560143 Λ; B7 8996pif.doc/008 五、發明說明（I ) 發明所屬的技術領域 、本發明係有成網路之複數台電腦的應用 程式、此應用程式所存取的常駐資料分散配置於個別電腦 上的網路系統、主記憶體常駐資料的管理方法以及實現此 管理方法之電腦可讀取的記憶媒體。 習知的技術 在電腦系統中，一般而言，爲了提升處理的速度， 採取將應用資料常駐在主記憶體的方法。而且，主記憶體 常駐資料的備份，通常儲存在碟式記憶體內，在電腦啓動 時’藉著由磁碟載入主記憶體常駐資料以設定初始値。 將主記憶體常駐資料寫入磁碟其目的係作爲備份， 由必須等待磁碟的輸入/輸出(Input/Output，I/O)而使得處理 速度降低的觀點，主記憶體常駐資料並不能每次都更新。 一般而言，主記憶體常駐資料寫入磁碟’係以一定 的週期進行，在主記憶體常駐資料寫入磁碟結束的時點， 主記憶體常駐資料與磁碟資料具有相同的狀態，其後，到 下一次的主記憶體常駐資料寫入之前，由於應用程式會對 主記憶體常駐資料進行更新，因此使得磁碟資料的狀態與 主記憶體常駐資料的狀態相差懸殊。 因此，在電腦啓動時，即使載入寫入於磁碟中的備 份資料以設定主記憶體，亦有可能無法復原電腦在停止時 點的主記憶體常駐資料。 … 而且，在分散電腦系統中，主記憶體常駐資料係由 4 本紙!Jk X 7艾適 W ’l· W .(I. (cnS)A 1 各（:M〇 X "Ά A ) I I I I I I I I · I — — I I I I ·Λ^_— I I I (請先Μ讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印於 560143 Λ7 H7 8996pif.doc/〇〇8 五、發明說明（7) 個別的電腦單位獨立管理，並無法確立於分散電腦系統全 體管理分散主記憶體常駐資料的結構。 (請先閱讀背面之注意事項再填寫本頁) 因此，作爲用以高速存取之主記憶體常駐資料’此 資料於存在的電腦上係爲高速進行區域的存取，關於藉由 網路之其他電腦所傳出的存取，則無法發揮主記憶體常駐 的局速存取特性。 發明所要解決的問題 本發明係用以解決上述的問題，本發明的目的係於 分散電腦系統中，在電腦啓·動時設定主記憶體常駐資料之 初始値時，並不依存預先寫入磁碟之主記憶體常駐資料的 備份資料，在提供一種設定主記憶體初始値之方法的同 時’並提供於分散電腦系統中，由應用程式隱蔽主記憶體 常駐資料存在於何處的電腦，並對應用程式提供主記憶體 常駐資料之可能限度的高速存取性。 解決問題的手段560143 Λ; B7 8996pif.doc / 008 V. Description of the invention (I) The technical field to which the invention belongs, the present invention is an application program with multiple computers in a network, and the resident data accessed by this application program is dispersedly arranged in individual Network system on computer, management method of resident data of main memory, and computer-readable storage medium implementing the management method. Conventional technology In computer systems, in general, in order to increase the processing speed, the application data is resident in the main memory. Moreover, the backup of the resident data in the main memory is usually stored in the disk memory. When the computer is started, the resident data in the main memory is loaded from the disk to set the initial volume. The purpose of writing the resident data of the main memory to the disk is as a backup. From the viewpoint of waiting for the input / output (I / O) of the disk to reduce the processing speed, the resident data of the main memory cannot Updated all times. Generally speaking, the main memory resident data writing to the disk is performed at a certain period. At the end of the main memory resident data writing to the disk, the main memory resident data and disk data have the same state. Later, before the next main memory resident data is written, the application program will update the main memory resident data, so the state of the disk data is very different from the state of the main memory resident data. Therefore, when the computer is started, even if the backup data written on the disk is set to set the main memory, the main memory resident data of the computer at the time of stopping may not be restored. … Moreover, in a decentralized computer system, the main memory resident data is composed of 4 papers! Jk X 7 Ai Shi W 'l · W. (I. (cnS) A 1 each (: M〇X " Ά A) IIIIIIII · I — — IIII Λ ^ _— III (please read the notes on the back before filling this page) Employee Consumer Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs printed on 560143 Λ7 H7 8996pif.doc / 〇〇8 V. Description of the invention ( 7) Individual computer units are managed independently and cannot be established in the structure of the distributed computer system to manage the resident data of the distributed main memory. (Please read the precautions on the back before filling this page.) Therefore, it is the master for high-speed access. Memory resident data 'This data is accessed at high speed on the existing computer. Regarding the access sent by other computers on the network, the local speed access characteristics of the main memory cannot be used. Problems to be Solved by the Invention The present invention is intended to solve the above-mentioned problems. The object of the present invention is to disperse the computer system. When the computer is started and started, the initial time of the resident data in the main memory is set, and it does not depend on the pre-written magnetic dish Backup data of the main memory resident data, while providing a method to set the initial size of the main memory, and provided in a distributed computer system, the application hides the computer where the main memory resident data exists, and The application provides the highest possible speed of access to the main memory resident data.
經濟部智慧財產局員工消費合作枝印S 爲了解決上述的問題，本發明申請專利範圍第1項 的網路系統’其特徵爲在構成網路之複數台電腦的應用程 式、此應用程式所存取的主記憶體常駐資料分散配置在個 別的電腦之網路系統中，於個別電腦具備有··接受經由應 用程式所傳出之交易要求的交易接受手段、實行所接受交 易要求的交易實行手段、將所接受交易要求發送至其 腦的交易發送手段、接收其他電腦所傳出之交易要求的= 5 560143 經濟部智慧財產局員工消費合作社印忽 8 9 9 6 p i f. d 〇 c / 0 0 8 五、發明說明（>)) 易接收手段、運用主記憶體常駐資料之所在位置資訊的系 統運用手段。 交易接受手段接受經由應用程式所傳出的交易要求 後，則將此些交易要求傳遞至交易實行手段。交易實行手 段係詢問系統運用手段以決定交易的實行方法，並將此交 易要求傳送至交易發送手段。交易發送手段係詢問系統運 用手段，以算出必須發送的.其他電腦，並對應此電腦的交 易接收手段發送交易要求。交易接收手段係將所接收之其 他電腦所傳出的女易要求傳送至交易實行手段。 依本發明的話，在構成分散系統的全部電腦中，由 於主記憶體常駐資料能夠一致化，而能夠由應用程式隱蔽 主記億體常駐資料，而且，由於頻繁進行更新之主記憶體 常駐資料的應用程式與更新對象的主記憶體常駐資料係配 置在相同的電腦，而能夠更高速的進行主記憶體資料的存 取。 *本發明申請專利範圍第2項的網路系統，在申請專 利範圍第1項所記載的網路系統中，更具備有用以保持由 其他電腦接收的交易要求連續性的第1交易回復手段。 第1交易接受手段接受經由應用程式所傳出的交易 要求後’則將此些交易要求傳遞至交易實行手段。交易實 行手段係詢問系統運用手段以決定交易的實行方法，並將 此交易要求傳送至交易發送手段。交易發送手段係詢問系 統運用手段，以算出必須發送的其他電腦，並對應此電腦 的交易接收手段發送交易要求。交易接收手段係將接收其 6 ^ 氏队乂^ 通川屮阄 W 家丨f; ·?- (CNS)A 1 (:MU X，):.父 ¥ ) ----Iw!!-^p^ !!— 訂丨丨.一丨丨丨線 (請先閲讀背面之注意事項再填寫本頁) 560143 8996pif.doc/008 A7 B7 經濟部智慧財產局員工消費合作社印製 五、發明說明（g:) 他電腦所傳出的交易要求傳送至交易實行手段。此處交易 的連續性產生問題的場合，交易實行手段使用交易回復手 段以回復交易的連續性。 依本發明的話’即使在發生網路故障妨害而遺失交 易的場合，能夠由遺漏的交易再度實行交易，由於能夠將 全部的電腦保持在相等數値的狀態，於分散電腦系統中能 夠提供更高可靠度的結構。· 本發明申請專利範圍第3項的網路系統，在申請專 利範圍第2項所記載的網路系統中，更具備有在電腦啓動 時將其他動作中電腦的主記憶體常駐資料爲基礎，設定爲 本身的主記憶體常駐資料之初始値的初始値設定手段。 在啓動電腦時，且需要設定主記憶體常駐資料之初 始値時’初始値設定手段將用以設定初始値的交易要求傳 送至交易發送手段。交易發送手段係詢問系統管理機能， 以算出必須發送的其他電腦，將交易要求發送至此電腦的 交易接收手段。 交易接收手段將接收的交易要求傳遞至初始値設定 手段。初始値設定手段基於要求形成設定主記憶體常駐資 料之初始値所必要的交易，並傳送至交易發送手段。 交易發送手段係將由交易實行手段所傳送、用以設 定主記憶體常駐資料之初始値的交易要求，發送至必須設 定主記憶體常駐資料初始値之電腦的交易接收手段。 交易接收手段將所接收之用以設定主記憶體常駐資 料之初始値的交易要求傳送至交易實行手段。交易實行手 7 本紙張尺度適用中國國家標準(CNS)A4HTS"x 297公釐) -----------裝----1---訂-------- (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印％ 560143 8996pif doc 008 Λ7 B7 五、發明說明（（） 段依照所傳送至的交易要求，實行本身的區域主記憶體常 駐資料相關的交易處理，以設定主記憶體常駐資料的初始 値。 依本發明的話，在構成分散電腦系統的複數台電腦 之間’由於在沒有磁碟的電腦亦可以設定主記憶體常駐資 料的初始値，主記憶體常駐資料可以互相使用備份資料。 而且，由於係由線上更新之其他電腦的主記億體常駐資料 以形成本身電腦的i記體常駐資料，而能夠將電腦啓動 ‘時的最新値，設定爲主記憶體常駐資料的初始値。 本發明申請專利範圍第4項的網路系統，在申請專 利範圍第3項所記載的網路系統中，在設定本身之主記憶 體常駐資料的初始値期間，更具備有儲存由其他電腦接收 之交易要求的交易保持手段。 在電腦啓動時，主記憶體常駐資料的初始値設定係 依照前述申請專利範圍第3項的方法進行。交易保持手段 在設定主記億體常駐資料的初始値期間’亦儲存由其他電 腦接收之交易要求，在主記憶體常駐資料的初始値設定結 束後，將儲存的交易要求傳送至交易實行處理。 依本發明的話，在電腦啓動時，在本身的主記憶體 常駐資料的初始値設定之際，由於在設定期間所接收的父 易要求亦不會遺失，而能夠確保交易的連續性。 本發明申請專利範圍第5項的網路系統’在申π円專 利範圍第3項與第4項其中任一項所記載的網路系統中， 其特徵爲更具備有在判斷第1交易回復手段無法回復時， 8 K—W---------------k---- (請先閱讀背面之注意事項再填寫本頁) 560143 H7 8996pif.doc/〇〇8 五、發明說明（g ) 使用初始値設定手段所形成的交易以進行交易回復的第2 交易回復手段。 依本發明的話，在發生暫時性的網路故障使得交易 要求發生遺漏，且第丨回復方法無法回復，而產生長時間 的交易要求遺漏場合，能夠提供回復的手段。 而且’在無法在實行交易時’亦能^使構成分散電 腦系統各電腦的主記憶體常駐資料的狀態一致化。 本發明申請專利範圍第6項的網路系統，在申請專 利範圍第1項的網路系統中，其特徵爲更具備有切換在任 口電腦更新初始的主gS憶體常駐資料之主記憶體常駐資 料更新權的更新權管理手段。 依本發明的話，持有主記憶體常駐資料的電腦系統 能夠建構多重化系統以及重複系統。 本發明申請專利範圍第7項記載之分散主記憶體資 料的管理方法，其特徵爲在具有應用程式與藉由應用程式 存取的主記憶體常駐資料，且管理互相以網路連接之複數 台電腦的各主記憶體常駐資料的方法中，其中一電腦接受 輕由應用程式所傳出的交易要求，以記憶有主記憶體常駐 貪料所在資訊爲基準，抽出必須發送交易要求的其他電 腦，向前述其他的電腦發送前述交易要求，前述其他電腦 接收前述一電腦所發送的交易要求，對前述其他電腦的主 胃己憶體常駐資料實行交易。 