Delta

DDD-centric event-sourcing library for the JVM.

What's in a name?

del·ta

/ˈdeltə/

symbol symbol: δ; symbol: Δ; symbol: delta 1. MATHEMATICS variation of a variable or function. 2. MATHEMATICS a finite increment.

- or -

Delta encoding

Delta encoding is a way of storing or transmitting data in the form of differences (deltas) between sequential data

Design principles

• Non-blocking code, Future based API.
• Pluggable architecture. Storage and messaging are implementation neutral. With multiple supported implementations, it's easier to use with existing infrastructure. Current implementations:
  • Storage:
    • Generic JDBC, with MySQL, and PostgreSQL adaptations. Custom adaptions are straightforward.
    • MongoDB
    • Cassandra
  • Messaging:
    • Redis Pub/Sub
    • Hazelcast Topic
• No third-party dependencies (except for storage/messaging implementations)
• Minimally invasive. Large degree of freedom in implementation. Uses type classes instead of inheritance. No annotation hell.
• Fast, scalable. Written for very large data sets; supports both horizontal and vertical scaling.
• Customizable. Attempts to provide sane defaults, but enables high degree of customization.

Why event-sourcing?

• Complete data history. Intrinsically complete and correct audit trail of any and all changes.
• Very fast. Append-only writes. Prebuilt reads (push and pull).
• Scalable; effortless sharding.
• No object-relational mismatch.
• No RDBMS schema upgrades.
• Fully separate OLTP/OLAP.
• Easy and flexible authentication. Any and all updates are done through explicit, use case specific, commands, each of which can have custom authentication, if required.