Skip to main content

Running Ember using Java

Prerequisites​

Access TimeBase using one of the following options:

  • To access market prices, Ember requires the Enterprise version of TimeBase.
  • To run TimeBase locally, use the TimeBase Initializr. You can also connect Ember to an existing TimeBase server running nearby.

Make sure you have Java JDK 11 or 17 installed.

The TimeBase client inside Ember requires the JDK package (rather than the JRE package) of Java.

Install​

To install Ember:

  1. Extract the Ember distribution package deltix-ember-version.zip into a designated installation directory. For example: /home/deltix/ember. Various launch scripts described on this page can be found under the bin subdirectory of your installation.

  2. Set an environment variable called EMBER_HOME that defines the Ember home directory where configuration files will live:

Linux: 

export EMBER_HOME=/home/deltix/emberhome

Windows: 

set EMBER_HOME=D:\Deltix\QuantServer\emberhome

Configure​

To configure Ember:

  1. Create a configuration file called ember.conf under EMBER_HOME and add the following block:

    timebase.settings {  
        url = "dxtick://localhost:8011"  
    }

    You may need to correct the host, port, and logon credentials to match your TimeBase server.

    Ember configuration files use the HOCON format described here.

  2. Tell Ember which TimeBase streams have market data:

    pricing.settings {  
      liveSubscription {  
        streams = ["COINBASE", "BINANCE"]  
      }  
    }

    Make sure the stream names match the exchanges you configured in the TimeBase Initializer.

  3. Configure a simple exchange simulator so you can send test trades. Add the following configuration fragment to enable this trading simulator as a SIM destination:

    connectors {
      SIM : ${sim} {
      }
    }

Start​

Finally, start Ember using the script bin/ember:

    ______          __
   / ____/___ ___  / /_  ___  _____      Version:  0.2-SNAPSHOT  
  / __/ / __ `__ \/ __ \/ _ \/ ___/      Timebase: 5.2.6x.44760  
 / /___/ / / / / / /_/ /  __/ /          Home:     D:\Deltix\QuantServer\ember  
/_____/_/ /_/ /_/_.___/\___/_/           Copyright (c) 2017 Deltix, Inc.

29 Sep 12:49:30.358422 INFO [trade-engine] Log is replayed, messages read 0
29 Sep 12:49:30.361264 INFO [trade-engine] Startup complete. This server is ready to process new requests.

Monitor Trade Flow​

Ember has an app, Ember Monitor, that can monitor events inside Ember OMS. This includes the state of orders, trading connectors, risk limits, and internal metrics like queue fill rate.

Ember Monitor can be started using the bin/ember-monitor script and accessed on port 8988:

Ember Monitor screenshot

tip

Ember Monitor needs the same EMBER_HOME environment variable we set up earlier to find Ember Journal and access system configuration.

Congratulations, Ember is running!

Further Steps​

Separate Output Directory​

In addition to the environment variable EMBER_HOME, you can use EMBER_WORK to define an Ember work directory. The work directory is used for all output files (i.e., transaction journal, log files, counters, etc.).

Linux: 

export EMBER_WORK=/mnt/nvme/ember

Windows: 

set EMBER_WORK=M:/ember

We recommend keeping the work directory separate from the home directory. For a higher performance work directory, it can be placed on faster storage (NVMe, high IOPS volume, etc.).

tip

Ember Monitor and other satellite Ember processes share the EMBER_WORK and EMBER_HOME environment variables.

Deploy an Algorithm​

Trading Algorithms are the building blocks of an Execution Server. Let's deploy the "ICEBERG" algorithm that takes a large order and slices it into smaller chunks visible on execution venue.

Add the following block into ember.conf to deploy the ICEBERG algorithm implemented by the deltix.ember.service.algorithm.samples.iceberg package:

algorithms {  
  ICEBERG : ${template.algorithm.default} {  
    factory = deltix.ember.service.algorithm.samples.iceberg.IcebergSampleAlgorithmFactory  
    queueCapacity = 4M  
    settings {  
      destinationId = SIM
    }  
  }  
}

You can read more about the design of algorithms and deployment parameters in the Algorithm Developer's Guide.

Configure a Trade Connector​

Trade connectors route orders to execution venues and bring back trading events.

For example, let's assume we have a connection to a CME iLink MSGW gateway. The following configuration fragment enables the CME connector and defines connection parameters:

connectors {  
  CME : ${CMEMSGW} {  
    settings {
      senderCompId = ROGUE1N
      targetCompId = CME  
      host=69.50.112.141  
      port=61529  
    }  
  }  
}

The above connector is registered under the name CME (case-sensitive). This definition uses the predefined connector type "cmemsgw" that converts Ember orders to CME MSGW iLink format and talks to CME iLink.

note

The Execution Server is integrated with 100+ different execution venues like CME.

You can read more about the design of trading connectors and deployment parameters in the Trading Connector Developer's Guide and the Configuration Reference Manual.

EmberPack​

If you attempt to restart Ember to apply the changes we just made, you will notice a Class Not Found error because Ember won't be able to find the CME connector code. The connector lives in the EmberPack container. This container extends Ember with an additional layer containing compiled versions of all supported trading connectors.

Send a Test Order​

There are several ways to send test orders:

Further Reading