[go: up one dir, main page]

JP2000112835A - Memory device, data recording / reproducing device, and computer system - Google Patents

Memory device, data recording / reproducing device, and computer system

Info

Publication number
JP2000112835A
JP2000112835A JP10280316A JP28031698A JP2000112835A JP 2000112835 A JP2000112835 A JP 2000112835A JP 10280316 A JP10280316 A JP 10280316A JP 28031698 A JP28031698 A JP 28031698A JP 2000112835 A JP2000112835 A JP 2000112835A
Authority
JP
Japan
Prior art keywords
memory
data
parity
array
recording medium
Prior art date
Legal status (The legal status 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 status listed.)
Withdrawn
Application number
JP10280316A
Other languages
Japanese (ja)
Inventor
Satoyuki Ono
智行 小野
Atsushi Miyauchi
敦 宮内
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sony Corp
Original Assignee
Sony Corp
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 Sony Corp filed Critical Sony Corp
Priority to JP10280316A priority Critical patent/JP2000112835A/en
Publication of JP2000112835A publication Critical patent/JP2000112835A/en
Withdrawn legal-status Critical Current

Links

Landscapes

  • Techniques For Improving Reliability Of Storages (AREA)

Abstract

(57)【要約】 【課題】 データバッファメモリのリスト管理等の手間
を減らすことができ、更にHDDが故障してデータが正
しく読み出せない場合であっても、データバッファメモ
リに対して通常時と等価なアクセスを行うことを可能に
する。 【解決手段】 インタークネクトI/F回路11と、デ
ータメモリ14D及びパリティメモリ14Pに分けられ
たメモリアレイ14と、アレイ14を複数の個々のアド
レス空間にマッピングするアドレスマッピング設定部1
3と、その設定に基づいてアレイ14のアクセスを制御
するメモリアクセス制御部12と、データとパリティの
比を設定するデータ・パリティ比設定部15と、その比
に基づいてパリティを計算するパリティ演算部16を設
け、メモリ14Dへのデータライト後にパリティをメモ
リ14Pにライトアクセスする。更に、パリティメモリ
マッピング設定部17を設け、データ再構築時の効率を
改善する。
(57) [Summary] [PROBLEMS] To reduce the trouble of managing a list of a data buffer memory, etc., and even in a case where an HDD fails and data cannot be read correctly, the data buffer memory is normally read. Access that is equivalent to SOLUTION: An interconnect I / F circuit 11, a memory array 14 divided into a data memory 14D and a parity memory 14P, and an address mapping setting unit 1 for mapping the array 14 into a plurality of individual address spaces.
3, a memory access control unit 12 for controlling access to the array 14 based on the setting, a data / parity ratio setting unit 15 for setting a data-parity ratio, and a parity operation for calculating a parity based on the ratio. A unit 16 is provided to write-access the parity to the memory 14P after writing data to the memory 14D. Further, a parity memory mapping setting unit 17 is provided to improve the efficiency at the time of data reconstruction.

Description

【発明の詳細な説明】DETAILED DESCRIPTION OF THE INVENTION

【0001】[0001]

【発明の属する技術分野】本発明は、例えば複数のHD
D(ハードディスクドライブ)によって並列冗長構成さ
れたディスクアレイ装置等に好適なメモリ装置、そのメ
モリ装置を使用したデータ記録再生装置及びコンピュー
タシステムに関する。
[0001] The present invention relates to, for example, a plurality of HDs.
The present invention relates to a memory device suitable for a disk array device or the like configured in parallel with D (hard disk drive), a data recording / reproducing device using the memory device, and a computer system.

【0002】[0002]

【従来の技術】従来より、ディスクI/Oの高速化及び
信頼性向上のための技術として、例えば複数のHDDに
よって並列冗長構成されたディスクアレイ装置が知られ
ている。
2. Description of the Related Art Conventionally, as a technique for improving the speed and reliability of a disk I / O, for example, a disk array device having a plurality of HDDs in a parallel redundant configuration has been known.

【0003】このディスクアレイ装置を構成方法によっ
て体系的に分類したのが、カリフォルニア大学バークレ
ー校のパターソンらによる論文”A Case for Redundant
Arrays of Inexpensive Disks(RAID)[csd-87-391]”で
ある。すなわち、この論文によれば、1から5のRAI
Dレベルが定義されている。
[0003] A systematic classification of the disk array devices according to the configuration method is described in a paper "A Case for Redundant" by Patterson et al. Of the University of California, Berkeley.
Arrays of Inexpensive Disks (RAID) [csd-87-391] ”. That is, according to this paper, RAI of 1 to 5
D level is defined.

【0004】RAID−1は、2台のディスクを組み合
わせることにより、ミラーリングを可能にし、当該ミラ
ーリングにより2台のディスクに同じものを書き、片方
に障害が発生しても処理の続行を可能にする。一方のデ
ィスクに障害が発生しても、他方のディスクには正常な
データが残っている。
[0004] RAID-1 enables mirroring by combining two disks, writes the same data on two disks by the mirroring, and enables continuation of processing even if one of the disks fails. . Even if one disk fails, normal data remains on the other disk.

【0005】RAID−2は、N台のデータディスクに
対し、ハミングコードによるECCデータを別のディス
クに記録しておき、ディスクの障害だけでなくデータに
誤りがあった場合でも訂正できるようにしている。
[0005] In RAID-2, ECC data based on a Hamming code is recorded on another N disks, so that not only disk failures but also data errors can be corrected. I have.

【0006】RAID−3は、ミラーリングで完全に二
重化する代わりに、N台のデータディスクに対し、1台
のパリティディスクを持つことにより、1台のドライブ
に障害が発生しても処理の続行を可能にしたものであ
る。1ビット若しくは1バイト単位でストライピングを
行う。データディスクのうち1台が故障しても、他のド
ライブからパリティを逆合成して、元のデータを作り出
すことにより、データを失うことなく処理が続行でき
る。もちろん、パリティディスクが故障しても、データ
ディスクが正常であれば、データアクセスには問題な
い。但し、RAID−5より構成が単純であるが、書き
込みにおいてパリティディスクにアクセスが集中するな
ど、アクセスの対称性に欠ける。
[0006] RAID-3 has one parity disk for N data disks instead of completely duplexing by mirroring, so that the processing can be continued even if one drive fails. It is made possible. Striping is performed in units of 1 bit or 1 byte. Even if one of the data disks fails, the parity can be decomposed from the other drives to generate the original data, so that the processing can be continued without losing the data. Of course, even if the parity disk fails, there is no problem in data access as long as the data disk is normal. However, although the configuration is simpler than that of RAID-5, it lacks access symmetry, for example, access concentrates on the parity disk in writing.

【0007】RAID−4は、RAID−3で小さかっ
たストライピングの単位をセクタ若しくはそれ以上にし
たものであり、このRAID−4もまとめて、RAID
−3と呼ぶことが多く、RAID−4という表現はあま
り使われない。
In RAID-4, the striping unit smaller than that of RAID-3 is set to a sector or more, and this RAID-4 is also collectively referred to as RAID-4.
-3, and the expression RAID-4 is rarely used.

【0008】RAID−5は、N台+1のデータディス
クに対し、N台のデータと1台のパリティを固定的に割
り当てるのではなく、等分にパリティを分散させる(各
ディスクにパリティを割り当てる)ことにより、2ヶ所
以上の同時書き換えを可能にしたものである。RAID
−3でパリティディスクに集中するディスクアクセス
を、パリティディスクを分散させることにより、ディス
クアクセスの分散を行ってスループットの向上を目指し
ている。
In RAID-5, N data and one parity are not fixedly allocated to N + 1 data disks, but parity is equally distributed (parity is allocated to each disk). This enables simultaneous rewriting at two or more locations. RAID
In -3, the disk access is concentrated on the parity disk, and the parity disk is distributed so that the disk access is distributed to improve the throughput.

【0009】上述したRAIDを実装する場合、様々な
実装方法が存在するが、ハードウェアで支援する場合に
は、パリティメモリが用いられる。このパリティメモリ
については、特開平10−105476号公報にパリテ
ィブロックの生成装置が開示されている。すなわち、こ
の公報記載のパリティブロックの生成装置は、複数のデ
ータワードからパリティデータを生成するものであり、
第1のパリティメモリと、第2のパリティメモリと、前
記複数のデータワードを受信するように構成され、前記
第1のパリティメモリ及び第2のパリティメモリに結合
され、前記パリティデータを生成するためのパリティ生
成データワードをそこに選択的に記憶し、またそこから
選択的にロードするためのパリティ・制御ユニット(デ
ィスクアレイコントローラ)とからなるものである。
[0009] There are various mounting methods for implementing the above-mentioned RAID, but a parity memory is used when hardware is supported. With respect to this parity memory, an apparatus for generating a parity block is disclosed in Japanese Patent Application Laid-Open No. 10-105476. That is, the parity block generation device described in this publication generates parity data from a plurality of data words.
A first parity memory, a second parity memory, and a memory configured to receive the plurality of data words and coupled to the first parity memory and the second parity memory to generate the parity data And a parity and control unit (disk array controller) for selectively storing therein and selectively loading therefrom.

【0010】[0010]

【発明が解決しようとする課題】ここで、代表してRA
ID−3を適用し、ディスクアレイを構成するHDDの
ドライブ数をN+1台とした場合を例に挙げて、従来の
ディスクアレイ装置のディスクアレイコントローラによ
る制御動作の流れを説明する。なお、このRAID−3
のディスクアレイ装置は、図示は省略するが、その基本
構成として、少なくとも、入力/出力用の単一のバッフ
ァメモリと、全体を制御するディスクアレイコントロー
ラと、複数のHDDと、これら複数のHDDとの間でデ
ータの送受を制御するHDDI/Fコントローラと、H
DDの書き込み/読み出しデータ用のバッファメモリ
と、パリティデータ用のパリティメモリとを有する。
Here, RA is representatively used.
The flow of the control operation by the disk array controller of the conventional disk array device will be described by taking, as an example, the case where ID-3 is applied and the number of HDDs constituting the disk array is N + 1. This RAID-3
Although not shown, the basic configuration of the disk array device is at least a single buffer memory for input / output, a disk array controller for controlling the whole, a plurality of HDDs, and a plurality of HDDs. An HDD I / F controller for controlling data transmission and reception between the
It has a buffer memory for DD write / read data and a parity memory for parity data.

【0011】図18には、当該RAID−3の従来のデ
ィスクアレイ装置における書き込み時の動作のフローチ
ャートを示す。
FIG. 18 shows a flowchart of the write operation in the conventional RAID-3 disk array device.

【0012】この図18において、先ず、ディスクアレ
イコントローラは、ステップS101として、単一のバ
ッファメモリに入力されて蓄積されている入力データ
を、書き込み要求に応じて、ストライプユニットサイズ
のN個のデータに分割(ストライピング)して読み出
す。
In FIG. 18, first, in step S101, the disk array controller converts the input data input to and accumulated in a single buffer memory into N data of a stripe unit size in response to a write request. And read them out.

【0013】また、ディスクアレイコントローラは、ス
テップS102として、パリティメモリ上にストライプ
ユニットサイズのパリティウィンドウを作成すると共
に、ステップS103として、N個のデータを書き込む
ためのデータバッファメモリとパリティメモリ上のパリ
ティウィンドウとを対応付ける。
In step S102, the disk array controller creates a stripe window of a stripe unit size in the parity memory. In step S103, the disk array controller creates a parity buffer in the data buffer memory and the parity memory for writing N data. Associate with a window.

【0014】次に、ディスクアレイコントローラは、ス
テップS104として、N個のデータをデータバッファ
メモリに書き込むと同時に、パリティデータを生成して
パリティメモリ上のパリティウィンドウに書き込む。
Next, at step S104, the disk array controller writes the N pieces of data into the data buffer memory and, at the same time, generates parity data and writes it into the parity window on the parity memory.

【0015】その後、ディスクアレイコントローラは、
ステップS105として、データバッファメモリからデ
ータを読み出させると共に、パリティメモリのパリティ
ウィンドウからパリティデータを取り出し、HDDI/
Fコントローラに送る。そして、HDDI/Fコントロ
ーラは、データバッファメモリからのN個のデータ及び
パリティメモリからのパリティデータを、各HDDに書
き込む。
Thereafter, the disk array controller:
In step S105, the data is read from the data buffer memory, the parity data is extracted from the parity window of the parity memory, and the HDDI /
Send to F controller. Then, the HDD I / F controller writes the N data from the data buffer memory and the parity data from the parity memory to each HDD.

【0016】上述したように、RAID−3の従来のデ
ィスクアレイ装置において書き込みを行う場合、パリテ
ィメモリを使用してパリティを計算するために、ディス
クアレイコントローラは、書き込み要求のあったデータ
を1/Nに分割し、パリティメモリと連動して動作する
データバッファメモリにデータを書き込む必要がある。
したがって、N個のデータバッファを管理するためのデ
ータ構造が必要となっている。
As described above, when data is written in a conventional disk array device of RAID-3, the disk array controller calculates the parity using the parity memory, so that the data requested to be written is divided by 1 /. It is necessary to divide the data into N and write data to a data buffer memory that operates in conjunction with the parity memory.
Therefore, a data structure for managing N data buffers is required.

