Browse Prior Art Database

Algorithm for Drawing Hierarchical Representations of Asynchronous Traces

IP.com Disclosure Number: IPCOM000117619D
Original Publication Date: 1996-Apr-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 107K

Publishing Venue

IBM

Related People

Lehr, TF: AUTHOR

Abstract

Disclosed is a process for efficient drawing of computing traces. Computing traces often contain more information than can be elegantly managed or understood using standard textual representations. Tools for visualizing trace data represent large traces in intellectually manageable ways. Efficient drawing of the visual trace representations is important for such tools to be usable.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Algorithm for Drawing Hierarchical Representations of Asynchronous
Traces

      Disclosed is a process for efficient drawing of computing
traces.  Computing traces often contain more information than can be
elegantly managed or understood using standard textual
representations. Tools for visualizing trace data represent large
traces in intellectually  manageable ways.  Efficient drawing of the
visual trace representations  is important for such tools to be
usable.

      Programmers trace software in order to understand the behavior
of programs.  Traces consisting of time-stamped logs of events
executed by software are used to debug functional and performance
problems.  Because traces frequently yield hundreds or thousands of
pages of text output, post processing programs are used to translate,
filter and condense the information in more manageable forms.
Visualization of traces is one of the resulting forms.  This
disclosure addresses the problem of drawing events quickly by visual
post processors.

      PieScope is a trace visualization tool which collates and draws
events along parallel time-lines.  It is used internally within IBM*
to diagnose application and system performance problems.

      The drawing of trace events must be correct, versatile and fast
if the visualization tool is to be usable.  A problem with drawing
events is that in a typical trace, the distribution of
event-durations spans several orders of time magnitude, making some
events too small to be seen.

      Other state-of-the-art trace visualization tools often
calculate the size of every event before determining whether to draw
it.  This is an inefficient method for it does not take advantage of
the common instrinsic topological properties of trace data.  The
inefficiency results in slow draw times and simplistic, monolithic
visual data representations akin to assigning equal importance to
molecules and solar systems at every level of magnification.

      This solution for removing the drawing inefficiencies assumes
that the trace data has been ordered in a hierarchical tree-like
graph in which the closer a node i...