Starling is McLeod Moores' technology platform for market risk and analytics. It is a fork of the OpenGamma platform (2.2) that has been independently developed by us in private. We are releasing our changes back to the community and will now work in public.
Our aims and principles:
- We won't break APIs unless we're correcting behaviour and when we replace APIs we will provide a clear migration path. We value your participation.
- We want to make the common use cases much easier to use, and the more complex configurations still possible.
- We want to promote the use of the Analytics library as a stand-alone component for when existing infrastructure is already in place in your application.
- We want to improve the documentation and invite others to participate.
Starling continues support for fully streaming, real-time analytics, while simplifying the API particularly for the more common use-cases like on-demand calculations using daily snapshots and historical data provided by the user.
Additionally, while preserving the abililty to scale to a multi-node micro- services style configuration, we have focussed on making simpler, single-node setups easier to run by separating the micro-services dependencies into separate packages. This yields large savings in memory and extraneous dependencies in simple use-cases that make it much easier to embed Starling in existing applications.
Firstly you need to make sure Apache Maven and Git are installed and working. Version 3.0.4 or later of Maven is required.
The Starling Platform is open source software using the Apache License v2. The company behind Starling also offers support, consultancy and some additional commercial components.
The source code can be cloned using git from GitHub:
git clone https://github.com/McLeodMoores/starling.git
The source code must be compiled before use. This will build multiple jar files and install them into your local Maven repository. Simply run this command from the root directory of the source code:
mvn install
The command above will run unit tests. These can be skipped to save time if desired:
mvn install -DskipTests
A core component in Starling is known as the "engine". For production, the engine is typically customized, however two example engine configurations are pre-supplied, one with Bloomberg support and one with simulated market data.
To run the example engine, change to the examples/examples-simulated directory and run the following commands:
cd examples/examples-simulated
mvn starling:server-init -Dconfig=fullstack
mvn starling:server-run -Dconfig=fullstack
Note that the "server-init" command only needs to be run once - the server can be stopped and restarted without having to repopulate the databases.
Wait for the server to start:
and then point your browser to
http://localhost:8080 to see the web user interface:
Go to http://localhost:8080/jax/components to get a sense of
the underlying power of the system, available via REST. Some of the components that are available can be seen below:
Clicking on "Analytics" opens a UI that shows calculation results when a view is selected:
An explanation of what the example views calculate can be found here.
For more information go to http://www.mcleodmoores.com