【0017】次に、図19には、RAID−3の従来の
ディスクアレイ装置における読み出し時の動作のフロー
チャートを示す。
FIG. 19 is a flowchart showing a read operation in the conventional RAID-3 disk array device.

【0018】この図19において、HDDI/Fコント
ローラは、ステップS111として、N個のデータが書
き込まれている各HDDからそれらN個のデータを読み
出すと共に、パリティデータが書き込まれているHDD
から当該パリティデータを読み出す。
In FIG. 19, in step S111, the HDD I / F controller reads the N data from each of the HDDs to which the N data has been written and also writes the parity data to the HDD.
From the parity data.

【0019】次に、ディスクアレイコントローラでは、
ステップS111として、これらN個のデータをデータ
バッファメモリに読み出す(書き込む)と同時に、パリ
ティメモリのパリティウィンドウにおいてパリティデー
タを用いた排他的論理和(XOR)演算を実行する。
Next, in the disk array controller,
In step S111, at the same time as reading (writing) these N pieces of data into the data buffer memory, an exclusive OR (XOR) operation using the parity data is executed in the parity window of the parity memory.

【0020】ここで、ディスクアレイコントローラは、
ステップS113として、HDDが壊れているか否かの
判定を行い、壊れていると判定した時はステップS11
4に、壊れていないときはステップS115に処理を進
める。
Here, the disk array controller is:
In step S113, it is determined whether or not the HDD is broken.
If it is not broken, the process proceeds to step S115.

【0021】ディスクアレイコントローラは、ステップ
S113にてHDDが1台壊れていると判定したとき、
ステップS114にて壊れていないHDDからのN−1
個のデータとパリティデータを単一バッファメモリに読
み出し(書き込み)、データを再構築する。
When the disk array controller determines that one HDD is broken in step S113,
N-1 from the HDD not broken in step S114
The data and parity data are read (written) into a single buffer memory, and the data is reconstructed.

【0022】一方、ディスクアレイコントローラは、ス
テップS113にてHDDが壊れていないと判定したと
き、ステップS115にてそれら壊れていないHDDか
らのN個のデータを単一バッファに読み出して(書き込
んで)データを再構築する。
On the other hand, when the disk array controller determines in step S113 that the HDDs are not damaged, the disk array controller reads (writes) the N data from the non-damaged HDDs into a single buffer in step S115. Reconstruct the data.

【0023】この図19にて説明したように、RAID
−3の従来のディスクアレイ装置においてデータを読み
出す場合、書き込み時と同様に、ディスクアレイコント
ローラは、N+1台のHDDから読み出したデータをパ
リティメモリを含むN+1個のデータバッファメモリに
読み出した(書き込んだ)後に、N個のデータバッファ
メモリを管理しているデータ構造を利用してデータを一
つに纏めることになる。
As described with reference to FIG.
When reading data in the conventional disk array device of No.-3, the disk array controller reads data read from N + 1 HDDs into N + 1 data buffer memories including a parity memory in the same manner as when writing data. ) Later, the data is combined into one using the data structure managing the N data buffer memories.

【0024】更に、HDDの内の例えば1台が故障した
ような場合、パリティメモリに保持されたパリティデー
タを用いることで、失われたデータを回復することはで
きるが、このときのディスクアレイコントローラは、故
障したHDDからのデータが入っているデータバッファ
メモリの代わりに、対応するパリティメモリへ毎回アク
セスする必要があり、したがって、故障していない場合
の読み出し動作に加えて、データバッファメモリを管理
しているデータ構造を変更する必要がある。
Further, when, for example, one of the HDDs fails, the lost data can be recovered by using the parity data held in the parity memory. Needs to access the corresponding parity memory every time instead of the data buffer memory that contains the data from the failed HDD, and therefore manages the data buffer memory in addition to the read operation when there is no failure. You need to change the data structure you are using.

【0025】そこで、本発明はこのような状況に鑑みて
なされたものであり、データバッファメモリのリスト管
理等の手間を減らすことができ、更にHDDの故障など
でデータが正しく読み出せない場合であっても、パリテ
ィメモリに対して通常時と等価なアクセスを行うことが
可能なメモリ装置と、そのメモリ装置を使用したデータ
記録再生装置及びコンピュータシステムを提供すること
を目的とする。
Therefore, the present invention has been made in view of such a situation, and can reduce the trouble of managing the data buffer memory list and the like. It is an object of the present invention to provide a memory device capable of performing equivalent access to a parity memory as usual, and a data recording / reproducing device and a computer system using the memory device.

【0026】[0026]

【課題を解決するための手段】本発明のメモリ装置は、
単一又は複数のインターコネクトに接続されるインター
コネクト接続手段と、物理メモリであるメモリアレイ
と、メモリアレイを複数の個々のアドレス空間にマッピ
ングするマッピング設定手段とマッピングの設定に基づ
いてメモリアレイのアクセスを制御するメモリアクセス
制御手段とを有することにより、上述した課題を解決す
る。
A memory device according to the present invention comprises:
Interconnect connection means connected to one or more interconnects, a memory array as a physical memory, mapping setting means for mapping the memory array to a plurality of individual address spaces, and access to the memory array based on the setting of the mapping. By having the memory access control means for controlling, the above-mentioned problem is solved.

【0027】また、本発明のメモリ装置は、単一又は複
数のインターコネクトに接続されるインターコネクト接
続手段と、物理メモリであるメモリアレイと、メモリア
レイの一部からなるサブメモリアレイ若しくは別個のサ
ブメモリアレイを使用し、設定によりメモリアレイに対
する任意の単位アクセスを対応するサブメモリアレイへ
のアクセスに変換するサブメモリアレイマッピング設定
手段とを有することにより、上述した課題を解決する。
Further, the memory device of the present invention comprises an interconnect connection means connected to a single or a plurality of interconnects, a memory array as a physical memory, and a sub memory array formed of a part of the memory array or a separate sub memory. The above-mentioned problem is solved by using an array and having a sub memory array mapping setting means for converting an arbitrary unit access to the memory array into an access to a corresponding sub memory array by setting.

【0028】ここで、本発明のメモリ装置は、メモリア
レイをデータメモリとパリティメモリに分けて使用、或
いは、メモリアレイの一部からなるサブメモリアレイ若
しくは別個のサブメモリアレイをパリティメモリとして
使用し、データとパリティの比を設定するデータ・パリ
ティ比設定手段と、そのデータとパリティの比の設定に
基づいて一のアドレス空間からの任意の連続したN個の
単位ライトアクセスに対してパリティを計算するパリテ
ィ演算手段とを設け、データメモリへのデータのライト
後に、パリティをパリティメモリに対してライトアクセ
スすることにより、上述した課題を解決する。
Here, the memory device of the present invention uses the memory array by dividing it into a data memory and a parity memory, or uses a sub memory array that is a part of the memory array or a separate sub memory array as a parity memory. Data / parity ratio setting means for setting a data / parity ratio, and calculating parity for any continuous N unit write accesses from one address space based on the setting of the data / parity ratio The above-described problem is solved by providing a parity operation unit that performs write access to the parity memory after writing data to the data memory.

【0029】次に、本発明のデータ記録再生装置は、本
発明のメモリ装置と、複数の記録媒体からなる記録媒体
アレイと、記録媒体アレイを制御する記録媒体アレイ制
御手段と、記録媒体アレイと記録媒体アレイ制御手段と
の間のデータ送受を制御する記録媒体アレイインターフ
ェイス手段とを有し、記録媒体アレイ制御手段と記録媒
体アレイインターフェイス手段とが別々のアドレス空間
を使用することにより、上述した課題を解決する。
Next, a data recording / reproducing apparatus according to the present invention comprises a memory device according to the present invention, a recording medium array including a plurality of recording media, recording medium array control means for controlling the recording medium array, and a recording medium array. The recording medium array interface means for controlling data transmission and reception between the recording medium array control means and the recording medium array control means, wherein the recording medium array control means and the recording medium array interface means use different address spaces; Solve.

【0030】さらに、本発明のコンピュータシステム
は、本発明のメモリ装置と、本発明のデータ記録再生装
置とを有し、メモリ装置及びデータ記録再生装置をイン
ターコネクトを介して接続してなることにより、上述し
た課題を解決する。
Further, a computer system according to the present invention includes the memory device according to the present invention and the data recording / reproducing device according to the present invention, and is configured by connecting the memory device and the data recording / reproducing device via an interconnect. The above-mentioned problem is solved.

【0031】[0031]

【発明の実施の形態】本発明の好ましい実施の形態につ
いて、図面を参照しながら説明する。
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described with reference to the drawings.

【0032】図1には、本発明実施の形態のメモリ装置
1が適用されるシステムの第1の構成例を示す。
FIG. 1 shows a first configuration example of a system to which the memory device 1 according to the embodiment of the present invention is applied.

【0033】この図1に示すシステムは、CPU2、ホ
スト/PCI(Peripheral Component Interconnect)
ブリッジ部3、メインメモリ4、入出力装置5等を有す
る通常のコンピュータに組み込まれている構成と、ディ
スクアレイコントローラ6、メモリ装置1、HDDI/
Fコントローラ7及び複数のHDD8等を有するデータ
記録再生装置であるディスクアレイ装置の構成とが、そ
れぞれPCIバスで接続されてなるコンピュータシステ
ムである。
The system shown in FIG. 1 has a CPU 2, a host / PCI (Peripheral Component Interconnect).
A configuration incorporated in a normal computer having a bridge unit 3, a main memory 4, an input / output device 5, etc., a disk array controller 6, a memory device 1, an HDDI /
The configuration of a disk array device, which is a data recording / reproducing device having an F controller 7 and a plurality of HDDs 8 and the like, is a computer system in which each is connected by a PCI bus.

【0034】CPU2はシステム全体の動作を制御し、
ホスト/PCIブリッジ部3はPCIバスとCPU2の
間のデータ送受の受け渡しを行うと共に、メインメモリ
4を制御するメモリコントローラの機能も含む。入出力
装置5は、キーボードやマウス等の操作入力装置や、デ
ィスプレイ等の表示装置、外部機器と接続される各種コ
ネクタ等を含む。
The CPU 2 controls the operation of the entire system,
The host / PCI bridge unit 3 exchanges data between the PCI bus and the CPU 2 and also includes a function of a memory controller that controls the main memory 4. The input / output device 5 includes an operation input device such as a keyboard and a mouse, a display device such as a display, various connectors connected to external devices, and the like.

【0035】CPU等にて構成されるディスクアレイコ
ントローラ6は、PCIブリッジ部を含み、ホストコン
ピュータから受け取ったデータを、後述するメモリ装置
1及びHDDI/Fコントローラ7を介して複数の各H
DD8にデータを書き込ませたり、またHDD8から読
み出されてHDDI/Fコンピュータ7及びメモリ装置
1を介したデータをホストコンピュータに送る。
The disk array controller 6 including a CPU and the like includes a PCI bridge unit, and transfers data received from a host computer to a plurality of H devices via a memory device 1 and an HDD I / F controller 7 described later.
Data is written to the DD 8, and data read from the HDD 8 and transmitted through the HDD I / F computer 7 and the memory device 1 is sent to the host computer.

【0036】HDDI/Fコントローラ7は、複数のH
DD8の管理及びデータの記録/再生を制御し、ディス
クアレイコントローラ6やメモリ装置1との間でデータ
の送受を行う。
The HDD I / F controller 7 has a plurality of H
It controls the management of the DD 8 and the recording / reproduction of data, and sends and receives data to and from the disk array controller 6 and the memory device 1.

【0037】メモリ装置1は本発明の主要部であり、そ
の構成及び動作を以下に説明する。
The memory device 1 is a main part of the present invention, and its configuration and operation will be described below.

【0038】図2には、本発明の第1の実施の形態のメ
モリ装置1の概略構成を示す。
FIG. 2 shows a schematic configuration of the memory device 1 according to the first embodiment of the present invention.

【0039】この第1の実施の形態のメモリ装置1は、
インターコネクトI/F回路11を介して単一または複
数のインターコネクトに接続されるものであって、複数
のアドレス空間にマッピング可能であり、物理メモリを
個々のアドレス空間にどのようにマッピングするかを設
定できるアドレスマッピング設定部13と、当該アドレ
スマッピング設定部13による設定に基づいて動作する
メモリアクセス制御部12と、物理メモリであるメモリ
アレイ14とを有してなることを特徴としている。
The memory device 1 according to the first embodiment includes:
It is connected to a single or a plurality of interconnects via an interconnect I / F circuit 11, can be mapped to a plurality of address spaces, and sets how to map a physical memory to each address space. It is characterized by having an address mapping setting unit 13 that can be used, a memory access control unit 12 that operates based on the settings made by the address mapping setting unit 13, and a memory array 14 that is a physical memory.

