WO1984001451A1 - Systeme ameliore de correction d'erreurs - Google Patents

Systeme ameliore de correction d'erreurs Download PDF

Info

Publication number
WO1984001451A1
WO1984001451A1 PCT/US1983/001573 US8301573W WO8401451A1 WO 1984001451 A1 WO1984001451 A1 WO 1984001451A1 US 8301573 W US8301573 W US 8301573W WO 8401451 A1 WO8401451 A1 WO 8401451A1
Authority
WO
WIPO (PCT)
Prior art keywords
sectors
data
sector
group
error correction
Prior art date
Application number
PCT/US1983/001573
Other languages
English (en)
Inventor
Peter N Yianilos
Gregory N Hullender
Original Assignee
Peter N Yianilos
Gregory N Hullender
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Peter N Yianilos, Gregory N Hullender filed Critical Peter N Yianilos
Publication of WO1984001451A1 publication Critical patent/WO1984001451A1/fr

Links

Classifications

    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
    • G11B20/10—Digital recording or reproducing
    • G11B20/18—Error detection or correction; Testing, e.g. of drop-outs

Definitions

  • This invention is for the digital encoding of information to provide improved error correction system.
  • the present invention introduces a practical scheme for correcting errors on the same scale at which they are detected.
  • the medium is divided into sectors for the pur ⁇ pose of error detection.
  • each sector contains sufficient additional information to correct all common errors and detect more than 99.99% of all possible errors.
  • the prior art corrections will appear trivial to one skilled in the art.
  • sectors are collected into groups for the purpose of error correction.
  • One sector in each group is reserved for error correction purposes.
  • the contents of this sector are computed from the other sectors ' in such a way that the contents of any sector in the group may be computed from the other sectors. This is most easily accomplished with the exclusive-or function.
  • This scheme greatly improves the situation with the floppy disk because most errors from handling and age tend to be localized to one or two sectors, but such errors are easily corrected with the invented system.
  • SUBST2TU7E SHEET error correction method for digital data on a recording medium that retrieves lost data in an entire sector of data recorded in sectors where the sectors are arranged in groups
  • a further object of this invention is to provide an error correction method where the data is recorded in sectors where the sectors are arranged in groups and where an error correction sector is provided for each group.
  • a further object of this invention is to provide an error correction method where the data is recorded in sectors where the sectors are arranged in groups and where an error correction sector is provided for each group and no group data is contained in two adjacent sectors whereby lost data is retrievable.
  • An additional object of this invention is to provide data in the error correction sector for one group from computation of data in other sectors of the same group in such a way that the contents of any sector in the group
  • * ' may be computed from the other sectors in the same group.
  • FIG. 1 illustrates generalized operating system software for input/output purposes.
  • Figure 2 illustrates the geometry of Xerox 820 disk with recovery sectoring.
  • Figure 3 illustrates the structure of internal buffers for Xerox 820.
  • Figure 4 illustrates the geometry of 1MB PC disk with recovery sectoring.
  • Figure 5 illustrates the geometry of IBM. PC disk with recovery sectoring and tracking.
  • Figure 6 illustrates the geometry of half-inch magnetic tape with recovery blocking.
  • Figure 7 illustrates record using XOR computation for correction.
  • FIG. 1 illustrates the basic structure of the I/O portion, the input/output portion
  • the file manager 12 is a high-level component of the operating system and handles a wide variety of mass- storage devices without any concern about their physical structure. There is only one file manager in any given operating system, and all mass-storage I/O requests go through it.
  • Device drivers 14, 14' are low-level components of the operating system which direct the physical process of Storing and retrieving data on the devices they serve. There is a different device driver for each type of mass- storage device, although one device driver usually handles all the devices of a particular type.
  • the invented system software represents a substan ⁇ tial improvement over the device drivers used in the prior - art; there is no change whatsoever in the file manager or the user software. It will be apparent to one skilled in the art that these device drivers may be implemented in either hardware or software.
  • the implementation on the Xerox 820 operates under the CP/M operating system and consists of a replacement of the device driver for the eight-inch floppy disks, shown in Figure 2.
  • This implementation illustrates the full power of the invention and is completely transparent to CP/M users and programmers. It is the preferred implementation. These are the physical specifications of the
  • Track zero 22 contains the bootstrap, and tracks one and two 24 are reserved for the operating system; the remainder of the disk is available for user data.
  • the protection of the invented software is not extended to the three system tracks because failure in the system tracks on any disk can be surmounted simply by booting the system from any underaged disk and then switching disks. This works because the system area is exactly the same on all disks.
  • the IBM PC implementation (below) describes a method for extending protection to the system area when it is necessary to do so.
  • Figure 2 illustrates the allocation of sectors in the area protected by the invented system. Odd numbered sectors comprise the odd group; even numbered sectors com- prise the even group. User data are stored in sectors stored in sectors 1-22; error correction information is stored in sectors 23 (for the odd group) and 24 (for the even group) .
  • the "DELETED" field can contain "NOT-DELETED,” “DEL 1,” or “DEL 2.” On a per- feet disk, all sectors are "NOT DELETED.”
  • the internal buffers contain all the data in all the sectors from a particular group.