依本發明的話，能夠圓滑的管理具有應用程式與藉 由應用程式存取的主記憶體常駐資料，且管理互相以網路 9 (cns)ai cmo- ^ v ) -----------------------^---- (請先閲讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印於 經濟部智慧財產局員工消費合作社印巧 560143 8996pif.doc 008 ___ B7__ 五、發明說明（)） 連接之複數台電腦的各主記憶體常駐資料。 本發明申請專利範圍第8項記載之記憶媒體，其特 徵爲在應用程式與藉由應用程式存取的主記憶體常駐資 料，且用以管理互相以網路連接之複數台電腦的各主記憶 體常駐資料的電腦可讀取記憶媒體中，記憶有用以在電腦 實行：接受經由應用程式所傳出的交易要求之處理，以記 憶有主記憶體常駐資料所在賓訊爲基礎，抽出必須發送交 易要求的其他電腦之處理，向前述其他電腦發送前述接受 的交易<要求之處理，接收交易要求之處理，實行前述接收 的交易要求之處理的程式。 由於本發明係爲記憶有申請專利範圍第7項所記載 之主記憶體常駐資料的管理方法且電腦可讀取的記憶媒 體，依本發明的話，藉由由記憶媒體讀出的程式而能夠實 現主記億體資料管理系統。 圖式之簡單說明 第1圖所繪示爲構成本發明的網路裝置之各手段關 係的方塊圖； 第2圖所繪示爲第一實施例網路系統(區域更新）之方 塊圖的上段； 第3圖所繪示爲第一實施例網路系統（區域更新）之方 塊圖的下段； 第4圖所繪示爲第一實施例網路系統（遠端更新）之方 塊圖的上段； ------------in —--It·--— (請先閱讀背面之注意事項再填寫本頁> 本紙適川 t W W孓H?- (CNS)A丨 λ、V1 ) 8996pif.doc/008 ^ _ B / —, — 五、發明說明（¾ ) 第5圖所繪示爲第一實施例網路系統（遠端更新）之方 塊圖的下段； 第6圖所繪示爲第一實施例區域更新與遠端更新的 時序流程圖； 第7圖所繪示爲第二實施例之區域更新的交易回復 方法之方塊圖的上段； 第8圖所繪示爲第二賓施例之區域更新的交易回復 方法之方塊圖的下段； 第9圖所繪示爲第三實施例的主記憶體資料初始設 定方法之方塊圖的上段； 第10圖所繪示爲第三實施例的主記憶體資料初始設 定方法之方塊圖的下段； 第Π圖所繪示爲第三實施例的主記憶體常駐資料之 版本的時序流程圖方塊圖； 第12圖所繪示爲第四實施例的交易保持方法之方塊 圖的上段； 第13圖所繪示爲第四實施例的交易保持方法之方塊 圖的下段； 第Η圖所繪示爲第四實施例之主記憶體常駐資料的 初始設定中之交易要求處理方法的時序流程圖； 第I5圖所繪示爲第五實施例之交易長時間遺漏的回 復方法之方塊圖的上段； 第1 6圖所繪示爲第五實施例之交易長時間遺漏的回 復方法之方塊圖的下段； -------*1! | (請先閱讀背面之注意事項存填寫本頁) 本纸张K/i適用中阀四家標，(L'NS)Al叹恪GK) 公¥ ) 560143 Λ7 B7 8996pif.doc/008 五、發明說明（q ) 第17圖所繪示爲第六實施例之主記憶體常駐資料的 重複系統構成方法之方塊圖的上段；以及 第18圖所繪示爲第六實施例之主記憶體常駐資料的 重複系統構成方法之方塊圖的下段。 圖式之標記說明 1 :原始電腦（Original Computer) 2 :複製電腦（Replication computer) 3:具有更新權限的系統管理程式 4、111、211 :系統管理資料 11 :主記憶體常駐資料(原始） 12、 22 :交易接受佇列 13、 23 :交易接受程式 14、 24 :交易實行佇列 15、 25 :承諾等待佇列 * 16、26 :交易實行程式 17、 27 :交易發送佇列 18、 28 :交易發送程式 19、 29 :交易接收佇列 21 :主記憶體常駐資料(複製） 110、210 :交易接收程式 112、212 :應用程式 1Π、213 :交易遺漏檢測處理 114、214 :交易再建構處理 ---— — — — — — — — · I--111— (請先閱讀背面之注意事項再填寫本頁> 經濟部智慧財產局員工消費合作社印€ 本纸讥通川屮阄a家m a、Ns)..\ 1从格>上厂公*¥ ) 經濟部智慧財產局員工消費合作社印於 560143 8996pif.doc/008 __B7_ 五、發明說明（P ) 115、 215 :交易結束佇列 116、 216 :初始値設定程式 117、 217 :版本設定處理 118、 218 :初始値設定交易要求的交接處理 119、 219 :交易保持佇列 12〇、220 :初始値設定交易要求處理 3〇1 :交易接受手段 - 302 :交易實行手段 303 :交易發送手段 3 04 ·父易接收手段 305 :交易運用手段 306 :第1交易回復手段 307 :初始値設定手段 308 :交易保持手段 309 :第2交易回復手段 * 310 :主記憶體常駐資料的更新權限管理手段 較佳實施例 以下基於圖式對本發明的實施例做說明。 第1圖所繪示爲構成本發明的網路系統之各手段關 係的方塊圖。各手段的關係如下所詳述。 第一實施例 第2圖與第3圖所繪示爲本發明的網路系統之方塊 ---1 — lllllli^e^ is— II — — — (請先閲讀背面之注意事項再填寫本頁) 本纸依K度適川屮阀W家代(CNS)A 1 %格X 乂、V ) 經濟部智慧財產局員工消費合作社印於 560143 8996pif.doc/008 __]E_______ 五、發明說明（H ) 圖，且兩圖互相連接。第2圖係在複數台電腦構成的分散 電腦系統中，搭載有主記憶體原始資料11的唯一電腦1， 第3圖所繪示爲複製此原始資料，搭載有主記憶體複製資 料21的複數台電腦中的其中一台2。 在個別電腦中，交易接受手段301係由交易接受佇 列12、22以及交易接受程式13、23所構成。 交易實行手段302係由交易實行佇列14、24、承諾 等待佇列15、25以及交易實行手段16、26所構成。 交易發送手段303係由交易送信佇列Π、27以及交 易發送程式18、28所構成。 交易接收手段304係由交易接收佇列19、29與交易 接收程式110、210所構成。 系統運用手段305係由系統管理資料111、211所構 成。 電腦1與電腦2個別有主記憶體常駐資料11、21， 兩者的差別係在電腦1所搭載的主記憶體常駐資料11爲 原始値，此係基於應用程式的要求，進行最初的全部更新 要求，相對於此，電腦2所搭載的主記憶體常駐資料21 、爲複製値’在應用程式更新原始資料後，反映更新內容之 原始資料的複製。 在電腦1與電腦2中，個別具有存取主記憶體常駐 資料11與主記憶體常駐資料21的應用程式n2與應用程 式 212。 在主記憶體常駐資料爲原始値的電腦】中，應用程 14 本纸爪 度適川屮W W孓d- (CNS)A 1从袼C U)/八、¥ ) !! i I ! I I e (請先閲讀背面之注意事項再填窝本頁) 560143 五 Λ7 Β7 經濟部智慧財產局員工消費合作社印於 8996pif.doc ,〇〇8 發明說明（) 式112對主記憶體常駐資料11更新時所進行的區域更新 處理說明如下。 電腦1的應用程式112，將主記憶體常駐資料11的 更新作爲交易要求，輸入電腦1的交易接受佇列12。 電腦1的交易接受程式13，由電腦1的交易接受佇 列12取出應用程式112所輸入的交易要求，在輸入電腦1 的交易實行佇14的同時·，將要求交易之應用程式112 的資訊輸入電& 1的承諾等待佇列15。 電腦1的交易實行程式16,由電腦1的交易實行佇 列14取出交易要求，並對照系統管理資料in，以使交易 處理的對象瞭解本身的區域主記憶體常駐資料，並對本身 的區域主記憶體常駐資料(原始）11進行交易處理。 電腦1的父易實行程式16檢索電腦1的承諾等待 宁 列15，取出要求進行交易之應用程式112的資訊，將交易 結束的含意通知應用程式112。 •電腦1的交易實行程式16附上具有本身所實行交易 要求的記號，輸入電腦1的交易發送佇列17。 電腦1的交易發送程式18係參照記載有是否搭載原 始資料以及複製資料之電腦的系統管理資料m，算出必 須傳送交易要求的電腦亦即是搭載複製資料的電腦，對此 電腦的交易接收佇列29輸入由電腦1的交易發送丨宁列π 所取出的交易要求。 電腦2的交易接收程式210取出輸入於電腦1的交 易發送程式18中的交易要求，並輸入電腦2的交易實行 --— mllllll^w^ · I I---II — — — (請先閱讀背面之注意事項再填寫本頁) 560143 Λ7 B7 8996pif.doc/008 五、發明說明（丨冷） 料211，由於判定爲非對應本身的區域複製資料，本身不 實行交易，並將此要求輸入電腦2的交易發送佇列27。 電腦2的交易發送程式28係由電腦2的交易發送佇 列27取出交易要求，並參照系統管理資料211，對必須傳 送此交易要求的電腦，亦即是成爲交易對象、搭載原始資 料的電腦，對此電腦1的交易接收佇列19輸入交易要求。 電腦1的交易接收程式110取出輸入電腦2的交易發 送程式28中的交易要求，並輸入電腦1的交易實行佇列 14 〇 電腦1的交易實行程式16,由電腦1的交易實行佇 列14取出交易要求。所取出的交易要求對照系統管理資 料111，由於對本身的區域主記憶體常駐資料（原始）11判 定爲有交易要求，因此對本身的區域主記憶體常駐資料11 進行交易處理。 電腦1的交易實行程式16,本身附上實行之印記， ’以將交易要求輸入電腦1的交易發送佇列π。 電腦1的交易發送程式18係參照系統管理資料111， 算出必須傳送交易要求的電腦亦即是搭載複製資料的電 腦，對此電腦的交易接收佇列輸入由電腦1的交易發送佇 列17所取出的交易要求。 此交易要求所必須送達的電腦群中的其中一台即爲 電腦2。 電腦2的接收程式210取出輸入電腦1的交易發送程 式18中的交易要求，輸入電腦2的交易實行佇列24。 -----— — — — — — --------^» — — — — — — 1— ^ (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印於 560143 Λ7 H7 8996pif.doc/008 五、發明說明（丨、） 電腦2的交易實行程式26由電腦2的交易實行佇列 24取出交易要求，由於判定爲以電腦1進行的更新要求， 因此基於此交易要求，對本身的區域主記憶體常駐資料(複 製）21進行交易處理，並檢索電腦2的承諾等待佇列25， 取出要求進行交易之應用程式212的資訊，將交易結束此 含意通知應用程式112。以上爲遠端更新的說明。 上述區域更新與遠端更新的2個方法彙整繪示爲第6 圖的時序流程圖，並說明如下。 •首先，由搭載有原始値之主記憶體常駐資料11的電 腦1其應用程式II2，藉由更新交易要求，由電腦1之第 21號的交易要求’發送至搭載有複製値之主記憶體常駐資 料的電腦2至電腦η的全部電腦。 其次’藉由搭載有複製値之主記憶體常駐資料21的電 腦2’將交易要求發送至無號碼且搭載有原始値之主記憶 體常駐資料的電腦1。 •接收此交易要求，並由電腦1將第22號交易要求， 發送至搭載有複製値之主記憶體常駐資料的電腦2至電腦 η的全部電腦。 再者，由搭載有原始値之主記憶體常駐資料^的電 腦1，將第23號交易要求，發送至搭載有複製値之主記憶 體常駐資料的電腦2至電腦η的全部電腦。 依此藉由具有原始資料之電腦1的應用程式112對更 新的主記憶體常駐資料（原始値）11的變更，反映至其他全 部電腦的主記憶體常駐資料（複製値）21，而且來自具有複 18 _______I-----------^----11111^ (請先閲讀背面之注意事項再填寫本頁) 濟 部 智 慧 財 產 局 員 工 消 f 560143 Λ7 B7 8996pif. doc/〇〇^ 五、發明說明（A ) 製資料之電腦2等的應用程式的交易要求’首先以原始的 電腦1處理，其後反映至包含電腦2的全部電腦。 (請先閱讀背面之注意事項再填寫本頁) 依此，能夠實現即使由應用程式並不知道與有初始 主記憶常駐資料的場所，亦能夠進行此存取的環璘° 而且，複製資料的更新交易，首先對原始資料進行 最初的更新交易，再將此結果反映至複製資料，爲單方向 的反映方法。 因此，構成分散系統之電腦的應用程式所爹求的更 -新交易要求，在具有原始資料的電腦，依接受順序實行的 話較佳，由於實行時不需進行調停，而能夠實現高速的交 易環境。 第二實施例 依第7圖以及第8圖，說明本發明的主記憶體資料 的區域更新之交易回復方法。尙且，第7圖以及第8圖互 相連接。 經濟部智慧財產局員工消費合作社印緊 兩圖係在第2圖以及第3圖說明的進行區域處理的 狀況中，由網路因爲暫時性的故障損害所導致交易要求遺 漏或是順序錯亂的狀況中回復交易的方法。 於第7圖與第8圖中，第1交易回復手段306係由 交易實行程式一部分的交易遺漏檢測處理113、213，以及 遺漏交易以後的交易再建構處理114、214，以及交易結束 佇列115、215所構成。 電腦1的交易實行程式16在交易實行結束後將實行 19 適用中阀四家捸莩（CNS)Al叹格公¥ ) 經濟部智慧財產局員工消費合作社印纪 560143 8996pif.doc/008 ___B7_____ 五、發明說明（Π) 的交易要求輸入交易結束佇列115。 以應用程式1處理交易要求的狀態，如同前述第2 圖以及第3圖所說明，於此處未圖示。 以電腦1的交易實行程式16所處理的交易要求，藉 由電腦1的交易發送程式18，發送至電腦2的交易接收程 式 210 〇 由此電腦1的交易發送程式18往電腦2的交易接收 程式210之交易要求通信，係藉由網路進行，因此交易要 求可能由於網路的暫時性故障損害而在網路上遺失。 本實施例具有主記憶體常駐資料之電腦的交易實行 程式’在對本身的區域主記憶體常駐資料進行處理的交易 要求’附上表示交易實行順序的交易編號。因此，交易要 求遺漏或者是順序錯亂可藉由編號的遺漏以檢測出。 具體而言，在交易實行程式中，由於係追加調查交 易連續性的處理，藉由電腦1的交易實行程式16附上編 號的交易要求，一旦發生遺漏則能夠由電腦2的交易實行 程式26檢測出。 電腦2的交易實行程式26，一旦檢測出交易要求的 遺漏，則將此意旨經由電腦2的交易送信程式28以及電 腦1的交易接收程式110,通知電腦1的交易實行程式16。 電腦1的交易實行程式16，由電腦2的交易實行程 式26，接獲交易要求遺漏的通知，則由電腦丨的交易結束 佇列115中取出遺漏的交易要求以後的交易要求，經由電 腦1的交易發送程式18以及電腦2的交易接收程式210， 20 本紙队K/免滷丨丨]f W Η家(CNS)A 1从恪公¥ ) --------------------1-------— (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印€ 560143 8996pif.doc/008 1 _H7_ 五、發明說明u&) 傳送至電腦2的交易實行程式26。 電腦2的交易實行程式26處理由電腦1的交易實行 程式16所傳送之遺漏的交易要求以後的交易要求。 先前第4圖以及第5圖所說明的，在以遠端更新交 易要求的場合，由於亦由具有原始主記憶體常駐資料之電 腦1的交易實行程式16附上交易編號，交易的遺漏係由 電腦2的交易實行程式26·檢測出，其後的回復處理，亦 即是遺漏的交易票求以後之交易要求的處理，，與第7圖以 及第8圖所說明的區域更新的交易回復場合相同。 依照以上的方法，藉由主記憶體常駐資料的區域更 新以及遠端更新，由應用程式隱蔽主記憶體的所在位置之 主記憶體資料管理方法中，在將原始資料與複製資料保持 在相同內容的結構中，當發生網路故障損害等而導致交易 遺失的場合，由於能夠由遺失的交易再度實行交易，因而 能夠保持原始資料與複製資料的內容一致。 第三實施例 依照第9圖與第10圖，說明本發明之主記憶體常駐 資料的初始値設定方法。 尙且，第9圖與第10圖係互相連接。 於兩圖中，主記憶體常駐資料的初始値設定手段307 係由初始値設定程式Π6、216，交易實行程式一部份的版 本設定處理117、217，以及交易接收程式一部份的初始値 設定交易要求的交接處理118、218所構成。 --------------------^----1---- (請先閱讀背面之注意事項再填寫本頁) __ 本纸汛K度遶川屮阀闷孓標羋（CNSM 1 乂恪UM0、上厂〉> ¥ ) 560143 8996pif.doc/008 Λ7 B7 五 經濟部智慧財產局員工消費合作社印纪 、發明說明（θ) 電腦1與電腦2係爲如同前述第一實施例所說明的 進行遠端更新與區域更新。 在啓動電腦2的場合，電腦2的初始値設定程式216， 係藉由參照系統管理資料211，決定是否由構成分散電腦 系統的其他電腦中之任意電腦，設定主記憶體常駐資料的 初始値，用以設定主記憶體常駐資料21之初始値的交易 要求，輸入電腦2的交易發送佇列27。 此處藉由參照系統管理資料211，藉由電腦1的主記 憶體常駐資料，設定電腦2之主記憶體常駐資料的初始値。 