【0040】すなわち、当該第1の実施の形態のメモリ
装置1は、あるインターコネクトに接続されるものであ
るため、そのインターコネクトへのインターフェイス回
路としてインターコネクトI/F回路11を備え、メモ
リアレイ14へのメモリアクセスのためのメモリアクセ
ス制御部12を備えている。
That is, since the memory device 1 according to the first embodiment is connected to a certain interconnect, the memory device 1 includes an interconnect I / F circuit 11 as an interface circuit to the interconnect, and the memory array 14 A memory access control unit 12 for memory access is provided.

【0041】また、この第1の実施の形態のメモリ装置
1は、ある範囲のアドレス空間を持ち、インターコネク
ト側からインターフェイスI/F回路11及びメモリ制
御回路12を介して、そのアドレス空間に自由にアクセ
スすることが可能になされている。
The memory device 1 according to the first embodiment has a certain range of address space, and can freely enter the address space from the interconnect side via the interface I / F circuit 11 and the memory control circuit 12. Access has been made possible.

【0042】アドレスマッピング設定部13は、メモリ
アレイ14上の連続したメモリ空間を、任意のメモリア
クセスに対してマッピングするための機構である。ここ
での設定がメモリアクセス制御部12に通知され、メモ
リアクセス時に反映される。
The address mapping setting section 13 is a mechanism for mapping a continuous memory space on the memory array 14 to an arbitrary memory access. The setting here is notified to the memory access control unit 12 and is reflected at the time of memory access.

【0043】例えば、当該メモリ装置1におけるアドレ
ス空間をA空間とB空間の2つに分け、また例えば、メ
モリアレイ14を4×4のアレイに見立てた場合、A空
間からのアクセスに対しては、 アドレス0 アドレス1 アドレス2 アドレス3 アドレス4 アドレス5 アドレス6 アドレス7 アドレス8 アドレス9 アドレス10 アドレス11 アドレス12 アドレス13 アドレス14 アドレス15 のようにマッピングし、B空間からのアクセスに対して
は、 アドレス0 アドレス4 アドレス8 アドレス12 アドレス1 アドレス5 アドレス9 アドレス13 アドレス2 アドレス6 アドレス10 アドレス14 アドレス3 アドレス7 アドレス11 アドレス15 のようにマッピングすることが可能となる。
For example, when the address space in the memory device 1 is divided into two spaces, A space and B space. For example, when the memory array 14 is regarded as a 4 × 4 array, the access from the A space is Address 0 Address 1 Address 2 Address 3 Address 4 Address 5 Address 6 Address 7 Address 8 Address 9 Address 10 Address 11 Address 12 Address 13 Address 14 Address 15 0 Address 4 Address 8 Address 12 Address 1 Address 5 Address 9 Address 13 Address 2 Address 6 Address 10 Address 14 Address 3 Address 7 Address 11 Address 15

【0044】このように、本発明では、同じメモリ領域
に対する2つの連続アクセスを別々の結果にすることが
可能になっている。本発明では、この機能を利用して、
ディスクアレイコントローラにおける動作効率を向上さ
せるようにしている。
As described above, according to the present invention, two consecutive accesses to the same memory area can have different results. In the present invention, using this function,
The operation efficiency of the disk array controller is improved.

【0045】次に、図3には、本発明の第2の実施の形
態のメモリ装置1の概略構成を示す。なお、この図3の
各構成要素において、図2の各構成要素と同様の機能を
有するものにはそれぞれ同一の指示符号を付してその説
明は省略する。
Next, FIG. 3 shows a schematic configuration of a memory device 1 according to a second embodiment of the present invention. 3 having the same functions as those of FIG. 2 are designated by the same reference numerals and the description thereof will be omitted.

【0046】当該第2の実施の形態のメモリ装置1は、
メモリアレイ14をデータメモリ14Dとパリティメモ
リ14Pに分けて使用するようにしており、さらに、デ
ータとパリティの比を設定(N:1の整数Nを設定)す
るデータ・パリティ比設定部15と、そのデータ・パリ
ティ比の設定に基づいて、あるアドレス空間からの任意
の連続したN個の単位ライトアクセスに対してパリティ
を計算するパリティ演算部16とを備え、データライト
の後に上記パリティをパリティメモリ14Pに対してラ
イトアクセスすることを特徴としている。したがって、
メモリアクセス制御部12は、パリティメモリ14Pの
アクセス制御も行う。
The memory device 1 according to the second embodiment comprises:
The memory array 14 is divided into a data memory 14D and a parity memory 14P, and further used is a data / parity ratio setting unit 15 for setting a data / parity ratio (setting an integer N of N: 1). A parity operation unit 16 for calculating a parity for an arbitrary continuous N unit write accesses from a certain address space based on the setting of the data / parity ratio. It is characterized in that write access is made to 14P. Therefore,
The memory access control unit 12 also controls access to the parity memory 14P.

【0047】なお、図3の例は、データ・パリティ比を
例えば4:1とした場合を示している。この場合、図1
のHDD8は、少なくともデータ用の4台のHDDとパ
リティ用の1台のHDDの5台となる。
The example of FIG. 3 shows a case where the data / parity ratio is, for example, 4: 1. In this case, FIG.
HDDs 8 are at least four HDDs for data and one HDD for parity.

【0048】この第2の実施の形態のメモリ装置1の具
体的な動作は以下のようである。
The specific operation of the memory device 1 according to the second embodiment is as follows.

【0049】ここで、PCIバスのようなインターコネ
クト上を伝送されるデータの単位はワード単位であるこ
とが普通である。したがって、ホストコンピュータから
前記ディスクアレイコントローラ6に送られて来るライ
トデータが、例えば図3に示すデータD0,データD
1,データD2,データD3のように、単位データ(ワ
ード単位のデータ)のバースト転送として送られてくる
とした場合、当該第2の実施の形態のメモリ装置1は以
下のような手順で動作する。
Here, a unit of data transmitted on an interconnect such as a PCI bus is usually a word unit. Therefore, the write data sent from the host computer to the disk array controller 6 is, for example, data D0 and data D0 shown in FIG.
Assuming that the data is sent as a burst transfer of unit data (data in units of words), such as 1, data D2 and data D3, the memory device 1 according to the second embodiment operates in the following procedure. I do.

【0050】すなわち、メモリアレイ14のデータメモ
リ14Dを4×4のアレイに見立て、ホストコンピュー
タから当該メモリ装置1へのアクセスが、上述したA空
間からのアクセス順序に準ずるとした場合、先ず、第1
のステップとして、初期状態でのデータメモリ14D上
には、 データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し のようにデータが無く、また、パリティ演算部16はク
リア(0)されている。
That is, if the data memory 14D of the memory array 14 is regarded as a 4 × 4 array and the access from the host computer to the memory device 1 follows the access order from the space A, first, 1
In the initial state, there is no data on the data memory 14D No data No data No data No data No data No data No data No data No data No data No data No data No data No data No data There is no data, and the parity operation unit 16 is cleared (0).

【0051】次に、第2のステップとして、ホストコン
ピュータからディスクアレイコントローラ6にデータD
0のライトデータが送られてくると、データメモリ14
D上には、 データD0 データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し のようにアドレス0の領域にデータD0が格納され、ま
た、パリティ演算部16では、第1のステップにおける
初期値0とデータD0との排他的論理和演算が行われ
る。
Next, as a second step, the data D is transmitted from the host computer to the disk array controller 6.
When write data of 0 is sent, the data memory 14
Data D0 is stored in the area of address 0, such as data D0 No data No data No data No data No data No data No data No data No data No data No data No data No data No data No data Further, the parity operation unit 16 performs an exclusive OR operation of the initial value 0 and the data D0 in the first step.

【0052】次に、第3のステップとして、ホストコン
ピュータからディスクアレイコントローラ6にデータD
1のライトデータが送られてくると、データメモリ14
D上には、 データD0 データD1 データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し のようにアドレス1の領域にデータD1が格納され、ま
た、パリティ演算部16では、データD0とデータD1
との排他的論理和演算が行われる。
Next, as a third step, the data D is transmitted from the host computer to the disk array controller 6.
When the 1 write data is sent, the data memory 14
On D, data D1 is stored in the area of address 1, such as Data D0 Data D1 No data No data No data No data No data No data No data No data No data No data No data No data No data In addition, the parity operation unit 16 outputs data D0 and data D1.
And the exclusive OR operation is performed.

【0053】次に、第4のステップとして、ホストコン
ピュータからディスクアレイコントローラ6にデータD
2のライトデータが送られてくると、データメモリ14
D上には、 データD0 データD1 データD2 データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し のようにアドレス2の領域にデータD2が格納され、ま
た、パリティ演算部16では、データD0とデータD1
とデータD2の排他的論理和演算が行われる。
Next, as a fourth step, the data D is transmitted from the host computer to the disk array controller 6.
2 write data, the data memory 14
On data D, data D2 is stored in the area of address 2 like data D0 data D1 data D2 no data no data no data no data no data no data no data no data no data no data no data no data In addition, the parity operation unit 16 outputs data D0 and data D1.
And the data D2 are subjected to an exclusive OR operation.

【0054】次に、第5のステップとして、ホストコン
ピュータからディスクアレイコントローラ6にデータD
3のライトデータが送られてくると、データメモリ14
D上には、 データD0 データD1 データD2 データD3 データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し データ無し のようにアドレス3の領域にデータD3が格納され、ま
た、パリティ演算部16では、データD0とデータD1
とデータD2とデータD3の排他的論理和演算が行われ
る。これにより、パリティ演算は完了し、得られたパリ
ティデータがパリティメモリ14Pに書き込まれる。
Next, as a fifth step, the data D is transmitted from the host computer to the disk array controller 6.
3 is sent, the data memory 14
On D, data D3 is stored in the area of address 3, such as Data D0 Data D1 Data D2 Data D3 No data No data No data No data No data No data No data No data No data No data No data In addition, the parity operation unit 16 outputs data D0 and data D1.
The exclusive OR operation of data D2 and data D3 is performed. Thus, the parity calculation is completed, and the obtained parity data is written to the parity memory 14P.

【0055】その後、続いてデータD4,データD5,
データD6,・・・とデータが供給される場合には、第
1のステップの動作に戻り、以下上述と同様の動作を繰
り返す。ただし、この場合、メモリアレイ14(データ
メモリ14D)の1行目には先に供給されたデータD0
からデータD3が既に格納されているため、これらデー
タD4以降の各データは、データメモリ14Dの2行目
以降に記憶される。
Thereafter, subsequently, data D4, data D5,
When data D6,... Are supplied, the operation returns to the operation of the first step, and the same operation as described above is repeated. However, in this case, the data D0 supplied earlier is stored in the first row of the memory array 14 (data memory 14D).
Since the data D3 have already been stored, the data after the data D4 are stored in the second and subsequent rows of the data memory 14D.

【0056】上述したような方法をとれば、ライトオペ
レーション時に、いわゆるリードモディファイライト
(Read Modifi Write)動作を行わずに、パリティ演算
ができる。なお、リードモディファイライト動作とは、
同一メモリサイクル内に一つのメモリセルのデータを読
み出し、更に修正して書き換える動作である。
According to the above-described method, a parity operation can be performed without performing a so-called read modify write operation during a write operation. The read-modify-write operation is
This is an operation of reading data of one memory cell within the same memory cycle, and further revising and rewriting the data.

【0057】このようにして、メモリアレイ14へのラ
イト動作が完了した後、例えばB空間からのアクセスに
より、当該メモリアレイ14のデータメモリ14Dに格
納された各データD0〜D3(データD4以降のデータ
も格納しているときはそれらデータも)が読み出され、
HDDI/Fコントローラ7を介して4台のHDD8に
振り分けれて記録される。また、パリティメモリ14P
に格納されているパリティデータも同時に読み出され、
HDDI/Fコントローラ7を介してパリティ用のHD
D8に記録される。なお、図示は省略するが、上記5台
のHDD8を、説明の都合上、HDD80、HDD8
1、HDD82、HDD83、HDD8Pとして表した
場合、HDD80にはデータメモリ14Dの例えば左側
から1番目の縦列のデータが記録され、HDD81には
データメモリ14Dの例えば左側から2番目の縦列のデ
ータが、HDD83にはデータメモリ14Dの例えば左
側から3番目の縦列のデータが、HDD8Pにはパリテ
ィメモリ14Pのデータがそれぞれ記録される。以上に
より、第2の実施の形態の場合におけるディスクアレイ
装置へのデータの書き込み動作が完了する。
After the write operation to the memory array 14 is completed in this way, each of the data D0 to D3 (data D4 and subsequent data) stored in the data memory 14D of the memory array 14 is accessed by, for example, accessing from the B space. If you also store data, that data is also read)
The data is distributed to four HDDs 8 via the HDD I / F controller 7 and recorded. The parity memory 14P
Is also read out at the same time,
HD for parity via HDD I / F controller 7
Recorded in D8. Although not shown, the five HDDs 8 are referred to as HDDs 80 and 8 for convenience of explanation.
1, the HDD 82, the HDD 83, and the HDD 8P, the HDD 80 records, for example, the first column of data from the left side of the data memory 14D, and the HDD 81 stores the data of, for example, the second column from the left side of the data memory 14D. For example, data in the third column from the left of the data memory 14D is recorded in the HDD 83, and data in the parity memory 14P is recorded in the HDD 8P. As described above, the operation of writing data to the disk array device in the case of the second embodiment is completed.

