_io::infinispan"

Usage

Starting Infinispan Server

The easiest way to get started is using Docker:

docker run -it -p 11222:11222 \
  -e USER="admin" \
  -e PASS="password" \
  infinispan/server:latest

Authentication

Infinispan supports two authentication mechanisms:

• Digest Authentication (recommended for HTTP): More secure than basic auth over HTTP
• Basic Authentication (HTTPS only): Simple username/password authentication

Set the auth_mechanism parameter to either "digest" or "basic".

Documentation

• Infinispan Documentation
• Infinispan REST API Guide
• Infinispan Vector Search
• Infinispan Query DSL

Requirements

• Infinispan Server 16.0+ (with vector search support)

Configuration

Field	Type	Required	Default	Description
url	HttpUrl	No	http://localhost:11222/	Infinispan server URL (e.g., http://localhost:11222)
username	str | None	No		Authentication username
password	SecretStr | None	No		Authentication password
use_https	bool	No	False	Enable HTTPS/TLS connection
auth_mechanism	str	No	digest	Authentication mechanism: 'digest' or 'basic'
verify_tls	bool	No	True	Verify TLS certificates for HTTPS connections (set to False only for development/testing with self-signed certificates)
persistence	KVStoreReference	No		Config for KV store backend
persistence.namespace	str	No		Key prefix for KVStore backends
persistence.backend	str	No		Name of backend from storage.backends

Sample Configuration

url: ${env.INFINISPAN_URL:=http://localhost:11222}
username: ${env.INFINISPAN_USERNAME:=admin}
password: ${env.INFINISPAN_PASSWORD:=}
use_https: false
auth_mechanism: digest
verify_tls: true
persistence:
  namespace: vector_io::infinispan
  backend: kv_default