電腦2的交易發送程式28，藉由交易發送佇列27將 輸入電腦2的初始値設定程式216中之用以設定初始値的 交易要求取出，並傳送茔電腦1的交易接收程式110。 電腦1的交易接收程式110由電腦1的交易接收佇列 19取出用以設定電腦2之主記憶體常駐資料21之初始値 的交易要求，並傳送至電腦1的初始値設定程式116。 •電腦1的初始値設定程式116係基於本身的區域主記 憶體常駐資料11，以形成用以設定主記憶體常駐資料11 之初始値的交易，並輸入電腦1的交易發送佇列17。 此處藉由依照初始値設定程式116所形成之設定主記 憶體常駐資料I1之初始値的交易要求，主記憶體常駐資 料11的版本編號亦爲一致設定。 主記憶體常駐資料的版本’係藉由構成初始値設定 手段之交易實行程式一部份的主記憶體常駐資料之版本設 定處理117以設定。 (請先閲讀背面之注意事項再填寫本頁) --------^--------- 本紙ik丨丨丨屮阀闷孓丨ϋ (cNS)A1从格UM() 公¥ } 560143In order to solve the above-mentioned problems, the network system of the first patent application scope of the present invention is characterized by an application program on a plurality of computers constituting a network, and the application program stored in this application program. The resident data taken from the main memory is distributed and arranged in the network system of individual computers. Each individual computer has a transaction acceptance method that accepts the transaction request transmitted through the application program, and a transaction execution method that implements the accepted transaction request. 、 The sending method of sending the accepted transaction request to his brain, and receiving the transaction request sent by other computers = 5 560143 Imprint of the employee consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 8 9 9 6 pi f. D oc / 0 0 8 V. Description of the invention (>)) Easy-to-receive means, system operation means using location information of resident data of main memory. After the transaction acceptance means accepts the transaction requests sent out through the application program, it passes these transaction requests to the transaction execution means. The transaction execution means is to inquire the system to use the method to determine the transaction execution method, and transmit the transaction request to the transaction sending means. The transaction sending means is to query the system using the means to calculate the other computers that must be sent, and send the transaction request corresponding to the computer's transaction receiving means. The means of receiving the transaction is to transmit the request from the female computer received from other computers to the means of implementing the transaction. According to the present invention, in all the computers constituting the decentralized system, since the resident data of the main memory can be unified, the main program resident data can be hidden by the application program, and because the resident data of the main memory is frequently updated, The application and the main memory resident data of the update target are arranged on the same computer, and the main memory data can be accessed at a higher speed. * The network system according to the second patent application scope of the present invention includes the first network transaction recovery method that is useful to maintain the continuity of transaction requirements received by other computers in the network system described in the first patent application scope. After the first transaction acceptance means accepts the transaction request transmitted through the application program ', these transaction requests are passed to the transaction execution means. The transaction execution method is to query the system to use the method to determine the execution method of the transaction, and transmit the transaction request to the transaction sending method. The transaction sending means is to query the system to use the means to calculate other computers that must be sent, and send the transaction request corresponding to the computer's transaction receiving means. The transaction receiving means will receive its 6 ^ clan 乂 ^ 通 川 屮 阄 W home 丨 f; ·?-(CNS) A 1 (: MU X,) :. Father ¥) ---- Iw !!-^ p ^ !! — Order 丨 丨. 一 丨 丨 丨 line (please read the precautions on the back before filling this page) 560143 8996pif.doc / 008 A7 B7 Printed by the Intellectual Property Bureau Employees Consumer Cooperatives of the Ministry of Economic Affairs g :) The transaction request from his computer is transmitted to the transaction execution means. If there is a problem with the continuity of the transaction, the transaction implementation method uses the transaction reply method to restore the continuity of the transaction. According to the present invention, 'Even in the case of a network failure and a lost transaction, the transaction can be re-executed from the missing transaction. Since all computers can be maintained in the same state, it can provide higher levels in a distributed computer system. Structure of reliability. · The network system according to the third patent application scope of the present invention, and the network system described in the second patent application scope, is further provided with resident data of the main memory of the computer during other operations when the computer is started. It is set as the initial setting method of the initial data of the main memory resident data. When the computer is started, the initial time of the resident data in the main memory needs to be set. The “initial time” setting means transmits the transaction request for setting the initial time to the transaction sending means. The transaction sending means is a transaction receiving means that asks the system management function to calculate the other computers that must be sent, and sends the transaction request to this computer. The transaction receiving means passes the received transaction request to the initial setting means. The initial volume setting means is based on the transaction required to form the initial volume for setting the resident data of the main memory, and transmits it to the transaction sending means. The transaction sending means refers to the transaction receiving means transmitted by the transaction execution means for setting the initial data of the main memory resident data to the computer of the transaction receiving means that must set the initial data of the main memory resident data. The transaction receiving means transmits the received transaction request for setting the initial value of the resident data of the main memory to the transaction executing means. Handling of transactions 7 The paper size is applicable to China National Standard (CNS) A4HTS " x 297 mm) ----------- installation ---- 1 ----- order -------- (Please read the precautions on the back before filling this page) Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 560143 8996pif doc 008 Λ7 B7 V. Description of the Invention (() Paragraph shall be implemented in accordance with the transaction requirements to which it is transmitted. Memory resident data-related transactions are processed to set the initial resident data of the main memory. According to the present invention, the main memory can be set between a plurality of computers constituting a decentralized computer system. At the beginning of the resident data, the resident data in the main memory can use the backup data with each other. Moreover, since the resident data of the main body of the other computer updated online forms the resident data of the i-computer of the own computer, the computer can be started The latest time of "Time" is set as the initial time of the resident data of the main memory. The network system