【0058】次に、図4には、本発明の第3の実施の形
態のメモリ装置1の概略構成を示す。なお、この図4の
各構成要素において、図2及び図3の各構成要素と同様
の機能を有するものにはそれぞれ同一の指示符号を付し
てその説明は省略する。
Next, FIG. 4 shows a schematic configuration of a memory device 1 according to a third embodiment of the present invention. 4 having the same functions as those in FIGS. 2 and 3 are denoted by the same reference numerals, and description thereof is omitted.

【0059】第3の実施の形態のメモリ装置1は、メモ
リアレイ14内の任意の一部の領域をサブメモリアレイ
として使用可能にすると共に当該サブメモリアレイのマ
ッピングを設定可能とし、このサブメモリアレイのマッ
ピング設定によって、メモリアレイ14に対する任意の
単位アクセスを対応するサブメモリアレイへのアクセス
に変換可能にしたことを特徴としている。すなわち、こ
の第3の実施の形態では、メモリアレイのマッピング設
定により、サブメモリアレイをメモリアレイ14内の任
意の領域にマッピングすることを可能にしている。な
お、サブメモリアレイは、メモリアレイ14とは別個に
設けられるものであってもよい。
The memory device 1 according to the third embodiment makes it possible to use an arbitrary part of the memory array 14 as a sub memory array and to set the mapping of the sub memory array. It is characterized in that an arbitrary unit access to the memory array 14 can be converted into an access to a corresponding sub memory array by setting the mapping of the array. That is, in the third embodiment, the sub memory array can be mapped to an arbitrary area in the memory array 14 by setting the mapping of the memory array. Note that the sub memory array may be provided separately from the memory array 14.

【0060】図4の例では、サブメモリアレイとして前
記パリティメモリ14Pを使用した例を挙げており、し
たがって、サブメモリアレイのマッピング設定手段とし
てパリティメモリマッピング設定部17を備え、当該パ
リティメモリマッピング設定部(サブメモリアレイマッ
ピング設定部)17により、メモリアレイ14に対する
任意の単位アクセスを対応するパリティメモリ(サブメ
モリアレイ)14Pへのアクセスに変換可能としてい
る。メモリアクセス制御部12は、アドレスマッピング
設定部13とパリティメモリマッピング設定部(サブメ
モリアレイマッピング設定部)17による設定に基づい
て動作することになる。
FIG. 4 shows an example in which the parity memory 14P is used as a sub memory array. Therefore, a parity memory mapping setting unit 17 is provided as a sub memory array mapping setting means. The unit (sub memory array mapping setting unit) 17 can convert an arbitrary unit access to the memory array 14 into an access to the corresponding parity memory (sub memory array) 14P. The memory access control unit 12 operates based on the settings by the address mapping setting unit 13 and the parity memory mapping setting unit (sub memory array mapping setting unit) 17.

【0061】なお、上述のようにサブメモリアレイとし
てパリティメモリ14Pを使用するのではなく、データ
メモリ14D内の任意の領域をサブメモリアレイとする
ことも可能である。また、この図4の例は、データ・パ
リティ比を例えば4:1とした場合を示している。この
場合のHDD8は、データ用の4台のHDDとパリティ
データ用の1台のHDDの5台となる。
Incidentally, instead of using the parity memory 14P as the sub memory array as described above, an arbitrary area in the data memory 14D can be used as the sub memory array. The example of FIG. 4 shows a case where the data / parity ratio is, for example, 4: 1. The HDDs 8 in this case are four HDDs for data and one HDD for parity data.

【0062】この第3の実施の形態のメモリ装置1をデ
ィスクアレイ装置に使用した場合の具体的な動作は以下
のようである。なお、図示は省略するが、上記5台のH
DD8を、HDD80、HDD81、HDD82、HD
D83、HDD8Pとして表した場合、HDD80はデ
ータメモリ14Dの左側から1番目の縦列のデータを記
録し、HDD81はデータメモリ14Dの左側から2番
目の縦列のデータを、HDD82はデータメモリ14D
の左側から3番目の縦列のデータを、HDD8Pはパリ
ティメモリ14Pのデータを、それぞれ記録するものと
する。
The specific operation when the memory device 1 according to the third embodiment is used in a disk array device is as follows. Although not shown, the above five H
DD8 is replaced by HDD80, HDD81, HDD82, HD
When represented as D83 and HDD8P, the HDD 80 records the data of the first column from the left side of the data memory 14D, the HDD 81 stores the data of the second column from the left side of the data memory 14D, and the HDD 82 stores the data of the data column 14D.
, The HDD 8P records data in the parity memory 14P.

【0063】これら5台のHDD8(80〜8P)から
なるディスクアレイに記録されているデータを読み出す
(リードする)場合を考える。
Consider a case where data recorded in a disk array composed of these five HDDs 8 (80 to 8P) is read (read).

【0064】HDD8(80〜8P)から読み出された
データは、B空間のアクセス順序によってメモリアレイ
14に書き込まれる。したがって、例えばHDD80か
ら読み出されたデータはデータメモリ14Dの左側から
1番目の縦列に書き込まれ、HDD81から読み出され
たデータはデータメモリ14Dの左側から2番目の縦列
に、HDD82ら読み出されたデータはデータメモリ1
4Dの左側から3目の縦列にそれぞれ書き込まれる。
The data read from the HDDs 8 (80 to 8P) are written to the memory array 14 in the B space access order. Therefore, for example, data read from the HDD 80 is written to the first column from the left of the data memory 14D, and data read from the HDD 81 is read from the HDD 82 to the second column from the left of the data memory 14D. Data stored in data memory 1
The data is written in the third column from the left of 4D.

【0065】ここで、例えばHDD81が故障した場合
を考える。このように、HDD81が故障した場合、デ
ータメモリ14Dの左側から2番目の縦列には正しいデ
ータが書き込まれていないことになる。この状態で、デ
ィスクアレイコントローラ6が、A空間のアクセス順序
によって当該データメモリ14Dのデータを読み出して
ホストコンピュータに渡すと、当該ホストコンピュータ
には間違ったデータが伝送されてしまうことになる。
Here, consider the case where the HDD 81 breaks down, for example. Thus, when the HDD 81 fails, correct data is not written in the second column from the left of the data memory 14D. In this state, if the disk array controller 6 reads the data in the data memory 14D according to the access order of the space A and passes it to the host computer, incorrect data will be transmitted to the host computer.

【0066】なお、一般のディスクアレイコントローラ
の中には、いわゆるオンザフライ(On the Fly)でデー
タを修正するために、データ用の複数のHDDのデータ
をデータ用のバッファメモリに書き込む際に、パリティ
用のバッファメモリにリードモディファイライト動作を
行うことで、HDDが1台故障していたとしても、その
故障HDDに記録されていたデータを、パリティ用のバ
ッファメモリに再生するようになされたものが存在して
いる。しかしながら、このようなシステムでは、データ
用のバッファメモリとパリティ用のバッファメモリは別
物として扱われるために、その故障HDDのデータを一
度別の領域にコピーしないと完全なデータを得ることが
できない。
It is to be noted that some general disk array controllers include a parity when writing data from a plurality of data HDDs to a data buffer memory in order to correct data on the fly. By performing a read-modify-write operation on the buffer memory for parity, even if one HDD fails, the data recorded in the failed HDD is reproduced in the buffer memory for parity. Existing. However, in such a system, the buffer memory for data and the buffer memory for parity are treated as different things, so that complete data cannot be obtained unless the data of the failed HDD is copied to another area once.

【0067】これに対して、本発明の第3の実施の形態
のメモリ装置1では、パリティメモリマッピング設定部
(サブメモリアレイマッピング設定部)17の設定によ
り、パリティメモリ(サブメモリアレイ)14Pを、メ
モリアレイ14のHDD81(故障HDD)に対応する
部分にマッピングすることができるので、一度設定を行
うだけで、HDDが故障していない場合と同様に、当該
メモリ(サブメモリアレイ)をアクセスすれば、正しい
データを読み出すことが可能になっている。すなわち、
メモリアクセス制御部12を介した場合に、パリティメ
モリ14Pの内容がメモリアレイ14の不正データ部分
にマッピングされて見えるので、図1のディスクアレイ
コントローラ6はHDDが故障していない場合と同じア
クセスをすることができる。
On the other hand, in the memory device 1 according to the third embodiment of the present invention, the parity memory (sub memory array) 14P is set by the setting of the parity memory mapping setting section (sub memory array mapping setting section) 17. Can be mapped to a portion of the memory array 14 corresponding to the HDD 81 (failed HDD), so that the setting is performed only once and the memory (sub-memory array) can be accessed in the same manner as in the case where the HDD has not failed. For example, correct data can be read. That is,
Since the contents of the parity memory 14P appear to be mapped to the incorrect data portion of the memory array 14 via the memory access control unit 12, the disk array controller 6 of FIG. 1 performs the same access as when the HDD has not failed. can do.

【0068】次に、図5には、本発明の上述した第1〜
第3の何れかの実施の形態のメモリ装置1を、実際の回
路基板上に実装し、PCIに接続する場合の大まかなブ
ロック構成を示す。
Next, FIG. 5 shows the above-described first to first embodiments of the present invention.
A rough block configuration in the case where the memory device 1 according to any one of the third embodiments is mounted on an actual circuit board and connected to PCI is shown.

【0069】この図5において、図中のSRAM(スタ
ティックRAM)46は前記メモリアレイ14のデータ
メモリ14Dに相当する。SRAM44,45は前記パ
リティメモリ14Pに相当し、これらSRAM44,4
5はバンク動作する。SRAM44,45にそれぞれ接
続されている2つのPLD(Programmable Logic Devic
e)42,43は、パリティメモリのコントローラと、
リードモディファイライト動作を行うためのパリティ演
算器と、前記B空間用のアクセスを行うRAIDアシス
ト回路を内蔵する。PCIバスに接続されるPLD41
は、前記アドレスマッピング設定部13、リードモディ
ファイライト動作を伴わないパリティ演算器と、PCI
バスに接続されるPCII/F回路と、前記A空間用の
アクセスを行うRAIDアシスト回路を内蔵する。PC
II/F回路は、例えば64ビット、33MHzで動作
する。上述したように、本発明実施の形態のメモリ装置
1においては、データメモリ14Dとパリティメモリ1
4Pに高速動作が可能なSRAMを使用し、アドレスの
行(ロウ)方向、列(カラム)方向に無関係に無駄なレ
イテンシ無しのアクセスが出来るようになっている。
In FIG. 5, an SRAM (static RAM) 46 in the figure corresponds to the data memory 14D of the memory array 14. The SRAMs 44 and 45 correspond to the parity memory 14P.
5 performs a bank operation. Two PLDs (Programmable Logic Devices) connected to the SRAMs 44 and 45, respectively.
e) 42 and 43 are controllers of the parity memory,
A parity arithmetic unit for performing a read-modify-write operation and a RAID assist circuit for accessing the B space are built in. PLD 41 connected to PCI bus
Is a parity calculator that does not involve a read-modify-write operation.
It incorporates a PCII / F circuit connected to the bus and a RAID assist circuit for accessing the A space. PC
The II / F circuit operates at, for example, 64 bits and 33 MHz. As described above, in the memory device 1 according to the embodiment of the present invention, the data memory 14D and the parity memory 1
An SRAM capable of high-speed operation is used for the 4P, and access can be performed without useless latency regardless of the row (row) direction and the column (column) direction of the address.

【0070】なお、本実施の形態において使用するメモ
リは、上記SRAMに限らず、DRAMやSDRAM、
RDRAMなどを使用することも可能である。
The memory used in the present embodiment is not limited to the above SRAM, but may be a DRAM, an SDRAM,
It is also possible to use RDRAM or the like.

【0071】次に、図6には、本実施の形態のメモリ装
置のアドレスマップの一例を示す。
Next, FIG. 6 shows an example of an address map of the memory device according to the present embodiment.

【0072】この図6において、A空間及びB空間のア
クセスは、前記図2で説明したのと同じである。実装す
るPLDでは、メモリアレイ14のサイズ、データ・パ
リティ比は設定可能なパラメータとなっている。
In FIG. 6, access to the A space and the B space is the same as that described in FIG. In the PLD to be mounted, the size of the memory array 14 and the data / parity ratio are settable parameters.

【0073】図7〜図10には、物理メモリ(メモリア
レイ14)のマッピングの一例を示し、ハードウェアの
実装をどのようにすればよいかを示している。なお、こ
の方法は、ハードウェアの実装は簡単であるが、データ
・パリティ比は2のべき乗に制限される。
FIGS. 7 to 10 show an example of mapping of the physical memory (memory array 14), and show how to implement hardware. In this method, the hardware implementation is simple, but the data parity ratio is limited to a power of two.