  • the invented system first updates the current group according to value of the MODIFIED flag. If the MODIFIED flag is set, the data field of that sector is written with the data in the buffer.
  • the ECC sector differs slightly from the rest in that it is always recom ⁇ puted before it is written to guarantee that it is up to date.
  • the system will retry up to ten times, after which the sector is con ⁇ sidered defective and is deleted. This is what the DELETED field in the sector preamble is for, but the preamble of a damaged sector is not a safe place to put such information. Accordingly, when a sector is deleted, a "one" is written in the DELETED field of the sector 14 sectors ahead on the disk and a "two" is written in the DELETED field of the sector 21
  • the system seeks to the appropriate track, reads the data from all the sectors in the group into the internal buffers, and sets all the MODIFIED flags to zero.
  • the data in the inter- nal buffer corresponding to the ECC sector is checked by
  • One 5-1/4 inch disk drive IBM 3740 format. 40 tracks per disk (single-sided) , numbered 0 to 39.
  • Track zero contains the bootstrap, the data are stored in tracks one through 24, and the program/operating system is stored in tracks 25 through 33.
  • Figure 4 illustrates the allocation of sectors in the area protected by the invented system. As in the
  • odd numbered sectors comprise the odd group; even numbered sectors comprise the even group.
  • Data are stored in sectors 1-8; error correction information is stored in sectors 9 (for the odd group) and 10 (for the even group) .
  • the "DELETED" field in the sector preamble is not used.
  • the principle of the error recovery plan is basically the same as with the Xerox 820; read errors on data sectors are recovered by XORing the surviving data sectors with the ECC sector.
  • the principal difference is that the IBM PC implementation does not allow the disk to be written; a special manufacturing program that is not distributed produces the disks which are sent to the customer and never modified.
  • the bootstrap is a short program which resides in sector one of track zero and which is read in automatically whenever the IBM PC is star ⁇ ted up. Because the boot block is read in by the ROMs already in the IBM PC, the above described inventive concepts cannot be used for this purpose. However, these ROMs will detect failure, using the methods of the prior art, and will attempt three times to read the boot block before giving up. Therefore, the manufacturing program writes multiple copies of the bootstrap in track zero and numbers then all sector one. Since the ROMs find sector one by reading the sector header, this technique causes the ROMs to try reading three different copies of the boot block instead of trying to read the same one three times. This system will always recover from any single error in the boot track, so it offers pro ⁇ tection as good as the invented system, although it wastes the entire boot track. However, the boot track is reserved
  • the bootstrap reads in the main program and oper ⁇ ating system using the invented system to correct errors, and the main program uses the invented system to read its data.
  • the invented system Upon receipt of a read request, the invented system seeks to the appropriate track and attempts to read the requested sector into an internal buffer. If that attempt succeeds, no further processing is done. If that attempt fails, the invented system attempts to read all the remaining sectors in the same group as the failing sector and computes their XOR into a second buffer. If it succeeds in this, the second buffer contains the data that was in the failing sector and the invented system returns that to the file manager. Because the sectors in each group alternate about the disk, the invented system can easily accomplish this while the disk rotates only once. If it fails to read any of the other sectors in the group, the invented system records which sectors it still needs to read.
  • the disk rotated around to the original sector again and the invented system tries again to read it. If that succeeds this time, it simply returns that data. If it fails again, the inven ⁇ ted system goes around the disk again and tries to read the sectors that failed the time before. If ten spins of the disk do not suffice, the system performs a disk reset and repeats the cycle. After ten disk resets, the system admits defeat and reports to the user that the- disk is defective. The immense number of retrys are justified by the
  • the invented system divides the sectors into two groups 40 as shown in Figure 5 the same as before.
  • it divides the tracks (excluding track zero) into three groups 42 of thir ⁇ teen tracks, called zero, one, and two, based on the remain— der obtained when dividing the track number by three as shown in Figure 5.
  • the innermost three tracks, (Nos. 37, 38, and 39) are used for error correction by writing track 37 with the logical XOR of the one group, track 38 with the XOR of the two group, and track 39 with the XOR of the three group.
  • S ⁇ _337.7 ⁇ 7_i SHEET will be apparent to one skilled in the art that the specific numbers in this section may be modified considerably.
  • ) Write the record number, starting with zero at the beginning of the tape, in the first two bytes and write the data in the other 2048. Collect the records into groups of sixteen, using the data field of the sixteenth record to contain the XOR of the other fifteen. Finally, interleave the groups such that each successive group element follows the previous one by sixteen records (see Figure 5) . On rereading the tape, -the invented system keeps a running XOR for each group, starting -over again every 256 records. Errors are detected by prior art methods. Any error of less than sixteen consecutive records out of every 256 is easily recovered by spacing over the bad section.
  • the invented system can tell exactly which records were missed and make a note of them. If any lost record belongs to a group that has already lost a record, no recovery is possible. When the 256 record set is complete, the contents of the lost records will be stored in their respective buffers.

