Software Failure and Reliability Assessment Tool (SFRAT)

A web oriented software reliabilty testing suite written in R. Made to parse data sets and predict failure models based on common statistical approaches using the shiny framework.

Lance Fiondella, University of Massachusetts Dartmouth

Allen Nikora, Jet propulsion laboratory/ California Institute of technology

Installation

Before installing R libraries. Please make sure you have some version of Perl installed. For Linux and Mac computers it is not usually required. For Windows machines you can download it here: http://strawberryperl.com/

Perl is required to import excel files into the tool

To install the required libraries, please run the following command in the R terminal

source("/Path/To/SRT/install_script.R")

Replace /Path/To/SRT/ with the path to the folder where you have extracted the tool. A window will pop up for you to select the server from which to download the packages. Select a server and the script will download and install packages.

To Run

library(shiny)  
runApp("/Path/To/SRT/")

This will launch the app on localhost on a random port and attempt to open a browser.

Deploy using docker image on Linux

• SRT is available as docker image and can be pulled from here (https://hub.docker.com/r/arhik/srt/)
• Install docker-engine (follow instructions on https://docs.docker.com/engine/installation/)
• Pull the image using
  $ docker pull arhik/srt
• deploy by portforwarding desired port on your system to the default 3838 port using
  $ docker run -d -p xxxx:3838 arhik/srt

Contribution:

The aim of this tool is to allow third party contributions both in terms of architecture and models.

Model Contributors Guide

---

• Link to contributors guide ()

Cons:

• PDF export of tables is not supported yet.
• Only NHPP Models are included.

TODO:

• Python port of this is in progress.

FOR USERS with JIRA Export Data

• Follow the script developed by MITRE, which is available at https://github.com/LanceFiondella/SRGM
  • To manipulate the JIRA export data to a form acceptable by the SFRAT
  • Using the JIRA export evaluate a Mean Time between Fault(failure) - MTBF and Mean Time to Fix Code(Repair)

Versioning:

• v1.0 is released.

Acknowledgement

This research was supported by (i) the Naval Air Systems Command (NAVAIR) through the Systems Engineering Research Center (SERC), a Department of Defense (DoD) University Affiliated Research Center (UARC) under Research Task 139: Software Reliability Modeling and (ii) the National Science Foundation under Grant Number (1526128).

Copyright and License

Code release under MIT LICENSE.