【0074】メモリアレイ14のマッピングにおいて
は、図7に示す行(ロウ)アドレスと列(カラム)アド
レスを入れ替えるようにマッピングを行い、RAIDの
実装をアシストする。
In the mapping of the memory array 14, mapping is performed so that the row (row) address and the column (column) address shown in FIG. 7 are exchanged, and the mounting of RAID is assisted.

【0075】A空間からのアクセスに対しては、図8中
の矢印に示すような流れでのアドレス割付となり、ま
た、B空間からのアクセスに対しては図9中の矢印に示
すような流れでのアドレス割付となる。
For access from the A space, address allocation is performed in the flow shown by the arrow in FIG. 8, and for access from the B space, the address allocation is performed as shown by the arrow in FIG. Address assignment.

【0076】なお、行(ロウ)アドレスと列(カラム)
アドレスのビット長を適当に固定すれば、図10に示す
ようなイメージとなる。
The row (row) address and column (column)
If the bit length of the address is appropriately fixed, an image as shown in FIG. 10 is obtained.

【0077】ここで、パリティ計算付きのA空間に対す
るライトアクセスは、図5のPLD41に含まれるA空
間用のRAIDアシスト回路によるパリティ生成を伴
い、データをデータメモリに書き込んだ後にパリティメ
モリに対するライトアクセスが発生する。パリティ計算
付きのB空間に対するライトアクセスは、図5のPLD
42,43に含まれるB空間用のRAIDアシスト回路
によるパリティ生成を伴う。
Here, the write access to the A space with the parity calculation involves parity generation by the RAID assist circuit for the A space included in the PLD 41 in FIG. 5, and the write access to the parity memory after writing the data to the data memory. Occurs. Write access to the B space with parity calculation is performed by the PLD shown in FIG.
This involves parity generation by the RAID assist circuit for the B space included in 42 and 43.

【0078】また、初期設定時におけるA空間とパリテ
ィ計算付きのA空間に対するリードアクセスの結果は等
価であり、B空間とパリティ計算付きのB空間に対する
リードアクセスに対しても同様である。
The results of the read access to the A space and the A space with parity calculation at the time of the initial setting are equivalent, and the same applies to the B space and the read access to the B space with parity calculation.

【0079】さらに、サブメモリアレイマッピング設定
部に対応するエラーアドレスレジスタへ適切な設定を行
うことによって、パリティ計算付きのA空間及びパリテ
ィ計算付きのB空間に対して指定したアドレスへのリー
ドアクセスは、対応するパリティメモリへのリードに変
換される。
Further, by appropriately setting the error address register corresponding to the sub memory array mapping setting unit, the read access to the address specified for the A space with parity calculation and the B space with parity calculation can be performed. Is converted to a read to the corresponding parity memory.

【0080】次に、本発明実施の形態のメモリ装置1に
おけるライト時の制御の流れとリード時の制御の流れに
ついて説明する。図11にはライト時の制御の流れを示
し、図12にはリード時の制御の流れを示す。
Next, the flow of control during writing and the flow of control during reading in the memory device 1 according to the embodiment of the present invention will be described. FIG. 11 shows a control flow at the time of writing, and FIG. 12 shows a control flow at the time of reading.

【0081】ライト動作及びリード動作を行う前に、デ
ィスクアレイコントローラ6は、メモリ装置1を使用す
るため、各アドレス空間のベースアドレスとサイズ及び
アドレス変換に関係のあるパラメータである単位サイ
ズ、カラムアドレスのビット長、ロウアドレスのビット
長を設定する。
Before performing the write operation and the read operation, the disk array controller 6 uses the memory device 1, so that the base address and size of each address space and the unit size and column address which are parameters related to the address conversion are used. And the bit length of the row address.

【0082】ここでは、説明の都合上、A空間ベースア
ドレスを0x0000 0000とし、パリティ計算付
きのA空間ベースアドレスを0x1000 0000、
B空間ベースアドレスを0x2000 0000、パリ
ティ計算付きのB空間ベースアドレスを0x3000
0000、パリティメモリベースアドレスを0x400
0 0000、メモリサイズを1MB、パリティメモリ
サイズを256KB、単位サイズを8バイト、カラムア
ドレス長を2ビット、ロウアドレス長を13ビットとす
る。
Here, for convenience of explanation, the A space base address is set to 0x0000 0000, the A space base address with parity calculation is set to 0x1000 0000,
0x2000 0000 for B space base address, 0x3000 for B space base address with parity calculation
0000, parity memory base address is 0x400
0 0000, the memory size is 1 MB, the parity memory size is 256 KB, the unit size is 8 bytes, the column address length is 2 bits, and the row address length is 13 bits.

【0083】図11に示すライト時の制御において、先
ず、ステップS1として、ディスクアレイコントローラ
6は、メモリアレイ14に対して、パリティ計算付きの
A空間でアドレス0x0000 0000から256K
Bのサイズのデータメモリ14Dを確保し、データをラ
イトする。同時に、カラムアドレス長が2ビットなの
で、4単位サイズ毎にパリティデータが計算され、パリ
ティメモリ14Pに書き込まれる。
In the control at the time of writing shown in FIG. 11, first, in step S1, the disk array controller 6 sends the memory array 14 from the address 0x0000 0000 to 256K in the A space with parity calculation.
A data memory 14D of size B is secured and data is written. At the same time, since the column address length is 2 bits, parity data is calculated for every four unit sizes and written to the parity memory 14P.

【0084】その後、ステップS2として、HDDI/
Fコントローラ7は、B空間のアドレス0x2000
0000から64KBのブロック単位、すなわちストラ
イプユニット単位でデータを読み出し、それらストライ
プユニット単位のデータを別々のHDDに書き込み、同
様にパリティデータもアドレス0x4000 0000
から64KBのストライプユニット単位で読み出してパ
リティデータ用のHDDに書き込む。
Thereafter, as step S2, the HDDI /
The F controller 7 has an address 0x2000 in the B space.
Data is read out in block units of 0000 to 64 KB, that is, in stripe unit units, and the data in stripe unit units is written to separate HDDs. Similarly, parity data has an address of 0x4000 0000.
The data is read in units of 64 KB stripe units and written to the HDD for parity data.

【0085】このように、ディスクアレイコントローラ
6は、ストライプユニット単位でメモリアレイ14を管
理し、それぞれのメモリ領域にデータを分割してライト
する必要がなく、また、HDDI/Fコントローラ7も
A空間とB空間の対応が既知であるので、パリティ計算
付きA空間のメモリ領域の先頭アドレスに対応するB空
間のアドレスからストライプユニット単位で読み出すこ
とで4台のデータ用のHDDと1台のパリティ用HDD
に矛盾なくデータを書く事が可能となっている。
As described above, the disk array controller 6 manages the memory array 14 in units of stripe units, and does not need to divide and write data in each memory area, and the HDD I / F controller 7 also has the A space. And the B space are known, so that four data HDDs and one parity HDD are read out from the address in the B space corresponding to the head address of the memory space in the A space with parity calculation in stripe unit units. HDD
It is possible to write data without contradiction.

【0086】リード動作は、前記ライト動作の全く逆を
行なえばよい。
The read operation may be exactly the reverse of the write operation.

【0087】すなわち、図12に示すリード時の制御に
おいて、先ず、ステップS10として、HDDI/Fコ
ントローラ7は、リード対象のストライプユニット単位
のデータとパリティデータを、全てパリティ計算付きB
空間のアドレス0x30000000からメモリアレイ
14に書き込む。この時、パリティメモリ14Pには全
てのHDDから読み出した64KBのストライプユニッ
ト単位のデータブロックの排他的論理和を取ったものが
格納されるが、その後の操作には関係しない。
That is, in the control at the time of reading shown in FIG. 12, first, in step S10, the HDD I / F controller 7 converts all data and parity data in units of stripe units to be read into parity data with parity calculation.
Write to the memory array 14 from the space address 0x30000000. At this time, the parity memory 14P stores an exclusive OR of the data blocks read from all HDDs in units of 64 KB stripe units, but does not relate to subsequent operations.

【0088】その後、後述するステップS11でノーと
判定されたときのステップS14において、ディスクア
レイコントローラ6は、パリティ計算付きA空間のアド
レス0x1000 0000から256KBのデータを
リードする。このように、ディスクアレイコントローラ
6は、分散されたメモリアレイのメモリ領域のギャザリ
ングを行なうことなく、単一のデータメモリ14Dを読
み出すだけで良い事になる。
After that, in step S14 when it is determined NO in step S11 described later, the disk array controller 6 reads 256 KB of data from the address 0x1000 0000 in the A space with parity calculation. As described above, the disk array controller 6 only needs to read the single data memory 14D without gathering the memory areas of the distributed memory arrays.

【0089】一方、ステップS11では、HDDが壊れ
ているか否かの判定が行われ、このステップS11にて
HDDが故障したときはステップS12の処理に進む。
On the other hand, in step S11, it is determined whether or not the HDD is broken. If the HDD has failed in step S11, the process proceeds to step S12.

【0090】ここで、例えば1台のHDDが故障した状
態で全てのHDDからのリードが終了した段階では、A
空間のアクセスでデータメモリ14Dをリードした場
合、故障HDDから読み出すべきデータのエリアは不正
となっているはずである。
Here, for example, at the stage where reading from all HDDs is completed in the state where one HDD has failed, A
If the data memory 14D is read by accessing the space, the area of the data to be read from the failed HDD must be incorrect.

【0091】従って、ディスクアレイコントローラ6
は、ステップS12において、エラーアドレスレジスタ
に故障HDDに対応する値が既に設定されているか否か
を判定し、設定されているときにはステップS14の処
理に進み、一方、設定されていないときにはステップS
13の処理に進んで、エラーアドレスレジスタに故障H
DDに対応する値をセットした後、ステップS14の処
理に進んでパリティ計算付きA空間からリードするよう
にしている。これにより、通常時のリードと全く同様の
操作で正しいデータを得ることが出来る。更に、エラー
アドレスレジスタへの設定は、最初にHDDの故障を検
出した場合のみ行なえば良く、その後のアクセスは全く
故障の無い場合と同じである。
Therefore, the disk array controller 6
Determines in step S12 whether or not a value corresponding to the failed HDD has already been set in the error address register. If it has been set, the process proceeds to step S14.
Proceeding to the process of No. 13, the error address register stores the fault H
After setting the value corresponding to DD, the process proceeds to step S14 to read from the A space with parity calculation. Thus, correct data can be obtained by exactly the same operation as in a normal read. Further, the setting to the error address register may be performed only when the failure of the HDD is detected first, and the subsequent access is the same as in the case where there is no failure.

【0092】ライト動作時に故障HDDが存在する場合
は、正常時と全く同じ動作の流れとなる。最終的にHD
DI/Fコントローラ7が故障したHDDに書き込み操
作を行なわないだけである。
If a failed HDD exists during the write operation, the operation flow is exactly the same as in the normal operation. Eventually HD
The only difference is that the DI / F controller 7 does not perform a write operation on the failed HDD.

【0093】以上の説明では、簡単のためにHDDの数
を2の冪乗に制約されるような実装方法で説明している
が、データパリティ比が2の冪乗に制限されないように
するための、より一般的なアドレス変換について以下に
説明する。
In the above description, the mounting method is described in which the number of HDDs is restricted to the power of 2 for simplicity. However, in order to prevent the data parity ratio from being restricted to the power of 2, A more general address translation will be described below.

【0094】ここで、具体的な例として、図13に示す
ようにメモリアレイ14のブロックサイズBSを48と
し、データパリティ比3:1(N=3)とし、また、ア
ドレス0から50までアクセスする場合を挙げて、実際
にアドレス計算を行ってみることにする。
Here, as a specific example, as shown in FIG. 13, the block size BS of the memory array 14 is set to 48, the data parity ratio is set to 3: 1 (N = 3), and addresses 0 to 50 are accessed. In this case, address calculation is actually performed.

【0095】すなわちこの場合のアドレス計算は、 Base=(Addr/BS)×BS+[Addr/(BS/N)]
%N Offset=Addr%(BS/N) XferAddr=Base+N×Offset のように表すことができる。なお、ブロックサイズBS
は、Nの整数倍である。これらのパラメータは全て整数
であり、行われる演算は全て整数演算である。
That is, the address calculation in this case is as follows: Base = (Addr / BS) × BS + [Addr / (BS / N)]
% N Offset = Addr% (BS / N) XferAddr = Base + N × Offset The block size BS
Is an integral multiple of N. These parameters are all integers, and the operations performed are all integer operations.

【0096】上記アドレス計算を実現するためのアドレ
ス変換機構としては、図14に示すような構成が考えら
れる。
As an address conversion mechanism for realizing the above address calculation, a configuration as shown in FIG. 14 can be considered.