of the fourth patent application scope of the present invention is set in the network system described in the third patent application scope. During the initial period of the resident data of the main memory, there is also a means for maintaining transactions for storing transaction requests received by other computers. When the computer is started, the initial setting of the resident data of the main memory is in accordance with item 3 of the aforementioned patent application scope. The transaction holding means also stores the transaction requests received by other computers during the initial period of setting the resident data of the main memory. After the initial setting of the resident data of the main memory is completed, the stored transaction request is transmitted to The transaction is processed. According to the present invention, when the computer is started, at the time of the initial setting of the resident data of its own main memory, since the parent request received during the setting period will not be lost, it can ensure the continuous transaction. The network system according to the fifth patent application scope of the present invention is characterized in that the network system described in any one of the third and fourth patent application scopes is further provided with the first judgment 8 K—W --------------- k ---- (Please read the precautions on the back before filling this page) 560143 H7 899 6pif.doc / 〇〇8 V. Description of the invention (g) The second transaction recovery method for transaction recovery using the initial transaction setting method. According to the present invention, a temporary network failure makes the transaction request Occurrence of omissions, which cannot be answered by the first reply method, and long-term transaction request omissions, can provide a means of reply. Moreover, the main memory of each computer constituting the decentralized computer system can be used 'when the transaction cannot be performed'. The status of the resident data is consistent. The network system according to the present invention in the scope of patent application No. 6 is characterized in that the network system in the first scope of the patent application is further equipped with a master gS memory that can be updated at any computer. Renewal right management method of resident data update right of main memory. According to the present invention, a computer system holding main memory resident data can construct a multiplexed system and a repetitive system. The method for managing decentralized main memory data described in item 7 of the scope of patent application of the present invention is characterized in that there is an application program and main memory resident data accessed by the application program and manages a plurality of computers connected to each other through a network. In the method of resident data of the main memory of the computer, one of the computers accepts the transaction request sent by the application lightly, and uses the information of the resident information stored in the main memory as a reference to extract other computers that must send a transaction request. The aforementioned transaction request is sent to the aforementioned other computer, and the aforementioned other computer receives the transaction request sent by the aforementioned computer, and executes the transaction on the resident data of the main computer of the aforementioned other computer. According to the present invention, it is possible to smoothly manage the resident data with the application and the main memory accessed by the application, and manage each other through the network 9 (cns) ai cmo- ^ v) -------- --------------- ^ ---- (Please read the notes on the back before filling out this page) Employee Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs printed on the employees' consumption of the Intellectual Property Bureau of the Ministry of Economic Affairs Cooperative cooperative printing 560143 8996pif.doc 008 ___ B7__ V. Description of the invention ()) Resident data of each main memory of multiple computers connected. The memory medium described in item 8 of the scope of patent application of the present invention is characterized by resident data in an application program and a main memory accessed by the application program, and is used to manage each main memory of a plurality of computers connected to each other through a network. In the computer-readable storage medium of the resident data, the memory is useful to implement on the computer: accept the processing of the transaction request sent through the application, based on the guest message where the resident data of the main memory is stored, and the transaction must be extracted and sent The processing of the other requested computer sends the aforementioned accepted transaction < requested processing to the aforementioned other computer, the processing of receiving the transaction request, and a program for implementing the processing of the aforementioned received transaction request. Since the present invention is a computer-readable storage medium for storing the resident data of the main memory as described in item 7 of the scope of patent application, according to the present invention, it can be realized by a program read from the storage medium. Master billion data management system. Brief Description of the Drawings Fig. 1 is a block diagram showing the relationship between various means constituting the network device of the present invention; Fig. 2 is a diagram showing the upper part of the block diagram of the network system (area update) of the first embodiment. ; Figure 3 shows the lower section of the block diagram of the network system (regional update) of the first embodiment; Figure 4 shows the upper section of the block diagram of the network system (remote update) of the first embodiment; ------------ in —-- It · --— (Please read the notes on the back before filling out this page> This paper is suitable for Sichuan WW 孓 H?-(CNS) A 丨 λ V1) 8996pif.doc / 008 ^ _ B / —, — V. Description of the invention (¾) Figure 5 shows the lower section of the block diagram of the network system (remote update) of the first embodiment; Figure 6 The sequence diagram of the area update and remote update of the first embodiment is shown in FIG. 7. The upper part of the block diagram of the transaction reply method of the area update of the second embodiment is shown in FIG. 7. The diagram in FIG. 8 is shown in FIG. The lower part of the block diagram of the transaction update method for the area update of the second guest example; FIG. 9 shows the initial setting method of the main memory data of the third embodiment Upper part of the block diagram; FIG. 10 shows the lower part of the block diagram of the initial setting method of the main memory data of the third embodiment; FIG. Π shows the version of the main memory resident data of the third embodiment Figure 12 is a block diagram of the timing chart of the fourth embodiment; Figure 12 is the upper section of the block diagram of the transaction holding method of the fourth embodiment; Figure 13 is the lower section of the block diagram of the transaction holding method of the fourth