consysT

consysT is a language and middleware which allows programmers to replicate data under specified consistency levels. The type system ensures safe mixing of consistency levels.

The language has a binding for Java. The implementation of the middleware is in Scala.

Find more information on our website at https://consysT-project.github.io.

How to install?

The project is built with Maven. Detailed instructions are found at https://consyst-project.github.io/install.html.

IntelliJ

In IntelliJ, you have to add the annotation processor manually.

1. Go to Preferences > Annotation Processors. You can try to check Obtain from project. IntelliJ may be able to retrieve the correct checker by default. If it can, you are already finished.

2. Add the compiled jar of the consys-type-checker project

3. Add the checker.jar from the CheckerFramework

4. Choose de.tu_darmstadt.consistency_types.checker.ConsistencyChecker as annotation processor.

External dependencies

These applications have to be installed and running to correctly use the system:

(For the Cassandra binding):

• Apache Cassandra 4.0.0-alpha3
• Zookeeper 3.5.6

Cassandra

You need Cassandra 4.0+ to run with Java 11.

If using ccm then use ccm setdir to set the correct Cassandra install directory.

Project overview

The code is structured in the following subprojects:

• consys-core: The implementation of the middleware built with (a) Akka, or (b) Cassandra and Zookeeper.
• consys-japi: The Java binding of the languages syntax.
• consys-compiler: Compiler plugin for the Java binding.
• consys-type-checker: Compiler plugin for the consistency type checker built with the Checker framework.
• consys-bench: General benchmark interface.
• integration-tests: Tests and demo for the completely integrated project.
• demos, examples, microbenchmarks: Benchmarks and demo applications that use consysT.

About

This project has been developed at the Reactive Programming Technology Group of Technische Universität Darmstadt.

The main developer is Mirko Köhler under supervision of Prof. Guido Salvaneschi.

We thank the many developers that helped with the project:

• PhD students
  • Nafise Eskandani
  • Pascal Weißenburger
• Students
  • Victor Schümmer and Jesper Schlegel
  • Martin Edlund
  • Matthias Heinrich and Julian Hindelang
  • Pascal Osterwinter