Browse Prior Art Database

First Failure IPDS Error Recovery

IP.com Disclosure Number: IPCOM000110344D
Original Publication Date: 1992-Nov-01
Included in the Prior Art Database: 2005-Mar-25
Document File: 3 page(s) / 183K

Publishing Venue

IBM

Related People

Mastie, SD: AUTHOR

Abstract

Many Intelligent Printer Data Stream (IPDS) errors are asynchronous, while others occur on various pages in the printer pipeline (anywhere from received to committed). Printers present these errors to IPDS Print Drivers with a Negative Acknowledgement (NACK).

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 38% of the total text.

First Failure IPDS Error Recovery

       Many Intelligent Printer Data Stream (IPDS) errors are
asynchronous, while others occur on various pages in the printer
pipeline (anywhere from received to committed).  Printers present
these errors to IPDS Print Drivers with a Negative Acknowledgement
(NACK).

      In the past, print drivers have had to back up and rediscover
some NACKs, so that they could precisely correlate the logical page
in error with the IPDS page counter; this is known as Error Print
State.

      This was terrible for performance, both because of
retransmitting some amount of data to rediscover the error and
because when using Error Print State a positive Acknowledge (ACK) is
required after every page sent to the printer.

      This invention allows the use of Error Print State to be
avoided during error recovery for IPDS printers.

      Utilizing the existing correlation in IPDS between fields in
the NACK and information which is supplied on the Begin Page
structured field, which defines each IPDS page, a correlation
mechanism is defined which allows Error Manager to identify or figure
out the page in error, and specify accurate repositioning, without
needing to cause the NACK a second time, or back in the datastream
any further than is necessary.

      Previous IPDS Error-Management schemes required that the Error
Manager back up and rediscover any NACK which was reported on a page
which did not have an ACK attached to it.  Some PSFs avoided the need
for this backing up (better known as Error Print State (EPS)) by
ACKing after every page for channel-attached printers (notably
PSF/MVS), but this constant ACKing causes performance problems for
other IPDS Print Drivers.  This invention allows immediate recovery
from all NACKs without backing up for IPDS printers.

      Utilization of this invention can significantly speed up IPDS
NACK recovery and subsequent retransmission of the datastream during
error recovery.  A special exception is the IBM 3820 printer, which,
because of IPDS deviations, requires special handling, and is covered
separately below.

      Fundamental to this invention is a new and unique exploitation
of bytes 20-23 inclusive, in 24 sense byte IPDS printers.  As is
specified in IPDS, NACK formats 0, 1, and 7 supply a page identifier
when a NACK occurs on a particular page in the datastream, and this
page identifier is the exact 4-byte binary data that was supplied on
the Begin Page (BP) command for that page (henceforth, the token ID).
However, no token ID is supplied in some cases, including format-2
NACKs, 3-sense-byte printers, when a NACK occurs independent of any
particular page (such as out of paper) and for some asynchronous
NACKs.

      This lack of consistent information in bytes 20-23 has
mitigated its use by other print drivers.  However, this invention
exploits this information when it is available--which it is in the
vast majority of synchronous I...