Browse Prior Art Database

Improved Communication Availability by Controlling Retry Counts and Times

IP.com Disclosure Number: IPCOM000122380D
Original Publication Date: 1991-Dec-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 3 page(s) / 97K

Publishing Venue

IBM

Related People

Dancker, GA: AUTHOR [+3]

Abstract

A Work Station Controller (WSC) provides attachment of multiple work stations to a computer system via some type of communication protocol. In most WSC communication protocols, transmission errors are retried. Transmission errors can occur on either a command (requesting the work station to perform a function) or on a poll (checking for completion of a previously sent command or polling for a new event, for example, a user keystroke). The error is considered permanent and the work station taken off-line after a set number of consecutive unsuccessful retries (retry count). This article describes how the retry count can be dynamic based on the type of error and availability of resources within the WSC. Also, the WSC will insert added delays between retry attempts for certain types of errors.

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

Improved Communication Availability by Controlling Retry Counts and
Times

      A Work Station Controller (WSC) provides attachment of
multiple work stations to a computer system via some type of
communication protocol.  In most WSC communication protocols,
transmission errors are retried.  Transmission errors can occur on
either a command (requesting the work station to perform a function)
or on a poll (checking for completion of a previously sent command or
polling for a new event, for example, a user keystroke).  The error
is considered permanent and the work station taken off-line after a
set number of consecutive unsuccessful retries (retry count).  This
article describes how the retry count can be dynamic based on the
type of error and availability of resources within the WSC.  Also,
the WSC will insert added delays between retry attempts for certain
types of errors.

      Work stations can include dependent displays, programmable work
stations, and printers.  In some protocols, the work stations are
allowed to cable-through, where multiple work stations can be
attached to the same work station attachment cable.  A typical local
WSC with attached work stations is shown in the figure.  This
disclosure also applies to remote WSCs which attach to the host
processor via a communication line instead of a system bus.

      Retry counts for potentially recoverable transmission errors
were set to a specific value (for example, 16) in previous WSC
designs.  When a transmission error was detected, a retry was issued
immediately for commands or on the next poll cycle for polls
(assuming the retry count had not been exceeded).

      This disclosure applies to potentially recoverable WSC
transmission errors which are retried up to the retry count. This
disclosure contains three concepts where all three concepts could be
used concurrently or in combination.
CATEGORIZING ERROR TYPES AND USING DIFFERENT RETRY COUNTS

      When a potentially recoverable error occurs, the error will be
categorized into one of two or more categories: those which are
typically recoverable (for example, parity), those which are
typically unrecoverable (for example, time out), and possibly an
in-between category (or categories). Then, different retry counts
will be use...