Cloud2Sim

An Adaptive and Distributed Architecture for Cloud and MapReduce Algorithms and Simulations.

Cloud2Sim proposes a distributed concurrent architecture to CloudSim simulations.

Building and Running Cloud2Sim

The easiest way to build and run the simulation is using the automated setup script.

1. Setup and Build

Run the setup script from the project root. This will compile the project with Maven and generate a run_sim.sh wrapper that handles all Java version-specific settings (like --add-opens for modern JDKs).

./setup.sh

2. Running Simulations

Use the generated run_sim.sh script to launch any simulation class. The script automatically manages the classpath and JVM arguments.

Static Simulation:

./run_sim.sh pt.inesc_id.gsd.cloud2sim.applications.main.statics.Simulator

Adaptive Scaling:

# Master node
./run_sim.sh pt.inesc_id.gsd.cloud2sim.applications.main.dynamics.ScalableSimulator

# Worker nodes
./run_sim.sh pt.inesc_id.gsd.cloud2sim.scale.ias.IntelligentAdaptiveScaler

Map-Reduce Simulations:

# Hazelcast
./run_sim.sh pt.inesc_id.gsd.cloud2sim.mapreduce.hazelcast.HzMapReduceSimulator

# Infinispan
./run_sim.sh pt.inesc_id.gsd.cloud2sim.infinispan.Initiator
./run_sim.sh pt.inesc_id.gsd.cloud2sim.mapreduce.infinispan.InfMapReduceSimulator

Adaptive Scaling

Initiate the ScalableSimulator on the master node:

java $JAVA_OPTS -classpath "lib/*:." pt.inesc_id.gsd.cloud2sim.applications.main.dynamics.ScalableSimulator

Initiate the IntelligentAdaptiveScaler on the worker nodes:

java $JAVA_OPTS -classpath "lib/*:." pt.inesc_id.gsd.cloud2sim.scale.ias.IntelligentAdaptiveScaler

Distributed Infinispan Simulation

Cloud2Sim leverages Infinispan's distributed data grid capabilities. You can run multiple instances of the Infinispan nodes to simulate a distributed environment.

1. Initiator: Acts as a cluster node. It can be used to hold data or serve as a persistent member of the grid.

   ./run_sim.sh pt.inesc_id.gsd.cloud2sim.infinispan.Initiator

2. Simulator: Joins the Infinispan cluster formed by the Initiator(s) and triggers the MapReduce tasks across all available nodes.

   ./run_sim.sh pt.inesc_id.gsd.cloud2sim.mapreduce.infinispan.InfMapReduceSimulator

Note: You can run one Initiator and one Simulator on the same machine, or distribute them across different machines by updating the JGroups configuration in conf/jgroups-tcp-config.xml with actual IP addresses.

Map-Reduce Configuration

Workload files and dataset configurations are managed in conf/cloud2sim.properties:

• loadFolder: Path to the input text files (e.g., conf/mapreduce/load).
• mapReduceSize: The number of lines to process per file.
• isVerbose: Set to true to see detailed word count results.

Using Hazelcast

Hazelcast supports distributed execution using an Initiator/Simulator pattern. You can form a cluster by running one or more Initiator nodes.

1. Initiator: Starts a standalone Hazelcast node to join the cluster.

   ./run_sim.sh pt.inesc_id.gsd.cloud2sim.hazelcast.HzInitiator

2. Simulator: Joins the Hazelcast cluster and executes the MapReduce simulation across all nodes.

   ./run_sim.sh pt.inesc_id.gsd.cloud2sim.mapreduce.hazelcast.HzMapReduceSimulator

Java Compatibility

Cloud2Sim is configured for Java 11 and newer. The run_sim.sh script automatically applies the necessary --add-opens flags required for Hazelcast and Infinispan to work on modern JDKs (9, 11, 17, 21+).

Using Infinispan

Follow the Initiator/Simulator pattern described above.

Citing Cloud2Sim

If you use Cloud2Sim in your research, please cite the following papers:

• Kathiravelu, P. & L. Veiga (2014). An Adaptive Distributed Simulator for Cloud and MapReduce Algorithms and Architectures. In IEEE/ACM 7th International Conference on Utility and Cloud Computing (UCC 2014), London, UK. pp. 79 – 88. IEEE Computer Society.

• Kathiravelu, P. & L. Veiga (2014). Concurrent and Distributed CloudSim Simulations. In IEEE 22nd International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS'14), Paris, France. pp. 490–493. IEEE Computer Society.