/IoTCOM

Verifying safety of interactions between IoT apps

Primary LanguageAlloyMIT LicenseMIT

Overview

In this repository you will find the protype code of IoTCOM, a model-checking tool to verify the safety of IoT Apps.

This work is published at ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA'20) and received ACM SIGSOFT Distinguished Paper Award.

Here you can find the draft paper

Here is the project website

IoTCOM Architecture

IoTCOM comprises two components:

1- Behavioral Rule Extractor (BRE): generates the behavioral models of the IoT apps (groovy and IFTTT).

2- Formal Analyzer (FA): relies on Alloy to perform formal verification to check the safety IoT apps.

picture

IoTCOM_Behavioral Rule Extractor

This component has been tested on MacOS and Linux. To use this component:

  1. Create an Eclipse project and clone this repository (you can find the instructions here).
  • Add all libraries (.jar files) in the BehavioralRuleExtractor/lib folder.
  1. Run the "ToAlloy.java" as "Java Application". Put the Goovy apps inside the directory input/sampleGroovyApps. More Groovy apps can be obtained from SmartThings Public GitHub Repo. The resulting Alloy models (.als files) will be generated and stored in /IoTCOM/BehavioralRuleExtractor/output/sampleApps. The models that are generatd using our tool are in the directory Dataset_RealWorldApps

IoTCOM_Formal Analyzer

The Formal Analyzer component can be built and run using Gradle. Java 8 or above is required. This component has been tested on MacOS and Linux.

To build an all-in-one jar file using Gradle:

  1. Clone the repository.
  2. Change to the FormalAnalyzer folder.
  3. Run ./gradlew customFatJar.
  4. Run java -jar build/libs/iotcom-all-1.0.0.jar [arguments]

The arguments are described in greater detail in the FormalAnalyzer Java file. The user can specify a variety of options, such as the directory for app files, the output directory, the size of the generated bundles, etc.

⚠️ Running with the defaults will take a very long time. Specify a value for the --appsdir argument on the command line.

As an example, to run the same analysis as our test files, invoke the jar like this:

java -jar build/libs/iotcom-all-1.0.0.jar \
  --appsdir models/test/ \
  --seed 1587400397957 \
  --bundlesizes 3 \
  --prplist t1

To run apps generated by the Behavioral Rule Extractor, specify the output directory of that tool instead (e.g., --appsdir ../BehavioralRuleExtractor/output/sampleApps).

The app was built using IntelliJ IDEA, but could also be imported into Eclipse using the Gradle import wizard.

Citation

@INPROCEEDINGS{iotcom,
    AUTHOR="M. Alhanahnah and C. Stevens and H. Bagheri",
    TITLE="Scalable Analysis of Interaction Threats in IoT Systems",
    BOOKTITLE="ISSTA 2020",
    YEAR="2020",
}