embodiment; The diagram shows a time sequence flow chart of the transaction request processing method in the initial setting of the main memory resident data of the fourth embodiment. Figure I5 shows the method of replying to the long missed transaction of the fifth embodiment. Upper part of the block diagram; Figure 16 shows the lower part of the block diagram of the method of replying to the long-missed transaction in the fifth embodiment; ------- * 1! | (Please read the precautions on the back first Save and fill in this page) This paper K / i is applicable to the four valves of the Chinese valve, (L'NS) Al Tan Ke GK) Public ¥) 560143 Λ7 B7 8996pif.doc / 008 V. Description of the invention (q) Drawing in Figure 17 A method of constructing a repeating system of resident data of the main memory shown in the sixth embodiment FIG upper block segment; and FIG. 18 depicted as main memory resident data of the sixth embodiment of the repeated block diagram showing the system configuration of the lower part of the method. Explanation of the mark of the drawing 1: Original Computer 2: Replication computer 3: System management program with update authority 4, 111, 211: System management data 11: Main memory resident data (original) 12 22: Transaction Acceptance Queue 13, 23: Transaction Acceptance Queue 14, 24: Transaction Execution Queue 15, 25: Commitment Waiting Queue * 16, 26: Transaction Execution Queue 17, 27: Transaction Send Queue 18, 28: Transaction sending program 19, 29: Transaction receiving queue 21: Main memory resident data (copy) 110, 210: Transaction receiving program 112, 212: Application 1Π, 213: Transaction omission detection processing 114, 214: Transaction reconstruction processing ---— — — — — — — — — I--111— (Please read the notes on the back before filling out this page > Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs ma, Ns) .. \ 1 Congge > Shangchanggong * ¥) Printed by the Consumers' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs at 560143 8996pif.doc / 008 __B7_ V. Description of the invention (P) 115, 215: Transaction completion queue 116, 216: Initial setting program 11 7, 217: Version setting processing 118, 218: Initial transaction setting transfer request processing 119, 219: Transaction holding queue 12 and 220: Initial transaction setting transaction request processing 301: Transaction acceptance means-302: Transaction execution Means 303: Transaction sending means 3 04Father easy receiving means 305: Transaction operation means 306: First transaction response means 307: Initial transaction setting means 308: Transaction holding means 309: Second transaction response means * 310: Main memory resident Preferred Embodiments of Means for Managing Authorization of Updating Data The following describes embodiments of the present invention based on drawings. FIG. 1 is a block diagram showing the relationship between various means constituting the network system of the present invention. The relationship between the means is described in detail below. Figures 2 and 3 of the first embodiment show the blocks of the network system of the present invention --- 1 — lllllli ^ e ^ is— II — — — (Please read the precautions on the back before filling this page ) This paper is suitable for Kawasaki valve (CNS) A 1% grid X 乂, V according to K degrees.) Printed on 560143 8996pif.doc / 008 __] E_______ by the Consumers ’Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs. ) Graph, and the two graphs are connected to each other. Fig. 2 is the only computer 1 in the decentralized computer system composed of a plurality of computers, which is loaded with the main data 11 of the main memory. One of the two computers 2. In some computers, the transaction acceptance means 301 is composed of transaction acceptance queues 12, 22 and transaction acceptance programs 13, 23. The transaction execution means 302 is composed of transaction execution queues 14, 24, commitment waiting queues 15, 25, and transaction execution means 16, 26. The transaction sending means 303 is composed of a transaction sending queue Π, 27 and a transaction sending program 18, 28. The transaction receiving means 304 is composed of the transaction receiving queues 19 and 29 and the transaction receiving programs 110 and 210. The system operation means 305 is composed of system management data 111 and 211. Computer 1 and Computer 2 each have main memory resident data 11, 21. The difference between the two is that the main memory resident data 11 carried in computer 1 is original. This is based on the requirements of the application program, and all the initial updates are performed. It is required that, on the other hand, the main memory resident data 21 carried in the computer 2 is a copy of the original data that reflects the updated content after the application updates the original data. The computer 1 and the computer 2 each have an application program n2 and an application program 212 that access the main memory resident data 11 and the main memory resident data 21. In a computer with main memory resident data as the original data], the application process is 14 paper claws, suitable for WW 孓 d- (CNS) A 1 from 袼 CU) //, ¥) !! i I! II e ( (Please read the precautions on the back before filling in this page) 560143 Five Λ7 Β7 Printed at 8996pif.doc, Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, 〇08 Description of the invention () Formula 112 is used to update the resident data of the main memory 11 The area update processing performed is described below. The application 112 of the computer 1 uses the update of the main memory resident data 11 as a transaction request, and enters the transaction acceptance queue 12 of the computer 1. Computer 1's transaction accepting program 13, the computer 1's transaction accepting queue 12 fetches the transaction request entered by the application 112, while inputting the computer 1's transaction execution 伫 14, · enters the information of the application 112 requesting the transaction Electric & 1's promise awaits queue 15. Computer 1's transaction execution program 16 takes the computer 1's transaction execution queue 14 to retrieve the transaction requirements, and compares the system management data in, so that the object of transaction processing knows the resident data of its own local main memory, and Memory resident data (raw) 11 for transaction processing. The parent computer of the computer 1 executes the program 16 to retrieve the promise of the computer 1 and waits for the queue 15 to retrieve the information of the application 112 requesting the transaction, and inform the application 112 of the end of the transaction. • The transaction execution program 16 of the computer 1 is attached with a token having a transaction request executed by itself, and is entered into the transaction transmission queue 17 of the computer 1. The transaction sending program 18 of the computer 1 refers to the system management data m of the computer that records whether the original data and the copied data are installed, and calculates the computer that must transmit the transaction request, that is, the computer equipped with the copied data. 29 Enter the transaction request with the transaction sent from Ninglie π from computer 