Browse Prior Art Database

Storage Device Error Detection and Fault/FRU Isolation via Concurrent Longitudinal Redundancy Codes Checking

IP.com Disclosure Number: IPCOM000118820D
Original Publication Date: 1997-Jul-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 4 page(s) / 101K

Publishing Venue

IBM

Related People

Lamb, K: AUTHOR [+2]

Abstract

General computing structures are shown in Fig. 1, which are commonly used in computer and data communication applications to detect (and/or correct) physical failures which might occur in hardware. From several sources, information packets can be written into a storage, which may be read out to several targets at some later time. There are different protection schemes for the storage: o The one shown in a) portion of Fig. 1 uses parity to detect errors in the data bits of a stored data word. o In b), Error Correcting Codes (ECC) are used, which allows to detect and correct data bit errors but requires a higher implementation cost than a).

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

Storage Device Error Detection and Fault/FRU Isolation via Concurrent
Longitudinal Redundancy Codes Checking

      General computing structures are shown in Fig. 1, which are
commonly used in computer and data communication applications to
detect (and/or correct) physical failures which might occur in
hardware. From  several sources, information packets can be written
into a storage, which may be read out to several targets at some
later time.  There are  different protection schemes for the storage:
  o  The one shown in a) portion of Fig. 1 uses parity to detect
      errors in the data bits of a stored data word.
  o  In b), Error Correcting Codes (ECC) are used, which allows
      to detect and correct data bit errors but requires a higher
      implementation cost than a).  Typical for both schemes is
      that they protect from data bit errors within a word of
      the storage, but it has neither correction nor detection
      capabilities for physical faults in the Rd Addr logic,
      the Wr Addr logic, or the write control.
  o  In c), a scheme is shown which is typical for data
      communication applications, where data packets are
      protected by Longitudinal Redundancy Codes (LRC).  The
      LRC is calculated over all data words from data word 0
      to data word l.
     Thus, it allows detection failures between data words,
      but the drawback of that approach is that an error could
      be detected only after all data has been read out and a
      LRC-check could be done.  But, in many computer applications,
      part of the data has been already sent to the target (and is
      used by the target), before the LRC-checking could be done.
 ...