JCON 2021

MicroStream is a fundamentally new persistence framework for storing complex Java object graphs natively, which means the object graph is stored on disk as it is in RAM. Vice versa it allows you to load any single subgraph on-demand that is then merged with your object graph in ram automatically. This leads to fantastic benefits: You can use object graphs as an in-memory database. An object graph is a powerful multi-model data structure. You can use any Java type, collections, and formats like JSON. Matching this, Java provides the perfect query language. With Java Streams, searching even huge object graphs will take only microseconds - up to 1000x faster than comparable SQL queries and even much faster than reading data directly from a local cache. DB-specific data structures or formats become meaningless for Java. No more expensive mappings. No more data-type conversion. No more DB-specific query language. Only one data model: Java classes (POJOs). No additional caching framework required. No more inconvenient object copies. Simple architecture and type-safe query code. Core-Java only instead of DB-specific concepts. As an end result, you will get out an ultra-fast Java in-memory database app or microservice. MicroStream is in productive use for more than 5 years and becomes now open source and part of the popular microservice frameworks Helidon and Open Liberty.