1. The transaction receiving program 210 of the computer 2 takes out the transaction request inputted into the transaction sending program 18 of the computer 1 and enters the transaction execution of the computer 2 --- mllllll ^ w ^ · I I --- II — — — (Please read first Note on the back, please fill out this page again) 560143 Λ7 B7 8996pif.doc / 008 V. Description of the invention (丨 cold) Material 211, because it is judged to be non-corresponding to the area to copy the data, the transaction itself is not implemented, and this requirement is entered into the computer Transaction 2 of queue 27. The transaction sending program 28 of the computer 2 is to take out the transaction request from the transaction sending queue 27 of the computer 2 and refer to the system management data 211 to the computer that must transmit the transaction request, that is, the computer that becomes the transaction object and is equipped with the original data. Enter a transaction request for the transaction receiving queue 19 of this computer 1. The transaction receiving program 110 of the computer 1 takes out the transaction request entered in the transaction sending program 28 of the computer 2 and enters the transaction execution queue 14 of the computer 1. The transaction execution program 16 of the computer 1 is taken out by the transaction execution queue 14 of the computer 1. Transaction requirements. The retrieved transaction requirements are compared with the system management information 111. Since the local main memory resident data (original) 11 is determined to have transaction requirements, the local main memory resident data 11 is subject to transaction processing. The transaction execution program 16 of the computer 1 is attached with a stamp of execution, ′ to input the transaction request into the transaction sending queue π of the computer 1. The transaction sending program 18 of computer 1 refers to the system management data 111, and calculates the computer that must transmit the transaction request, that is, the computer equipped with the copied data. The transaction receiving queue input for this computer is taken out by computer 1's transaction sending queue 17 Transaction requirements. One of the computer groups that this transaction requires to be delivered is computer 2. The receiving program 210 of the computer 2 takes out the transaction request entered in the transaction sending program 18 of the computer 1, and enters the transaction execution queue 24 of the computer 2. ----- — — — — — — — -------- ^ »— — — — — — 1— ^ (Please read the notes on the back before filling out this page) Staff Consumption of Intellectual Property Bureau, Ministry of Economic Affairs Cooperatives printed on 560143 Λ7 H7 8996pif.doc / 008 V. Description of the Invention (丨,) Computer 2's transaction execution program 26 takes the computer 2's transaction execution queue 24 to take out the transaction request. Because it is determined to be an update request by computer 1, Therefore, based on this transaction request, perform transaction processing on the local main memory resident data (copy) 21, and retrieve the commitment of computer 2 to wait for queue 25, retrieve the information of the application 212 requesting the transaction, and end the transaction. Notification application 112. The above is the description of the remote update. The above two methods of regional update and remote update are shown as a sequence flow chart in FIG. 6 and described below. • First, the application program II2 of computer 1 equipped with resident data 11 of the main memory of the original 値 is sent to the main memory equipped with the copied 値 by the transaction request No. 21 of the computer 1 by updating the transaction request. All computers from computer 2 to computer resident. Secondly, 'the computer 2 equipped with the main memory resident data 21 of the copied volume is used to send the transaction request to the computer 1 having no number and the main memory resident data of the original volume. • Receive this transaction request, and computer 1 sends the transaction request No. 22 to all the computers from computer 2 to computer η equipped with the resident data of the main memory of the copy card. Furthermore, the computer 1 equipped with the resident data of the main memory of the original memory ^ sends the transaction request No. 23 to all the computers of the computer 2 to the computer η equipped with the resident data of the copied main memory of the memory. Accordingly, the application 112 of the computer 1 with the original data changes the updated main memory resident data (original data) 11 to the main memory resident data (reproduction data) 21 of all other computers. Reply 18 _______ I ----------- ^ ---- 11111 ^ (Please read the notes on the back before filling in this page) Employees of the Ministry of Economic Affairs and Intellectual Property Office f 560143 Λ7 B7 8996pif.doc / 〇 Ⅴ. V. Description of Invention (A) The transaction requirements of applications such as computer 2 which make data are processed first by the original computer 1, and then reflected to all computers including computer 2. (Please read the precautions on the back before filling in this page.) With this, it is possible to realize this access even if the application does not know and the place where the initial main memory is resident. Also, the copying of data Update transactions. First, the original update transaction is performed on the original data, and then the result is reflected to the copied data, which is a one-way reflection method. Therefore, the new-and-new transaction requirements demanded by the application programs of the computers constituting the decentralized system are better to implement in the order of acceptance on the computers with the original data. Since no mediation is required during the implementation, a high-speed transaction environment can be realized. . Second Embodiment A method of transaction recovery for area update of main memory data according to the present invention will be described with reference to Figs. 7 and 8. In addition, Figures 7 and 8 are connected to each other. The two pictures printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs are in the situation of regional processing illustrated in Figures 2 and 3, and the transaction request is missed or out of order due to temporary faults on the Internet. Method of replying to transaction. In Figures 7 and 8, the first transaction response means 306 is a transaction omission detection process 113, 213 which is part of the transaction execution program, and a transaction reconstruction process 114, 214 after the missing transaction, and a transaction end queue 115. , 215. Computer 1's transaction execution program 16 will be implemented after the transaction execution is completed. 