Abstract

Procédé amélioré de stockage et de récupération de données numériques. Des codes de détection d'erreurs sont utilisés dans l'art antérieur pour identifier des secteurs défaillants. Dans ce procédé amélioré, les secteurs sont recueillis en groupes et un secteur par groupe est réservé aux informations de corrections d'erreurs. Les informations dans ce secteur sont calculées de telle sorte que les informations dans tout secteur du groupe peuvent être calculées à partir des autres secteurs. Tout secteur défaillant peut ainsi être récupéré. Afin de réduire les probabilités de perdre deux secteurs du même groupe, les groupes sont répartis de manière qu'aucun groupe ne contienne deux secteurs adjacents.
PCT/US1983/001573 1982-10-06 1983-10-06 Systeme ameliore de correction d'erreurs WO1984001451A1 (fr)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US43307182A 1982-10-06 1982-10-06

Publications (1)

Publication Number Publication Date
WO1984001451A1 true WO1984001451A1 (fr) 1984-04-12

Family

ID=23718746

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1983/001573 WO1984001451A1 (fr) 1982-10-06 1983-10-06 Systeme ameliore de correction d'erreurs

Country Status (2)

Country Link
EP (1) EP0121565A1 (fr)
WO (1) WO1984001451A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1989010594A1 (fr) * 1988-04-22 1989-11-02 Amdahl Corporation Systeme de fichiers pour une pluralite de types de stockage

