GitLicense Build Status BCH compliance Codacy Badge Project Stats

IGUANA

IGUANA Logo

Youtube Tutorials:

IMAGE ALT TEXT HERE

ABOUT

Semantic Web is becoming more important and it's data is growing each day. Triple stores are the backbone here, managing these data. Hence it is very important that the triple store must scale on the data and can handle several users. Current Benchmark approaches could not provide a realistic scenario on realistic data and could not be adjustet for your needs very easily. Further on it was impossible to compare results for different benchmarks.

Iguana is an An Integrated Suite for benchmarking SPARQL which solves the issue. It provides an enviroment which ...

For further Information visit

iguana-benchmark.eu

Wiki

Getting Started

Prerequisites

You need to install Java 8 or greater and RabbitMQ Version 4.x.z or greater. In Ubuntu you can install these using the following commands

sudo apt-get install java
sudo apt-get install rabbitmq-server

RabbitMQ will be automatically started after you installed it.

Iguana Modules

Iguana consists of two modules!

  1. corecontroller: This will benchmark the systems
  2. resultprocessor: This will calculate the Metrics and save the raw benchmark results

Further on you need to install and start the message brocker RabbitMQ. It is needed for communication between the corecontroller and the resultprocessor.

corecontroller

The corecontroller will benchmark your system. It should be started on the same machine your Triple Store is started. It will be started as a daemon process in the background and you can send a benchmark configuration to this module (see below).

It will start the benchmark according to the benchmark configuration and will send data about each executed query to the resultprocessor. The data includes sent for each query includes:

resultprocessor

The resultprocessor will be started as a daemon too. Its behavior is widely configurable. By default it stores its result in a ntriple file. But you may configure it, to write the results directly to a Triple Store. On the processing side, it calculates various metrics.

Per run metrics:

Per query metrics:

You can change these in the resultprocessor configuration file.

If you use the basic configuration, it will save QMPH, NoQPH and QPS to a file called results_{{DATE_RP_STARTED}}.nt

Setup Iguana

Download

Please download the release zip Iguana_Release.zip from the newest release available here:

wget https://github.com/dice-group/IGUANA/releases/download/v2.1.0d/Iguana_Release.zip
unzip Iguana_Release.zip
cd Iguana_Release/

It contains the following files:

Start Daemons

Use the start script

./start-iguana.sh

Now the iguana daemons are running in the background.

Run Your Benchmarks

Create a Configuration

You can use the basic configuration we provide and modify it to your needs. For further information please visit our configuration and Stresstest wiki pages. For a detailed, step-by-step instruction please attend our tutorial.

Execute the Benchmark

Make sure you:

  1. started Iguana
  2. created your benchmark configuration
  3. started and setup the Triple Store you want to bench.

Assuming your benchmark configuration is called benchmark.config, you can start your benchmark now with:

./send-config.sh benchmark.config

It will send your configuration the corecontroller and start it.