Para

Para is a modular, Java-based backend framework that provides object persistence, search indexing and distributed caching for web, mobile and IoT applications. It exposes a RESTful JSON API and is designed for multitenant deployments and rapid prototyping.

Key Features

• Multitenancy: each application has its own storage, search index and cache isolation.
• RESTful JSON API with signed requests and stateless client authentication using JSON Web Tokens (JWT).
• Database-agnostic persistence layer with adapters for various data stores and lightweight support for embedded databases.
• Full-text search support via Lucene or Elasticsearch with automatic indexing of objects.
• Distributed and local caching options (Hazelcast or other cache implementations) for low-latency access.
• Pluggable security: LDAP/SAML/social login support, CSRF protections and resource-level permissions.
• Validation and robustness via JSR-303 / Hibernate Validator and optimistic locking support in DAOs.
• Modular architecture with dependency injection and plugin support, executable JAR with embedded servlet container, and metrics/monitoring integration.

Use Cases

• Rapidly build backend APIs for mobile and single-page web applications with per-app multitenancy.
• Add full-text search and indexing to application data using Lucene or Elasticsearch.
• Provide a unified backend for IoT or game clients that need scalable caching and flexible auth.

Limitations and Considerations

• JVM-centric: primarily a Java/Maven codebase and requires a Java runtime to run and build.
• Relies on external components for some functionality: search (Elasticsearch/Lucene) and distributed cache (Hazelcast) need to be provisioned and configured for production scaling.
• Advanced integrations and hardening (e.g., enterprise SSO, specific DB backends) require additional setup and testing per deployment.

Para is targeted at teams that want a tested, extensible backend framework rather than a managed SaaS API. It is suitable for prototyping and production deployments where a JVM-first, modular backend with search and caching is desirable.