Exelix ^TM LLC

Galileo^TM is the most advanced tool for dynamic fault tree analysis available today

	advanced methods of precise, scalable, efficient analysis of dynamic fault trees
	a serious concern for the production of correct and valid analysis results
	unparalleled ease of learning, ease of use, and seamless integration
	competitive terms enabled by advanced software design methods.

This page provides a little background on Galileo and then elaborates each of these aspects of the Galileo software. Read on for more information, including a screen shot of Galileo's graphical editing interface.

Background

Galileo was originally developed at the University of Virginia under contract to the NASA Langley Research Center (NAS1-99098 and NAS1-02076). The U.S. Government retains certain rights in the University software. Information on the research underlying Galileo is available here (a site Exelix neither controls nor can endorse). Exelix has retained the project and product name Galileo, which has gained wide visibility through the underlying research efforts. Galileo/Exelix delivers these advanced research results in a form appropriate for industrial use.

Leading-Edge Techniques for Precise, Scalable Analysis

Galileo supports efficient precise dynamic fault tree analysis. Modeling features of Galileo include multiple time-to-failure distributions (fixed, exponential, lognormal, and Weibull), static and dynamic coverage modeling, phased mission modeling, and sensitivity analysis.

In effect Galileo converts a given dynamic fault tree to an equivalent Markov chain for analysis by standard Markov methods. Galileo thus provides reliability engineers the ability to model fault-tolerant systems with complex redundancy management using high-level fault tree notations rather than unwieldy, low-level Markov chain techniques.

A problem with a naïve translation of fault trees to Markov chains is that it scales poorly: Markov chains grow exponentially in the number of events in a fault tree, making the naive approach unworkable, in general, due to overwhelming demand for memory.

Galileo provides leading-edge methods to overcome this problem without sacrificing exact solutions. To produce exact solutions efficiently, Galileo's engine decomposes a given dynamic fault tree into its independent sub-trees and it then solves the sub-trees independently using the most appropriate methods . The result is a vast reduction in state space, in general. If a sub-tree has no dynamic gates, Galileo solves it by even more efficient Boolean Decision Diagram (BDD) techniques. Galileo re-combines module results to produce an overall result.

Galileo’s combination of modularization, dispatching of sub-trees to the appropriate solvers, and the re-combination of module results into exact overall results puts Galileo in a class of its own as the leading-edge analytic tool for dynamic fault tree analysis.

Concern for Dependability of Computed Results

Software engineering methods today are inadequate to guarantee the absence of faults in the specifications or implementations of complex software systems, including fault tree analysis tools. It is thus inadvisable to make critical decisions solely on the basis of results computed by a single, computationally complex software tool, whether Galileo or a tool of another vendor. The user of any such engineering modeling and analysis tool must take responsibility for decisions based on the use of the tool.

Users of such tools must bear in mind that they might not always produce correct results. Research has shown several popular, commercial fault tree analysis tools to have systematically produced incorrect results in some cases. Common algorithmic shortcuts appear to produce common failure across several such tools. According to a 1999 study published in the IEEE Transactions on Reliability Engineering, "Testing two tools popular in the reliability engineering community revealed the same algorithmic error in both, despite their claimed ability to provide exact solutions." Source: Dugan, Sullivan, Coppit. Developing a low-cost, high-quality software tool for dynamic fault tree analysis. Transactions on Reliability, December 1999, pp. 49-59.

To provide a basis for additional (albeit still limited) confidence in Galileo's analysis, its specification and implementation employ certain advanced software engineering methods, including the formal, mathematical specification of the mapping functions from fault trees to Markov chains, and bounded, systematic testing of key elements of the Galileo engine against this specification.

Unprecedented Ease of Learning, Ease of Use, and Integration

Galileo's achievement of ease of learning, ease of use, and tight integration into standard environments is based on its unique-in-the-industry use of Microsoft Office and related applications as major user interface components. In particular, Microsoft's Visio^TM drawing tool, extended with our own high-level fault tree editing functions, provides most of Galileo's graphical editing, display, formatting, printing, cut-and-paste and related capabilities. You get all of this without paying for a non-standard graphical editor, costly to build and unnecessarily hard to use. To use Galileo's graphical editing features, you do need to have Visio^TM installed on your PC. Without Visio, you can use Galileo's integrated Microsoft Word editor to enter fault trees in a textual form.

Competitive Pricing and Licensing Terms

Please contact us for information on our very competitive pricing for Galileo/Exelix, available to early adopters of the tool.

[ Home ] [ Up ]

Exelix TM LLC