Vortex Series

Vortex is a lightweight distributed streaming computing framework. It is based on memory computing, which is suitable for high availability, high concurrency and real-time computing business scenarios.

Vortex is developed based on SpringBoot framework. It relies on tridenter component to build cluster. Vortex application expose extra independent TCP server port to accept external connections. Vortex application cluster member discover for each other and establish long-term connections through the net multicast to achieve high availability, decentralization and load balancing.

Components

The vortex series consists of three parts:

• vortex-common client side, vortex common classes, it makes a Java application as vortex client.

• vortex-spring-boot-starter server side, the core class of the vortex framework. it makes spring application as vortex cluster member.

• vortex-metrics-api a time series computing tool library, it is used to do relevant programming on time series computing based on memory

• vortex-metrics a web application based on time series computing tool. It is easy to be scalable and have high performance in realtime situation

Compatibility

1. Jdk8 (or later)
2. SpringBoot Framework 2.2.x (or later)
3. Redis 3.x (or later)
4. Netty 4.x (or later)

Install

Server Side

<dependency>
	<groupId>com.github.paganini2008.atlantis</groupId>
	<artifactId>vortex-spring-boot-starter</artifactId>
	<version>1.0-RC1</version>
</dependency>

Client Side

<dependency>
   <groupId>com.github.paganini2008.atlantis</groupId>
   <artifactId>vortex-common</artifactId>
   <version>1.0-RC1</version>
</dependency>

At present, following open source projects based on vortex framework:

1. Jellyfish
2. Greenfinger
3. Vortex metrics

Quick Start

How to use the vortex API in your application?

Server Side:

Implement Handler interface

@Slf4j
public class TestHandler implements Handler{

    @Override
    public void onData(Tuple tuple) {
        log.info(tuple.toString());
    }

}

Client Side:

String brokerUrl = "http://localhost:10010"; // Expose the location of Server Side
TransportClient transportClient = new HttpTransportClient(brokerUrl);
Tuple tuple = Tuple.newOne();
tuple.setField("clusterName", clusterName);
tuple.setField("applicationName", applicationName);
tuple.setField("host", host);
tuple.setField("message", "Hello World!");
transportClient.write(tuple);

How to retain historical metrics data ?

Default setting shows statistical time window is 1 minute and rollingly retain recent 60 records, discarding historical records once exceeding the size. If you want to save historical metrics data, You need to:

• Implementing interface io.atlantisframework.vortex.metric.api.MetricEvictionHandler to customize your persistence policy
• Implementing interface io.atlantisframework.jellyfish.http.MetricSequencerFactory or expanding class io.atlantisframework.jellyfish.http.DefaultMetricSequencerFactory and redefine a GenericUserMetricSequencer