19 Applicable to Zhongwan Shijiajia (CNS) Al Tange ¥) Employees ’Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, India 560143 8996pif.doc / 008 ___B7_____ V. The transaction of invention (Π) requires entry of a transaction end queue 115. The status of processing the transaction request by the application program 1 is not illustrated here, as described in the foregoing FIG. 2 and FIG. 3. The transaction request processed by the computer 1's transaction execution program 16 is transmitted to the computer 2's transaction receiving program 210 via the computer 1's transaction sending program 18 〇 From the computer 1's transaction sending program 18 to the computer 2's transaction receiving program The 210 transaction request communication is performed through the Internet, so the transaction request may be lost on the network due to the temporary failure of the network. In the present embodiment, the computer-implemented transaction execution program of the main memory resident data 'attaches a transaction number indicating the execution order of the transaction to the transaction request for processing the main memory resident data in its own area. Therefore, omission of transaction requirements or out of order can be detected by omission of numbers. Specifically, in the transaction execution program, since the transaction continuity is additionally investigated, the transaction execution program 16 of the computer 1 is attached with a numbered transaction request. Once the omission occurs, it can be detected by the transaction execution program 26 of the computer 2. Out. The transaction execution program 26 of the computer 2 is used to notify the transaction execution program 16 of the computer 1 via the transaction transmission program 28 of the computer 2 and the transaction reception program 110 of the computer 1 once the omission of the transaction request is detected. The transaction execution program 16 of computer 1 and the transaction execution program 26 of computer 2 receive the notification of missing transaction request, and the missing transaction request is taken out from the transaction end queue 115 of computer 丨 and the subsequent transaction requests are transmitted through computer 1 Transaction sending program 18 and computer 2's transaction receiving program 210, 20 Paper team K / halogen-free 丨 丨] f W Η 家 (CNS) A 1From the official ¥) ------------- ------- 1 -------— (Please read the notes on the back before filling out this page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economy € 560143 8996pif.doc / 008 1 _H7_ V. Invention Explanation u &) The transaction execution program 26 transmitted to the computer 2. The transaction execution program 26 of the computer 2 processes the missing transaction request transmitted by the transaction execution program 16 of the computer 1 and the subsequent transaction requests. As shown in Figure 4 and Figure 5, when the transaction request is updated remotely, because the transaction execution program 16 of the computer 1 with the original main memory resident data is also attached with the transaction number, the omission of the transaction is caused by The transaction execution program 26 of the computer 2 is detected, and the subsequent reply processing, that is, the processing of the missing transaction ticket for subsequent transaction requests, and the transaction update occasion updated with the area described in FIG. 7 and FIG. 8 the same. According to the above method, through the area update of the resident data of the main memory and the remote update, the main memory data management method in which the location of the main memory is hidden by the application keeps the original data and the copied data at the same content. In the structure, when a transaction is lost due to network failure or damage, etc., since the transaction can be re-executed from the lost transaction, the content of the original data and the copied data can be kept consistent. Third Embodiment A method for initial setting of resident data of the main memory according to the present invention will be described with reference to Figs. 9 and 10. Moreover, FIG. 9 and FIG. 10 are interconnected. In the two figures, the initial setting method 307 of the resident data of the main memory consists of the initial setting programs Π6, 216, the version setting processing of the transaction execution program 117, 217, and the initial setting of the transaction receiving program. It is composed of transfer processing 118 and 218 for setting transaction requirements. -------------------- ^ ---- 1 ---- (Please read the notes on the back before filling out this page) __ This paper floods K degrees around Kawasaki valve valve standard (CNSM 1 UM0, go to the factory> > ¥) 560143 8996pif.doc / 008 Λ7 B7 The 5th Ministry of Economic Affairs, Intellectual Property Bureau, Employees' Cooperatives, India, Inventory (θ) Computer 1 and computer 2 is to perform remote update and area update as described in the first embodiment. When the computer 2 is started, the initial program setting program 216 of the computer 2 determines whether to set the initial program of the main memory resident data by any computer among other computers constituting the distributed computer system by referring to the system management data 211. It is used to set the initial transaction request of the resident data 21 in the main memory, and enter the transaction transmission queue 27 of the computer 2. Here, by referring to the system management data 211 and the main memory resident data of the computer 1, the initial data of the main memory resident data of the computer 2 is set. The transaction sending program 28 of the computer 2 uses the transaction sending queue 27 to take out the transaction request entered in the initial setting program 216 of the computer 2 to set the initial request, and sends it to the transaction receiving program 110 of the computer 1. The transaction receiving program 110 of the computer 1 is obtained from the transaction receiving queue 19 of the computer 1 and uses it to set the initial request of the main memory resident data 21 of the computer 2 to the transaction request, and sends it to the initial request setting program 116 of the computer 1. • The initial program setting program 116 of the computer 1 is based on its own regional resident data 11 to form an initial transaction for setting the main memory resident data 11 and enters the computer 1 transaction sending queue 17. Here, by setting the transaction request of the initial master resident data I1 formed by the initial master setting program 116, the version number of the master memory resident data 11 is also consistent. The version of the main memory resident data 'is set by the version setting process 117 of the main memory resident data by a transaction execution program constituting the initial setting means. (Please read the precautions on the back before filling out this page) -------- ^ --------- This paper ik 丨 丨 丨 屮 VALVE 孓 丨 孓 (cNS) A1 Cong UM ( ) Public ¥} 560143
