T-Rex is a general purpose Complex Event Processing (CEP) Middleware designed to support an expressive language for rule definition while offering efficient processing mechanisms.
With the T-Rex tools users can create TESLA rules, publish events and subscribe to event notifications.
TESLA is a powerful yet simple complex event definition language.
TESLA includes operators to express:
- Content-based event filtering
- Customizable event selection and consumption policies
- Event sequences with timing constraints
- Negations (time-based and interval-based)
- Parameters
- Aggregates
Further details on TESLA can be found in the following paper: TESLA: a Formally Defined Event Specification Language, G. Cugola, A. Margara. In Proceedings of 4th ACM International Conference On Distributed Event-Based Systems (DEBS 2010). Cambridge, United Kingdom. July 12 - 15, 2010
The T-Rex software suite consists of 3 components: a C++/CUDA highly optimized engine, a C++ server and a Java client.
Further details on T-Rex algorithms and implementation can be found in the following paper: Complex Event Processing with T-Rex, G. Cugola, A. Margara. In Journal of Systems and Software. Volume 85 - Issue 8. August 2012.
More on the GPU implementation can be found in the following thesis: GTrex : implementing a CEP engine for the TESLA language on GPU, Daniele Rogora.
https://www.politesi.polimi.it/bitstream/10589/88428/1/Tesi.pdf
The client includes the code to actually perform the communication between the client itself and the server as well as a TESLA rule parser.
More on the specific syntax used for the parser can be found in the PDF document TESLA_parser_intro in this repository.
This is the core of the project, containing both the standard and the CUDA powered engine.
First of all build the configure script:
$ autoreconf --install
To configure:
$ ./configure
To build also the CUDA engine:
$ ./configure --enable-cuda
Build the binary library:
$ make
To install the library and the development headers:
$ make install
This is a server built with the TRex2 library that handles the communication between the engine and the clients.
First of all build the configure script:
$ autoreconf --install
To configure:
$ ./configure
Build the binary:
$ make
To install the library and the development headers:
$ make install
To run the server with the standard engine:
$ TRexServer
To run the server with the CUDA engine:
$ TRexServer -gpu
This project contains both a Java library to simply and efficiently take advantage of the T-Rex server and an example command line client developed with the library.
To build the example command line client:
$ ant jars
Run with:
$ java -jar TRex-client.jar
Other options can be seen with:
$ ant -p
This project contains a very simple proxy written in node.js, which converts REST calls into calls to the T-Rex server.
The proxy can be launched through the command:
$ nodejs trex-proxy.js
waits connections on port 8888, and connects to the T-Rex server running on localhost:50254 (the default). Use the following environment variables to change these defaults: TREX_HOST, TREX_PORT
The proxy exports five main REST services: one to connect, one to publish events, one to subscribe to events, one to unsubscribe, and one to get received events (which it stores, on behalf of clients). The first is a GET request, those to publish and subscribe are POST requests, the one to unsubscribe is a DELETE request, the latter is a GET request.
Clients are identified through an UUID they receive at connection time.
File "testjs.html" (served through the URL: http://proxy.machine:8888/testjs.html) provides a starting example of usage from Javascript. The javascript client-side library used by such an example can be found in file trex-client.js.
The license is usually specified at the beginning of each source file; where not explicitly stated assume GPLv3.
Gianpaolo Cugola
Alessandro Margara
Daniele Rogora