【0097】この図14において、入力されたアドレス
Addrは演算器51に送られる。また、N値設定レジスタ
52にはデータパリティ比を表すNの値が保持されてお
り、ブロックサイズ設定レジスタ53にはメモリアレイ
14のブロックサイズBSが保持されている。N値設定
レジスタ52からのN値は演算器51とN×オフセット
演算器56に送られ、ブロックサイズ設定レジスタ53
からのブロックサイズBSは演算器51に送られる。
In FIG. 14, the input address
Addr is sent to the arithmetic unit 51. Further, the N value setting register 52 holds the value of N representing the data parity ratio, and the block size setting register 53 holds the block size BS of the memory array 14. The N value from the N value setting register 52 is sent to the arithmetic unit 51 and the N × offset arithmetic unit 56, and the block size setting register 53
Is sent to the arithmetic unit 51.

【0098】演算器51では、アドレスAddrとブロック
サイズBSとN値とを用いて、ベースアドレスBaseとオ
フセットOffsetを演算する。ベースアドレスBaseは加算
器57へ送られ、オフセットOffsetはN×オフセット演
算器56に送られる。
The computing unit 51 computes a base address Base and an offset Offset using the address Addr, the block size BS, and the N value. The base address Base is sent to the adder 57, and the offset Offset is sent to the N × offset calculator 56.

【0099】N×オフセット演算器56では、オフセッ
トOffsetにN値を乗算し、その乗算結果のN×Offsetの
値を加算器57に送る。
The N × offset calculator 56 multiplies the offset Offset by the N value and sends the multiplied result of the N × Offset value to the adder 57.

【0100】加算器57では、ベースアドレスBaseとN
×Offsetの値を加算する。これにより、変換後のアドレ
スXferAddrが得られる。
At the adder 57, the base addresses Base and N
Add the value of × Offset. Thereby, the converted address XferAddr is obtained.

【0101】図15には、上記図14の構成におけるア
ドレス計算によって、入力アドレスAddrから生成される
ベースアドレスBaseとオフセットOffset、及び、変換後
のアドレスXferAddrの対応表を示す。
FIG. 15 shows a correspondence table between the base address Base generated from the input address Addr, the offset Offset, and the converted address XferAddr by the address calculation in the configuration shown in FIG.

【0102】次に、図16には、上述した本発明実施の
形態のメモリ装置1が適用されるシステムの第2の構成
例を示す。なお、この図16の各構成要素において、図
1の各構成要素と同様の機能を有するものにはそれぞれ
同一の指示符号を付してその説明は省略する。
Next, FIG. 16 shows a second configuration example of a system to which the above-described memory device 1 of the present invention is applied. 16 having the same functions as those of FIG. 1 are denoted by the same reference numerals, and description thereof is omitted.

【0103】この図16において、ディスクアレイコン
トローラ3は、PCIバスを介してホストコンピュータ
と接続される。当該ホストコンピュータからメモリ装置
1へのアクセスは、A空間からのアクセスである。
In FIG. 16, the disk array controller 3 is connected to a host computer via a PCI bus. Access from the host computer to the memory device 1 is access from the A space.

【0104】HDDI/Fコントローラ7a〜7cは、
それぞれ図1のHDDI/Fコントローラ7と同じ機能
を有し、それぞれに対して複数のHDD8a、8b、8
cが接続されている。これらHDDI/Fコントローラ
7a〜7cからメモリ装置1へのアクセスはB空間から
のアクセスである。
The HDD I / F controllers 7a to 7c are
Each has the same function as the HDD I / F controller 7 of FIG. 1, and a plurality of HDDs 8a, 8b, 8
c is connected. The access from the HDD I / F controllers 7a to 7c to the memory device 1 is an access from the B space.

【0105】次に、図17には、本発明実施の形態のメ
モリ装置1が適用されるシステムの第3の構成例を示
す。この図17は、本実施の形態のメモリ装置1が複数
のインターコネクトに接続される構成例を示している。
また、図17の構成例では、メモリ装置1として前記図
4と略々同様の構成を例に挙げている。なお、この図1
7の各構成要素において、図1及び図4の各構成要素と
同様の機能を有するものにはそれぞれ同一の指示符号を
付してその説明は省略する。
Next, FIG. 17 shows a third configuration example of a system to which the memory device 1 according to the embodiment of the present invention is applied. FIG. 17 shows a configuration example in which the memory device 1 of the present embodiment is connected to a plurality of interconnects.
In the configuration example of FIG. 17, the memory device 1 has a configuration substantially similar to that of FIG. Note that FIG.
7, components having the same functions as those of the components in FIGS. 1 and 4 are denoted by the same reference numerals, and description thereof is omitted.

【0106】この図17の構成においては、メモリ装置
1を複数のインターコネクトに接続した場合に必要な構
成として、PCII/F11C及び11Hを備えてい
る。PCII/F11Cはディスクアレイコントローラ
3側のPCIバスと接続され、PCII/F11HはH
DDI/Fコントローラ7側のPCIバスと接続されて
いる。各PCII/F11C及び11Hは、それぞれデ
ータ・パリティ比設定ブロック15とパリティメモリマ
ッピング設定ブロック17、メモリアクセス制御部1
2、アドレスマッピング設定部13と接続される。
In the configuration of FIG. 17, PCI / Fs 11C and 11H are provided as necessary components when the memory device 1 is connected to a plurality of interconnects. The PCIII / F11C is connected to the PCI bus on the disk array controller 3 side, and the PCIII / F11H is
It is connected to the PCI bus on the DDI / F controller 7 side. Each of the PCII / Fs 11C and 11H includes a data / parity ratio setting block 15, a parity memory mapping setting block 17, and a memory access control unit 1.
2. Connected to the address mapping setting unit 13.

【0107】図1と図16の構成例のように、1つのイ
ンターコネクトにメモリ装置1が接続されている場合に
は、例えばインターコネクトのバンド幅がメモリ装置1
のバンド幅より小さいと、メモリ装置1の性能を極限ま
で引き出すことができないのに対して、図17の構成の
ように、インターコネクトを複数に分けるようにすれ
ば、メモリ装置1を有効に使用することができる。例え
ば、RAMバスDRAMのような高いバンド幅を持つR
AMをメモリアレイ14に採用するような場合に効果が
ある。
When the memory device 1 is connected to one interconnect as in the configuration examples of FIGS. 1 and 16, for example, the bandwidth of the interconnect is
If the bandwidth is smaller than the bandwidth of the memory device 1, the performance of the memory device 1 cannot be maximized. On the other hand, if the interconnects are divided into a plurality as shown in FIG. 17, the memory device 1 is used effectively. be able to. For example, a high bandwidth R such as a RAM bus DRAM
This is effective when AM is used for the memory array 14.

【0108】上述したように、本実施の形態のメモリ装
置によれば、同一物理メモリアレイ14を複数のアドレ
ス空間にマッピングし、ディスクアレイコントローラ6
とHDDI/Fコントローラ7が別々のアドレス空間を
使用することにより、メモリアレイ14のリスト管理等
の手間を減らすことができ、さらに、HDDの故障など
でデータが正しく読み出せないような場合に、パリティ
メモリ14Pを適切なエリアにマッピングして、通常時
と等価なアクセスをすることが可能になっている。
As described above, according to the memory device of the present embodiment, the same physical memory array 14 is mapped to a plurality of address spaces, and the disk array controller 6
When the HDD I / F controller 7 and the HDD I / F controller 7 use different address spaces, it is possible to reduce the trouble of managing the list of the memory array 14 and the like. Further, when data cannot be read correctly due to a failure of the HDD or the like, The parity memory 14P is mapped to an appropriate area so that an access equivalent to a normal time can be performed.

【0109】なお、本実施の形態では、HDDアレイを
使用した例を挙げたが、例えばテープ状記録媒体を備え
たテープドライブアレイを用いた場合にも適用可能であ
る。
In the present embodiment, an example in which an HDD array is used has been described. However, the present invention is also applicable to a case in which, for example, a tape drive array having a tape-shaped recording medium is used.

【0110】[0110]

【発明の効果】以上の説明で明らかなように、本発明の
メモリ装置においては、単一又は複数のインターコネク
トに接続され、物理メモリであるメモリアレイを複数の
個々のアドレス空間にマッピングし、そのマッピングの
設定に基づいてメモリアレイのアクセスを制御すること
により、例えばデータバッファメモリのリスト管理等の
手間を減らすことができる。
As apparent from the above description, in the memory device of the present invention, a memory array which is connected to a single or a plurality of interconnects and is a physical memory is mapped to a plurality of individual address spaces, and By controlling access to the memory array based on the setting of the mapping, it is possible to reduce the trouble of, for example, managing the list of the data buffer memory.

【0111】また、本発明のメモリ装置は、単一又は複
数のインターコネクトに接続され、物理メモリであるメ
モリアレイの一部からなるサブメモリアレイ若しくは別
個のサブメモリアレイを使用し、設定によりメモリアレ
イに対する任意の単位アクセスを対応するサブメモリア
レイへのアクセスに変換することにより、例えばデータ
バッファメモリのリスト管理等の手間を減らすことがで
きる。
Further, the memory device of the present invention uses a sub memory array which is connected to a single or a plurality of interconnects and is a part of a memory array which is a physical memory or a separate sub memory array, and the memory array is set by setting. Is converted into an access to the corresponding sub-memory array, thereby reducing the trouble of, for example, managing the list of the data buffer memory.

【0112】ここで、本発明のメモリ装置において、メ
モリアレイをデータメモリとパリティメモリに分けて使
用し、データとパリティの比に基づいて一のアドレス空
間からの任意の連続したN個の単位ライトアクセスに対
してパリティを計算し、データメモリへのデータのライ
ト後に、パリティをパリティメモリに対してライトアク
セスすることにより、更に例えばHDD等の記録媒体の
故障などでデータが正しく読み出せない場合であって
も、パリティメモリに対して通常時と等価なアクセスを
行うことが可能である。
Here, in the memory device of the present invention, the memory array is divided into a data memory and a parity memory and used, and N consecutive unit write operations from one address space are performed based on the ratio of data and parity. The parity is calculated for the access, and after the data is written to the data memory, the parity is written to the parity memory, so that the data cannot be read correctly due to a failure of a recording medium such as an HDD. Even if there is, it is possible to access the parity memory equivalent to the normal time.

【0113】また、本発明のデータ記録再生装置におい
ては、本発明のメモリ装置を備え、複数の記録媒体から
なる記録媒体アレイを制御する記録媒体アレイ制御手段
と、記録媒体アレイと記録媒体アレイ制御手段との間の
データ送受を制御する記録媒体アレイインターフェイス
手段とが別々のアドレス空間を使用することにより、例
えばデータバッファメモリのリスト管理等の手間を減ら
すことができ、更に例えばHDD等の記録媒体の故障な
どでデータが正しく読み出せない場合であっても、パリ
ティメモリに対して通常時と等価なアクセスを行うこと
が可能である。
The data recording / reproducing apparatus of the present invention includes a memory device of the present invention, a recording medium array control means for controlling a recording medium array comprising a plurality of recording media, a recording medium array and a recording medium array control. The use of a separate address space between the recording medium array interface means and the recording medium array interface means for controlling data transmission / reception between the means can reduce the trouble of, for example, managing a list in a data buffer memory. Therefore, even if data cannot be read correctly due to a failure in the parity memory or the like, it is possible to access the parity memory equivalent to a normal access.

【0114】さらに、本発明のコンピュータシステムに
おいては、本発明のメモリ装置と本発明のデータ記録再
生装置とを備え、メモリ装置及びデータ記録再生装置を
インターコネクトを介して接続してなることにより、例
えばデータバッファメモリのリスト管理等の手間を減ら
すことができ、更に例えばHDD等の記録媒体の故障な
どでデータが正しく読み出せない場合であっても、パリ
ティメモリに対して通常時と等価なアクセスを行うこと
が可能である。
Further, the computer system of the present invention includes the memory device of the present invention and the data recording / reproducing device of the present invention, and the memory device and the data recording / reproducing device are connected via an interconnect. It is possible to reduce the work of managing the list of the data buffer memory and the like, and even if the data cannot be read correctly due to, for example, a failure of the recording medium such as the HDD, access to the parity memory equivalent to the normal access is performed. It is possible to do.

【図面の簡単な説明】[Brief description of the drawings]

【図1】本発明実施の形態のメモリ装置が適用されるシ
ステムの第1の構成例を示すブロック回路図である。
FIG. 1 is a block circuit diagram showing a first configuration example of a system to which a memory device according to an embodiment of the present invention is applied;

【図2】本発明の第1の実施の形態のメモリ装置の概略
構成を示すブロック回路図である。
FIG. 2 is a block circuit diagram illustrating a schematic configuration of the memory device according to the first embodiment of the present invention;

【図3】本発明の第2の実施の形態のメモリ装置の概略
構成を示すブロック回路図である。
FIG. 3 is a block circuit diagram illustrating a schematic configuration of a memory device according to a second embodiment of the present invention;

【図4】本発明の第3の実施の形態のメモリ装置の概略
構成を示すブロック回路図である。
FIG. 4 is a block circuit diagram illustrating a schematic configuration of a memory device according to a third embodiment of the present invention.

