Timeout in total ordering architecture

The approach to updating outdated data while preserving determinism in a total ordering architecture (such as consensus-based systems).

f(command[i -> j], timestampX) == f(command[i -> j], timestampY)

flowchart LR
    %% Producers
    subgraph P[" "]
        p1["producer"]
        p2["producer"]
        p3["producer"]
    end

    %% Command queue
    subgraph Q[" "]
        direction LR
        c1["command"]
        c2["command"]
    end

    %% Consumer section
    subgraph C[" "]
        direction TB
        loop["loop"]
        journal["journal"]
        logic["handle logic"]
        state["state"]
        timers["sortedTimers"]
        tick["tickCommand"]
        spawn["spawn tick interval"]

        spawn --> tick
        logic --> state
        state --> timers
    end

    %% Connections
    p1 -->|"publish command"| Q
    p2 -->|"publish command"| Q
    p3 -->|"publish command"| Q

    Q -->|"consume"| loop
    loop --> journal
    journal --> logic

    tick -->|"publish"| Q

Loading

Stacks

• Java 25
• Spring Boot 4.0.0-SNAPSHOT - Web framework
• RocksDB 9.4.0 - Embedded key-value database
• JCTools 4.0.5 - High-performance concurrent data structures
• Agrona 2.3.0 - High-performance data structures and utilities

Development

./gradlew bootRun -Dspring-boot.run.jvmArguments="--enable-native-access=ALL-UNNAMED"

Swagger

http://localhost:8080/swagger