Browse Prior Art Database

Distributed Hierarchical Error Structure Contextual Promotion Methodology

IP.com Disclosure Number: IPCOM000110336D
Original Publication Date: 1992-Nov-01
Included in the Prior Art Database: 2005-Mar-25
Document File: 2 page(s) / 59K

Publishing Venue

IBM

Related People

Bugg, A: AUTHOR [+6]

Abstract

In a distributed network client/server model environment, Error states exist perpetually, but often merely as transient states of stability; as networked exchange state information they necessarily anticipate the context of a peer or a client/server. In this environment, extraneous error conditions can seriously impede the detection and resolution of real, non-trivial problems, simply by overloading the recovery and logging facilities of the environment.

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

Distributed Hierarchical Error Structure Contextual Promotion Methodology

       In a distributed network client/server model environment,
Error states exist perpetually, but often merely as transient states
of stability; as networked exchange state information they
necessarily anticipate the context of a peer or a client/server. In
this environment, extraneous error conditions can seriously impede
the detection and resolution of real, non-trivial problems, simply by
overloading the recovery and logging facilities of the environment.

      In this invention, low-level error condition detections are
evaluated automatically to determine from the current context
whether they are non-trivial.  Current context is evaluated based on
system state, alternate path potential and correction code methods.

      Error conditions in a distributed environment are highly
volatile and often transient in nature.  The interactive and
interdependent complexities  of a LAN-based operating environment are
so great that they approach stochastic aspects at indeterminate
times.  Therefore, a sophisticated set of error tracking and scaling
heuristics are required in order that error conditions which are not
contextually significant do not precipitate corrective action.  The
situation is complicated by the fact that certain errors are
catastrophic at certain states, and are trivial at others.

      In this invention, a low-level failure condition is detected
(e.g., a file not found condition) and interpreted by the detecting
function.  If that function does not possess sufficient information
to determine the criticality of the failure, it defers external
reporting (logging) and error recovery procedures, and routes a
reporting structure to a higher level routine.  The higher-level
routines interpret the structure, and from current state information
(e.g., is...