【図5】本実施の形態のメモリ装置を実際の回路基板上
に実装し、PCIに接続する場合の大まかな構成を示す
ブロック回路図である。
FIG. 5 is a block circuit diagram showing a rough configuration in a case where the memory device according to the present embodiment is mounted on an actual circuit board and connected to PCI;

【図6】本実施の形態のメモリ装置のアドレスマップの
一例を示す図である。
FIG. 6 is a diagram illustrating an example of an address map of the memory device according to the present embodiment;

【図7】メモリアレイの行(ロウ)アドレスと列(カラ
ム)アドレスの説明に用いる図である。
FIG. 7 is a diagram used to explain a row (row) address and a column (column) address of a memory array.

【図8】A空間からのアクセスに対するメモリアレイの
マッピングの説明に用いる図である。
FIG. 8 is a diagram used to explain mapping of a memory array with respect to access from a space A;

【図9】B空間からのアクセスに対するメモリアレイの
マッピングの説明に用いる図である。
FIG. 9 is a diagram used to explain mapping of a memory array with respect to access from a B space.

【図10】行(ロウ)アドレスと列(カラム)アドレス
のビット長を適当に固定した場合のメモリアレイのマッ
ピングの説明に用いる図である。
FIG. 10 is a diagram used to explain mapping of a memory array when the bit lengths of a row (row) address and a column (column) address are appropriately fixed.

【図11】本実施の形態のメモリ装置におけるライト時
の制御の流れを示すフローチャートである。
FIG. 11 is a flowchart showing a control flow at the time of writing in the memory device of the present embodiment.

【図12】本実施の形態のメモリ装置におけるリード時
の制御の流れを示すフローチャートである。
FIG. 12 is a flowchart showing a flow of control at the time of reading in the memory device of the present embodiment.

【図13】メモリアレイのブロックサイズの一例とイン
ターコネクトからのデータの流れの説明に用いる図であ
る。
FIG. 13 is a diagram illustrating an example of a block size of a memory array and a flow of data from an interconnect.

【図14】より一般的なアドレス変換のためのアドレス
計算を実現するアドレス変換機構の構成例を示すブロッ
ク回路図である。
FIG. 14 is a block circuit diagram illustrating a configuration example of an address translation mechanism that implements address calculation for more general address translation.

【図15】図14の構成でのアドレス計算により入力ア
ドレスAddrから生成される、ベースアドレスBaseとオフ
セットOffset、及び変換後のアドレスXferAddrの対応表
を示す図である。
15 is a diagram showing a correspondence table between a base address Base, an offset Offset, and a converted address XferAddr generated from an input address Addr by the address calculation in the configuration of FIG. 14;

【図16】本発明実施の形態のメモリ装置が適用される
システムの第2の構成例を示すブロック回路図である。
FIG. 16 is a block circuit diagram showing a second configuration example of a system to which the memory device according to the embodiment of the present invention is applied;

【図17】本発明実施の形態のメモリ装置が適用される
システムの第3の構成例を示すブロック回路図である。
FIG. 17 is a block circuit diagram showing a third configuration example of a system to which the memory device according to the embodiment of the present invention is applied;

【図18】RAID−3の従来のディスクアレイ装置に
おける書き込み時の動作のフローチャートである。
FIG. 18 is a flowchart of an operation at the time of writing in a conventional RAID-3 disk array device.

【図19】RAID−3の従来のディスクアレイ装置に
おける読み出し時の動作のフローチャートである。
FIG. 19 is a flowchart of an operation at the time of reading in a conventional RAID-3 disk array device.

【符号の説明】[Explanation of symbols]

1 メモリ装置、 2 CPU、 3 ホスト/PCI
ブリッジ部3、 4メインメモリ4、 5 入出力装
置、 6 ディスクアレイコントローラ、 7HDDI
/Fコントローラ、 8 HDD、 11 インターコ
ネクトI/F回路、 12 メモリアクセス制御部、
13 アドレスマッピング設定部、14 メモリアレ
イ、 14D データメモリ、 14P パリティメモ
リ、15 データ・パリティ比設定部、 16 パリテ
ィ演算部、 17 パリティメモリマッピング設定部
1 memory device, 2 CPU, 3 host / PCI
Bridge unit 3, 4 main memory 4, 5 input / output device, 6 disk array controller, 7 HDDI
/ F controller, 8 HDD, 11 interconnect I / F circuit, 12 memory access control unit,
13 address mapping setting unit, 14 memory array, 14D data memory, 14P parity memory, 15 data / parity ratio setting unit, 16 parity operation unit, 17 parity memory mapping setting unit

─────────────────────────────────────────────────────
────────────────────────────────────────────────── ───

【手続補正書】[Procedure amendment]

【提出日】平成10年11月12日(1998.11.
12)
[Submission date] November 12, 1998 (1998.11.
12)

【手続補正1】[Procedure amendment 1]

【補正対象書類名】明細書[Document name to be amended] Statement

【補正対象項目名】0025[Correction target item name] 0025

【補正方法】変更[Correction method] Change

【補正内容】[Correction contents]

【0025】そこで、本発明はこのような状況に鑑みて
なされたものであり、データバッファメモリのリスト管
理等の手間を減らすことができ、更にHDDの故障など
でデータが正しく読み出せない場合であっても、データ
バッファメモリに対して通常時と等価なアクセスを行う
ことが可能なメモリ装置と、そのメモリ装置を使用した
データ記録再生装置及びコンピュータシステムを提供す
ることを目的とする。
Accordingly, the present invention has been made in view of such a situation, and can reduce the trouble of managing the list of the data buffer memory and the like. It is an object of the present invention to provide a memory device capable of performing an equivalent access to a data buffer memory as usual, and a data recording / reproducing device and a computer system using the memory device.

【手続補正2】[Procedure amendment 2]

【補正対象書類名】明細書[Document name to be amended] Statement

【補正対象項目名】0067[Correction target item name] 0067

【補正方法】変更[Correction method] Change

【補正内容】[Correction contents]

【0067】これに対して、本発明の第3の実施の形態
のメモリ装置1では、パリティメモリマッピング設定部
(サブメモリアレイマッピング設定部)17の設定によ
り、パリティメモリ(サブメモリアレイ)14Pを、メ
モリアレイ14のHDD81(故障HDD)に対応する
部分にマッピングすることができるので、一度設定を行
うだけで、HDDが故障していない場合と同様に、当該
メモリ(メモリアレイ)をアクセスすれば、正しいデー
タを読み出すことが可能になっている。すなわち、メモ
リアクセス制御部12を介した場合に、パリティメモリ
14Pの内容がメモリアレイ14の不正データ部分にマ
ッピングされて見えるので、図1のディスクアレイコン
トローラ6はHDDが故障していない場合と同じアクセ
スをすることができる。
On the other hand, in the memory device 1 according to the third embodiment of the present invention, the parity memory (sub memory array) 14P is set by the setting of the parity memory mapping setting section (sub memory array mapping setting section) 17. Can be mapped to a portion of the memory array 14 corresponding to the HDD 81 (failed HDD), so that the setting is performed only once, and the memory (memory array) can be accessed as in the case where the HDD has not failed. It is possible to read correct data. That is, since the contents of the parity memory 14P appear to be mapped to the incorrect data portion of the memory array 14 via the memory access control unit 12, the disk array controller 6 of FIG. Can access.

【手続補正3】[Procedure amendment 3]

【補正対象書類名】明細書[Document name to be amended] Statement

【補正対象項目名】0112[Correction target item name] 0112

【補正方法】変更[Correction method] Change

【補正内容】[Correction contents]

【0112】ここで、本発明のメモリ装置において、メ
モリアレイをデータメモリとパリティメモリに分けて使
用し、データとパリティの比に基づいて一のアドレス空
間からの任意の連続したN個の単位ライトアクセスに対
してパリティを計算し、データメモリへのデータのライ
ト後に、パリティをパリティメモリに対してライトアク
セスすることにより、更に例えばHDD等の記録媒体の
故障などでデータが正しく読み出せない場合であって
も、データバッファメモリに対して通常時と等価なアク
セスを行うことが可能である。
Here, in the memory device of the present invention, the memory array is divided into a data memory and a parity memory and used, and N consecutive unit write operations from one address space are performed based on the ratio of data and parity. The parity is calculated for the access, and after the data is written to the data memory, the parity is written to the parity memory, so that the data cannot be read correctly due to a failure of a recording medium such as an HDD. Even if there is, it is possible to access the data buffer memory equivalent to the normal access.

【手続補正4】[Procedure amendment 4]

【補正対象書類名】明細書[Document name to be amended] Statement

【補正対象項目名】0113[Correction target item name]

【補正方法】変更[Correction method] Change

【補正内容】[Correction contents]

【0113】また、本発明のデータ記録再生装置におい
ては、本発明のメモリ装置を備え、複数の記録媒体から
なる記録媒体アレイを制御する記録媒体アレイ制御手段
と、記録媒体アレイと記録媒体アレイ制御手段との間の
データ送受を制御する記録媒体アレイインターフェイス
手段とが別々のアドレス空間を使用することにより、例
えばデータバッファメモリのリスト管理等の手間を減ら
すことができ、更に例えばHDD等の記録媒体の故障な
どでデータが正しく読み出せない場合であっても、デー
タバッファメモリに対して通常時と等価なアクセスを行
うことが可能である。
Further, the data recording / reproducing apparatus of the present invention includes a memory device of the present invention, a recording medium array control means for controlling a recording medium array comprising a plurality of recording media, a recording medium array and a recording medium array control. The use of a separate address space between the recording medium array interface means and the recording medium array interface means for controlling data transmission / reception between the means can reduce the trouble of, for example, managing a list in a data buffer memory, and further, for example, a recording medium such as an HDD. Even if the data cannot be read correctly due to a failure in the data buffer, it is possible to access the data buffer memory equivalent to the normal access.

【手続補正5】[Procedure amendment 5]

【補正対象書類名】明細書[Document name to be amended] Statement

【補正対象項目名】0114[Correction target item name]

【補正方法】変更[Correction method] Change

【補正内容】[Correction contents]

【0114】さらに、本発明のコンピュータシステムに
おいては、本発明のメモリ装置と本発明のデータ記録再
生装置とを備え、メモリ装置及びデータ記録再生装置を
インターコネクトを介して接続してなることにより、例
えばデータバッファメモリのリスト管理等の手間を減ら
すことができ、更に例えばHDD等の記録媒体の故障な
どでデータが正しく読み出せない場合であっても、デー
タバッファメモリに対して通常時と等価なアクセスを行
うことが可能である。
Further, the computer system of the present invention includes the memory device of the present invention and the data recording / reproducing device of the present invention, and the memory device and the data recording / reproducing device are connected via an interconnect. It is possible to reduce the trouble of managing the list of the data buffer memory, etc. Further, even if the data cannot be read correctly due to, for example, a failure of the recording medium such as the HDD, the equivalent access to the data buffer memory as usual is performed. It is possible to do.

Claims (15)