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4209809A (en) * 1978-09-11 1980-06-24 International Business Machines Corporation Apparatus and method for record reorientation following error detection in a data storage subsystem
US4211997A (en) * 1978-11-03 1980-07-08 Ampex Corporation Method and apparatus employing an improved format for recording and reproducing digital audio
US4238852A (en) * 1978-04-17 1980-12-09 Sony Corporation Error correcting system
US4306305A (en) * 1978-10-23 1981-12-15 Sony Corporation PCM Signal transmitting system with error detecting and correcting capability
US4380071A (en) * 1981-02-02 1983-04-12 Sony Corporation Method and apparatus for preventing errors in PCM signal processing apparatus
US4393502A (en) * 1979-10-09 1983-07-12 Masato Tanaka Method and apparatus for communicating digital information words by error-correction encoding

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4238852A (en) * 1978-04-17 1980-12-09 Sony Corporation Error correcting system
US4209809A (en) * 1978-09-11 1980-06-24 International Business Machines Corporation Apparatus and method for record reorientation following error detection in a data storage subsystem
US4306305A (en) * 1978-10-23 1981-12-15 Sony Corporation PCM Signal transmitting system with error detecting and correcting capability
US4211997A (en) * 1978-11-03 1980-07-08 Ampex Corporation Method and apparatus employing an improved format for recording and reproducing digital audio
US4393502A (en) * 1979-10-09 1983-07-12 Masato Tanaka Method and apparatus for communicating digital information words by error-correction encoding
US4380071A (en) * 1981-02-02 1983-04-12 Sony Corporation Method and apparatus for preventing errors in PCM signal processing apparatus

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1989010594A1 (fr) * 1988-04-22 1989-11-02 Amdahl Corporation Systeme de fichiers pour une pluralite de types de stockage
US4993030A (en) * 1988-04-22 1991-02-12 Amdahl Corporation File system for a plurality of storage classes

Also Published As

Publication number Publication date
EP0121565A1 (fr) 1984-10-17

Similar Documents

Publication Publication Date Title
JP3405712B2 (ja) データ回復方法およびデータ記憶装置
US6367047B1 (en) Multi-level error detection and correction technique for data storage recording device
JP2514289B2 (ja) デ―タの修復方法およびシステム
EP0466296B1 (fr) Canal pour la régénération de données dans un réseau d'entraînement de disques tolérant des fautes et méthode pour y corriger les erreurs
US6532548B1 (en) System and method for handling temporary errors on a redundant array of independent tapes (RAIT)
US6728922B1 (en) Dynamic data space
US5404361A (en) Method and apparatus for ensuring data integrity in a dynamically mapped data storage subsystem
US4698810A (en) Data recording and reproducing system with error correction capability using ECC and CRC codes
US5757824A (en) Code error correction apparatus
US7464322B2 (en) System and method for detecting write errors in a storage device
JP4709485B2 (ja) オンドライブ統合化セクタ・フォーマットｒａｉｄの誤り訂正符号システムおよび方法
US5737344A (en) Digital data storage with increased robustness against data loss
JPS62164278A (ja) ２次記憶サブシステムのためのデイスクフオ−マツト
JPH0773602A (ja) 光ディスク装置
US6349359B1 (en) Method and apparatus for maintaining data consistency in raid
WO1993018589A1 (fr) Recuperation de donnees apres une defaillance de correction d'erreur
JPH0612895B2 (ja) 情報処理システム
US20040036997A1 (en) Method for writing streaming audiovisual data to a disk drive
US7386754B2 (en) Method and apparatus to improve magnetic disc drive reliability using excess un-utilized capacity
JP2002528839A (ja) データ記憶装置のための書換え方法及び装置
JPS63177219A (ja) データ転装装置
WO1984001451A1 (fr) Systeme ameliore de correction d'erreurs
JP3993921B2 (ja) 記憶装置
EP1069569A2 (fr) Recouvrement d'un paquet de données par superposition de mot de code
JPS6139286A (ja) デ−タ管理方法

Legal Events

Date Code Title Description
AK Designated states

Designated state(s): JP

AL Designated countries for regional patents

Designated state(s): AT BE CH DE FR GB LU NL SE