【特許請求の範囲】[Claims] 【請求項1】 単一又は複数のインターコネクトに接続
されるインターコネクト接続手段と、 物理メモリであるメモリアレイと、 上記メモリアレイを複数の個々のアドレス空間にマッピ
ングするマッピング設定手段と、 上記マッピングの設定に基づいて上記メモリアレイのア
クセスを制御するメモリアクセス制御手段とを有するこ
とを特徴とするメモリ装置。
1. An interconnect connection means connected to one or a plurality of interconnects; a memory array as a physical memory; a mapping setting means for mapping the memory array to a plurality of individual address spaces; And a memory access control means for controlling access to the memory array based on the memory device.
【請求項2】 上記メモリアレイをデータメモリとパリ
ティメモリに分けて使用し、 データとパリティの比を設定するデータ・パリティ比設
定手段と、 上記データとパリティの比の設定に基づいて、一のアド
レス空間からの任意の連続したN個の単位ライトアクセ
スに対してパリティを計算するパリティ演算手段とを設
け、 上記データメモリへのデータのライト後に、上記パリテ
ィを上記パリティメモリに対してライトアクセスするこ
とを特徴とする請求項1記載のメモリ装置。
2. A data / parity ratio setting means for setting a data / parity ratio using the memory array by dividing the data array into a data memory and a parity memory. A parity calculating means for calculating a parity for any continuous N unit write accesses from the address space; and after writing the data to the data memory, write the parity to the parity memory. The memory device according to claim 1, wherein:
【請求項3】 上記パリティメモリをサブメモリアレイ
として使用し、 上記マッピングの設定による上記メモリアレイに対する
任意の単位アクセスを対応するサブメモリアレイへのア
クセスに変換するサブメモリアレイマッピング設定手段
を設けることを特徴とする請求項2記載のメモリ装置。
3. A sub-memory array mapping setting means which uses the parity memory as a sub-memory array and converts an arbitrary unit access to the memory array by setting the mapping into an access to a corresponding sub-memory array. The memory device according to claim 2, wherein:
【請求項4】 単一又は複数のインターコネクトに接続
されるインターコネクト接続手段と、 物理メモリであるメモリアレイと、 上記メモリアレイの一部からなるサブメモリアレイ若し
くは別個のサブメモリアレイを使用し、設定により上記
メモリアレイに対する任意の単位アクセスを対応するサ
ブメモリアレイへのアクセスに変換するサブメモリアレ
イマッピング設定手段とを有することを特徴とするメモ
リ装置。
4. An interconnect connection means connected to one or a plurality of interconnects, a memory array which is a physical memory, and a setting using a sub memory array which is a part of the memory array or a separate sub memory array. And a sub memory array mapping setting means for converting an arbitrary unit access to the memory array into an access to a corresponding sub memory array.
【請求項5】 上記サブメモリアレイをパリティメモリ
として使用し、 データとパリティの比を設定するデータ・パリティ比設
定手段と、 上記データとパリティの比の設定に基づいて、一のアド
レス空間からの任意の連続したN個の単位ライトアクセ
スに対してパリティを計算するパリティ演算手段とを設
け、 データのライト後に、上記パリティを上記パリティメモ
リに対してライトアクセスすることを特徴とする請求項
4記載のメモリ装置。
5. A data / parity ratio setting means for setting a data / parity ratio using the sub memory array as a parity memory, and a data / parity ratio setting means for setting a data / parity ratio from one address space. 5. A parity calculating means for calculating a parity for any continuous N unit write accesses, wherein the parity is write-accessed to the parity memory after writing data. Memory device.
【請求項6】 物理メモリであるメモリアレイを複数の
個々のアドレス空間にマッピングすると共に、上記マッ
ピングの設定に基づいて上記メモリアレイのアクセスを
制御する可能なメモリ手段と、 複数の記録媒体からなる記録媒体アレイと、 上記記録媒体アレイを制御する記録媒体アレイ制御手段
と、 上記記録媒体アレイと上記記録媒体アレイ制御手段との
間のデータ送受を制御する記録媒体アレイインターフェ
イス手段とを有し、 上記記録媒体アレイ制御手段と記録媒体アレイインター
フェイス手段とが別々のアドレス空間を使用することを
特徴とするデータ記録再生装置。
6. A memory means capable of mapping a memory array as a physical memory to a plurality of individual address spaces and controlling access to the memory array based on the setting of the mapping, and a plurality of recording media. A recording medium array, recording medium array control means for controlling the recording medium array, and recording medium array interface means for controlling data transmission and reception between the recording medium array and the recording medium array control means; A data recording / reproducing apparatus, wherein the recording medium array control means and the recording medium array interface means use different address spaces.
【請求項7】 上記メモリ手段は、メモリアレイをデー
タメモリとパリティメモリに分けて使用し、データとパ
リティの比に基づいて、一のアドレス空間からの任意の
連続したN個の単位ライトアクセスに対してパリティを
計算し、データメモリへのデータのライト後に、パリテ
ィをパリティメモリに対してライトアクセスすることを
特徴とする請求項6記載のデータ記録再生装置。
7. The memory means uses a memory array by dividing it into a data memory and a parity memory, and performs arbitrary N unit write accesses from one address space based on a ratio of data and parity. 7. The data recording / reproducing apparatus according to claim 6, wherein a parity is calculated for the parity memory, and after the data is written to the data memory, the parity is write-accessed to the parity memory.
【請求項8】 上記メモリ手段は、上記パリティメモリ
をサブメモリアレイとして使用し、上記マッピングの設
定による上記メモリアレイに対する任意の単位アクセス
を対応するサブメモリアレイへのアクセスに変換するこ
とを特徴とする請求項7記載のデータ記録再生装置。
8. The memory means, wherein the parity memory is used as a sub-memory array, and an arbitrary unit access to the memory array by setting the mapping is converted into an access to a corresponding sub-memory array. The data recording / reproducing apparatus according to claim 7, wherein
【請求項9】 物理メモリであるメモリアレイの一部か
らなるサブメモリアレイ若しくは別個のサブメモリアレ
イを使用し、設定により上記メモリアレイに対する任意
の単位アクセスを対応するサブメモリアレイへのアクセ
スに変換可能なメモリ手段と、 複数の記録媒体からなる記録媒体アレイと、 上記記録媒体アレイを制御する記録媒体アレイ制御手段
と、 上記記録媒体アレイと上記記録媒体アレイ制御手段との
間のデータ送受を制御する記録媒体アレイインターフェ
イス手段とを有し、 上記記録媒体アレイ制御手段と記録媒体アレイインター
フェイス手段とが別々のアドレス空間を使用することを
特徴とするデータ記録再生装置。
9. A sub memory array which is a part of a memory array as a physical memory or a separate sub memory array is used, and an arbitrary unit access to the memory array is converted into an access to a corresponding sub memory array by setting. Possible memory means, a recording medium array comprising a plurality of recording media, a recording medium array control means for controlling the recording medium array, and controlling data transmission and reception between the recording medium array and the recording medium array control means A data recording / reproducing apparatus, comprising: a recording medium array interface means for performing a recording operation, wherein the recording medium array control means and the recording medium array interface means use different address spaces.
【請求項10】 上記メモリ手段は、上記サブメモリア
レイをパリティメモリとして使用し、データとパリティ
の比に基づいて、一のアドレス空間からの任意の連続し
たN個の単位ライトアクセスに対してパリティを計算
し、データメモリへのデータのライト後に、パリティを
パリティメモリに対してライトアクセスすることを特徴
とする請求項9記載のデータ記録再生装置。
10. The memory means uses the sub-memory array as a parity memory, and generates a parity for any continuous N unit write accesses from one address space based on a data / parity ratio. 10. The data recording / reproducing apparatus according to claim 9, wherein after calculating the data to the data memory, the parity is write-accessed to the parity memory.
【請求項11】 物理メモリであるメモリアレイを複数
の個々のアドレス空間にマッピングすると共に、上記マ
ッピングの設定に基づいて上記メモリアレイのアクセス
を制御するメモリ装置と、 複数の記録媒体からなる記録媒体アレイを制御する記録
媒体アレイ制御手段と、上記記録媒体アレイと上記記録
媒体アレイ制御手段との間のデータ送受を制御する記録
媒体アレイインターフェイス手段とを備え、上記記録媒
体アレイ制御手段と記録媒体アレイインターフェイス手
段とが別々のアドレス空間を使用するデータ記録再生装
置とを有し、 上記メモリ手段及びデータ記録再生装置をインターコネ
クトを介して接続してなることを特徴とするコンピュー
タシステム。
11. A memory device for mapping a memory array as a physical memory to a plurality of individual address spaces and controlling access to the memory array based on the setting of the mapping, and a recording medium comprising a plurality of recording media Recording medium array control means for controlling an array, and recording medium array interface means for controlling data transmission and reception between the recording medium array and the recording medium array control means, wherein the recording medium array control means and the recording medium array A computer system comprising: an interface unit and a data recording / reproducing device that uses different address spaces; and the memory unit and the data recording / reproducing device are connected via an interconnect.
【請求項12】 上記メモリ装置は、メモリアレイをデ
ータメモリとパリティメモリに分けて使用し、データと
パリティの比に基づいて、一のアドレス空間からの任意
の連続したN個の単位ライトアクセスに対してパリティ
を計算し、データメモリへのデータのライト後に、パリ
ティをパリティメモリに対してライトアクセスすること
を特徴とする請求項11記載のコンピュータシステム。
12. The memory device according to claim 1, wherein the memory array is divided into a data memory and a parity memory, and is used for an arbitrary number of consecutive unit write accesses from one address space based on a ratio of data and parity. 12. The computer system according to claim 11, wherein a parity is calculated for the parity, and after writing the data to the data memory, the parity is write-accessed to the parity memory.
【請求項13】 上記メモリ装置は、上記パリティメモ
リをサブメモリアレイとして使用し、上記マッピングの
設定による上記メモリアレイに対する任意の単位アクセ
スを対応するサブメモリアレイへのアクセスに変換する
ことを特徴とする請求項12記載のコンピュータシステ
ム。
13. The memory device, wherein the parity memory is used as a sub memory array, and an arbitrary unit access to the memory array by setting the mapping is converted into an access to a corresponding sub memory array. 13. The computer system according to claim 12, wherein
【請求項14】 物理メモリであるメモリアレイの一部
からなるサブメモリアレイ若しくは別個のサブメモリア
レイを使用し、設定により上記メモリアレイに対する任
意の単位アクセスを対応するサブメモリアレイへのアク
セスに変換可能なメモリ手段と、 複数の記録媒体からなる記録媒体アレイを制御する記録
媒体アレイ制御手段と、上記記録媒体アレイと上記記録
媒体アレイ制御手段との間のデータ送受を制御する記録
媒体アレイインターフェイス手段とを備え、上記記録媒
体アレイ制御手段と記録媒体アレイインターフェイス手
段とが別々のアドレス空間を使用するデータ記録再生装
置とを有し、 上記メモリ手段及びデータ記録再生装置をインターコネ
クトを介して接続してなることを特徴とするコンピュー
タシステム。
14. A sub memory array which is a part of a memory array which is a physical memory or a separate sub memory array, and an arbitrary unit access to the memory array is converted into an access to a corresponding sub memory array by setting. Possible memory means, recording medium array control means for controlling a recording medium array including a plurality of recording media, and recording medium array interface means for controlling data transmission and reception between the recording medium array and the recording medium array control means Wherein the recording medium array control means and the recording medium array interface means have a data recording / reproducing device using different address spaces, and the memory means and the data recording / reproducing device are connected via an interconnect. A computer system characterized in that:
【請求項15】 上記メモリ装置は、上記サブメモリア
レイをパリティメモリとして使用し、データとパリティ
の比に基づいて、一のアドレス空間からの任意の連続し
たN個の単位ライトアクセスに対してパリティを計算
し、データメモリへのデータのライト後に、パリティを
パリティメモリに対してライトアクセスすることを特徴
とする請求項14記載のコンピュータシステム。
15. The memory device according to claim 1, wherein the sub memory array is used as a parity memory, and a parity is provided for any continuous N unit write accesses from one address space based on a data / parity ratio. 15. The computer system according to claim 14, wherein is calculated, and after writing the data to the data memory, the parity is write-accessed to the parity memory.
JP10280316A 1998-10-01 1998-10-01 Memory device, data recording / reproducing device, and computer system Withdrawn JP2000112835A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP10280316A JP2000112835A (en) 1998-10-01 1998-10-01 Memory device, data recording / reproducing device, and computer system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP10280316A JP2000112835A (en) 1998-10-01 1998-10-01 Memory device, data recording / reproducing device, and computer system

Publications (1)

Publication Number Publication Date
JP2000112835A true JP2000112835A (en) 2000-04-21

Family

ID=17623305

Family Applications (1)

Application Number Title Priority Date Filing Date
JP10280316A Withdrawn JP2000112835A (en) 1998-10-01 1998-10-01 Memory device, data recording / reproducing device, and computer system

Country Status (1)

Country Link
JP (1) JP2000112835A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100470413B1 (en) * 2002-06-01 2005-02-05 솔리드 스테이트 시스템(주) Method for Partitioning Memory Mass Storage Device
JP2008065515A (en) * 2006-09-06 2008-03-21 Hitachi Ltd Computer system constituting remote I / O and I / O data transfer method

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100470413B1 (en) * 2002-06-01 2005-02-05 솔리드 스테이트 시스템(주) Method for Partitioning Memory Mass Storage Device
JP2008065515A (en) * 2006-09-06 2008-03-21 Hitachi Ltd Computer system constituting remote I / O and I / O data transfer method

Similar Documents

Publication Publication Date Title
JP3742494B2 (en) Mass storage device
JP3129732B2 (en) Storage array with copy-back cache
JP3304115B2 (en) Configurable redundant array storage
US6430702B1 (en) Fault tolerant memory
JP3181398B2 (en) Array type recording device
US5566316A (en) Method and apparatus for hierarchical management of data storage elements in an array storage device
US7984325B2 (en) Storage control device, data recovery device, and storage system
CN101681305B (en) Storage management system and method
JP2912802B2 (en) Disk array device failure handling method and device
US7861035B2 (en) Method of improving input and output performance of raid system using matrix stripe cache
US7464322B2 (en) System and method for detecting write errors in a storage device
US6832291B2 (en) Memory system for improving data input/output performance and method of caching data recovery information
JPH06208476A (en) Redundancy array parity cashing system
US5485598A (en) Redundant disk array (raid) system utilizing separate cache memories for the host system and the check data
US6343343B1 (en) Disk arrays using non-standard sector sizes
JPH06230903A (en) Disk array device failure recovery method and disk array device
JP2000112835A (en) Memory device, data recording / reproducing device, and computer system
JP3011020B2 (en) Array type disk system
KR100300836B1 (en) Data reconstruction method and data storage system
JP2857289B2 (en) Disk array device
JP3224741B2 (en) Data storage device
JPH1031563A (en) Storage device
GB2402803A (en) Arrangement and method for detection of write errors in a storage system
JP3128033B2 (en) Disk array device
JPH07328072A (en) Cache control method and information processing device

Legal Events

Date Code Title Description
A300 Withdrawal of application because of no request for examination

Free format text: JAPANESE INTERMEDIATE CODE: A300

Effective date: 20060110