經濟部智慧財產局員工消費合作社印K 8996pif.doc/008 五、發明說明（W ) 父易要求係如同發明的第一實施例所示的附上連續 的編5虎’此編5虎即爲主記億體常駐資料的版本編號。 亦即是如同第5圖所舉的實施例，電腦i、電腦2以 及電腦η的主記億體常駐資料，藉由處理個別的交易要求 而變化爲版本21、版本22、版本23。 電腦1的交易發送程式18係由電腦1的交易發送佇 列17 ’取出用以設定電腦2之主記憶體常駐資料21之初 始値的交易要求，並輸入電腦2的交易接收佇列29。 '電腦2的交易接收程式210係由電腦2的交易接收 疗列29 ’取出用以設定電腦2之主記憶體常駐資料21之 初始値的交易要求，並傳送至電腦2的初始値設定程式 216 ° 電腦2的初始値設定程式216係將設定電腦1的初 始値設定程式116所形成之主記憶體常駐資料11初始値 的交易要求，藉由本身的區域主記憶體常駐資料21進行 處理，以設定初始値。 此時於電腦2的主記憶體常駐資料21，在電腦1的 初始値設定程式116形成用以設定主記憶體常駐資料η 之初始値的交易要求之時點，設定電腦1之主記憶體常駐 資料11的版本。 電腦1在電腦2設定主記憶體常駐資料21的期間’ 亦處理接續而來的交易要求。因此，電腦2藉由電腦1的 交易要求以結束主記憶體常駐資料21的初始値設定後所 接收的交易要求，亦可能較設定主記憶體常駐資料21之 (請先閲讀背面之注意事項再填寫本頁)Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs, Consumer Consumption Cooperative K 8996pif.doc / 008 V. Description of the Invention (W) The parent's request is to attach a continuous series of 5 tigers as shown in the first embodiment of the invention. This series of 5 tigers is The version number of the resident data of Yiyi is recorded. That is, as in the embodiment shown in FIG. 5, the permanent data of the main body of computer i, computer 2 and computer n are changed to version 21, version 22, and version 23 by processing individual transaction requests. The transaction sending program 18 of the computer 1 is sent from the transaction sending queue 17 of the computer 1 'to take out the initial transaction request for setting the resident data 21 of the main memory of the computer 2 and enter the transaction receiving queue 29 of the computer 2. 'The transaction receiving program 210 of the computer 2 is the transaction receiving treatment line 29 of the computer 2.' Take out the transaction request for setting the initial data of the resident data 21 of the main memory of the computer 2 and send it to the initial program setting program 216 of the computer 2 ° The initial program setting program 216 of computer 2 will set the main memory resident data 11 formed by the initial program setting computer 116 of computer 1 to the transaction request of the initial initial program 11 of the main memory. Set the initial frame. At this time, at the time point of the main memory resident data 21 of the computer 2 and the initial request setting program 116 of the computer 1 to form the transaction request for setting the initial value of the main memory resident data η, the main memory resident data of the computer 1 is set. 11 version. The computer 1 sets the period of the main memory resident data 21 on the computer 2 and also processes subsequent transaction requests. Therefore, computer 2 uses the transaction request of computer 1 to end the initial setting of the resident data of main memory 21. The transaction request received after setting the resident data of main memory 21 may also be lower than that of resident data 21 of main memory (please read the precautions on the back before (Fill in this page)
_ n n ϋ n n I n 一^J· ϋ ϋ ·1 ϋ ϋ H ϋ I I 本紙依Κ度適川中阀阀家代(CJNS)A丨炚格（」U) χ」1)7公¥ ) 560143 8996pii'doc/008 i / — ___I£__ 五、發明說明（：)1 ) 初始値時的版本編號更加的新穎。 然而，此場合可藉由本發明第2實施例所述之暫時 性故障的交易回復處理，而能夠保證交易的連續性。 第11圖所繪示爲電腦1、電腦2以及電腦η的主記 憶體常駐資料的版本更新示意圖。 電腦1的主記憶體常駐資料11藉由處理交易要求而 形成版本20，由電腦Γ對電腦2以及電腦η發送附有交易 編號20的交易要求。 電腦2以及電腦η藉由接收、實行此交易要求，使 個別的主記憶體常駐資料的版本爲20。 在電腦2停止的場合，其後之電腦丨的版本更新至 21、22，伴隨著此主記憶體常駐資料的版本更新者，僅有 電腦η的主記憶體常駐資料。電腦n的主記憶體常駐資料 更新’係伴隨著電腦1的主記憶體常駐資料更新而成爲版 本 21 、 22 ° •電腦1的主記憶體常駐資料成爲版本22之後，電腦 2啓動的話，電腦2係對電腦1發送用以設定主記憶體常 駐資料Π之初始値的交易要求。 經濟部智慧財產局員工消費合作社印€ (請先閱讀背面之注意事項再填寫本頁) 藉由接收、實行設定由電腦1之主記憶體常駐資料11 之初始値的交易要求，電腦2的主記憶體常駐資料21的 初始値成爲設定的版本22。 其後，電腦1的主記憶體常駐資料更新爲版本23， 電腦2以及電腦η之主記憶體常駐資料的版本亦個別隨之 更新爲2 3。 24 本紙队Κ度⑨川t W W 'i:丨H (<JNS)A 1从格Cl〇 A ) 560143 Λ7 B7 8996pif.doc,/〇08 五、發明說明（〕？） 依照以上的方法，即使在無磁碟的電腦中，亦能夠 設定主記憶體常駐資料的初始値。由於能夠將其他電腦的 主記憶體常駐資料，藉由線上更新以形成本身的主記憶體 常駐資料，而能夠將電腦啓動時點的最新値設定爲主記憶 體常駐資料的初始値。 第四實施例 依照第12圖以及第13圖說明本發明的交易保持方 法。尙且第12圖與第13圖係互相連接。 於兩圖中，交易保持手段係由交易接收程式一部份 的初始値設定交易要求的交接處理118與交易保持佇列 119所構成。 於兩圖中，電腦2的初始値設定程式216係接受用 以設定電腦1的初始値設定程式116所形成之主記憶體常 駐資料11初始値的交易要求，經由電腦1的交易發送程 式_ 18以及電腦2的交易接收程式210接受、處理中的物 件於此處未圖示。 電腦2的交易接收程式210係作爲交易保持手段的 一部份，在本身的區域主記憶體常駐資料之初始値設定 中，混入用以設定初始値的交易要求之所接收交易要求， 不輸入電腦2的交易實行佇列24，而是進入電腦2的交易 保持佇列219。 在設定主記憶體常駐資料之初始値的交易要求全部 結束後，則電腦2的交易接收程式210由電腦2的交易保 (請先閲讀背面之注意事項再填寫本頁) --------^--------- 經濟部智慧財產局員工消费合作社印於 本纸過用 家d- (CNS)Al 从格上丨:· :、、V ) A7 560143 8996pif.doc/008 五、發明說明（力） 持佇列219中取出所保持的交易要求，藉由輸入電腦2的 交易實行佇列24，如同以往的以交易實行程式26處理交 易要求。 第14圖所繪示爲處理設定主記憶體常駐資料之初始 値期間所接收的交易要求的狀態。 電腦1的主記憶體常駐資料藉由處理交易要求而成 爲版本20，由電腦1對電腦2以及電腦η發送附上交易編 號20的交易要求。 電腦2以及電腦η接收、實行此交易要求，使個別 的版本爲20。 在電腦2停止的場合，其以後之電腦1的版本更新 至21，伴隨著此主記憶體常駐資料的版本更新者，僅有電 腦η的主記憶體常駐資料。電腦η的主記憶體常駐資料更 新，係伴隨著電腦1的主記憶體常駐資料更新而成爲版本 21° *電腦1的主記憶體常駐資料成爲版本21之後，啓動 電腦2的話，電腦2對電腦1發送用以設定主記憶體常駐 資料之初始値的交易要求。 藉由接收、實行來自電腦1的設定主記憶體常駐資 料之初始値的交易要求，電腦2的主記憶體常駐資料的初 始値設定爲版本21。 電腦2在實行由電腦1接收的設定主記億體常駐資 料21之初始値的交易要求期間，由電腦1接收附交易編 號22的交易要求與附交易編號23的交易要求的話’則將 26 本纸、乂墁⑨川屮内 w（CNS)A1 1 ) ____________--------^ (請先閱讀背面之注意事項再填寫本頁)_ nn ϋ nn I n ^ J · ϋ ϋ · 1 ϋ ϋ H ϋ II The paper is suitable for the Chuanzhong valve valve family generation (CJNS) A 丨 炚 grid ("U) χ" 1) 7 public ¥) 560143 8996pii 'doc / 008 i / — ___ I £ __ 5. Description of the invention (:) 1) The version number at the time of initial is more novel. However, in this case, the continuity of transactions can be guaranteed by the transaction failure processing of the temporary failure described in the second embodiment of the present invention. Figure 11 shows a schematic diagram of the version update of the resident data of the computer 1, computer 2 and computer n. The resident data 11 of the main memory of the computer 1 forms a version 20 by processing the transaction request, and the computer Γ sends a transaction request with a transaction number 20 to the computer 2 and the computer η. The computer 2 and the computer η receive and execute the transaction request, so that the version of the individual main memory resident data is 20. When the computer 2 stops, the version of the subsequent computer 丨 is updated to 21, 22, and with the version update of the resident data of the main memory, only the resident data of the main memory of the computer η is updated. Update of the main memory resident data of computer n 'is accompanied by the update of the main memory resident data of computer 1 to become version 21, 22 ° • After the main memory resident data of computer 1 becomes version 22, computer 2 starts, and computer 2 It is a transaction request sent to the computer 1 to set the initial volume of the resident data of the main memory. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs (please read the precautions on the back before filling this page) By receiving and implementing the initial transaction request set by the resident data 11 of the main memory of computer 1, the owner of computer 2 The initial version of the memory resident data 21 becomes the set version 22. Thereafter, the resident data of the main memory of computer 1 was updated to version 23, and the resident data of the main memory of computer 2 and computer η were also updated to 2 3 individually. 24 Paper team κ 度 ⑨ 川 t WW 'i: 丨 H (< JNS) A 1 Congge ClOA) 560143 Λ7 B7 8996pif.doc, / 〇08 5. Description of the invention (]?) According to the above method, Even in a diskless computer, you can set the initial size of the main memory resident data. Because the main memory resident data of other computers can be updated online to form its own main memory resident data, the latest time at which the computer is started can be set as the initial data of the main memory resident data. Fourth Embodiment The transaction holding method of the present invention will be described with reference to Figs. 12 and 13.尙 And Figure 12 and Figure 13 are interconnected. In the two figures, the transaction holding means is composed of the initial processing setting processing request 118 and the transaction holding queue 119 which are part of the transaction receiving program. In the two figures, the initial program setting program 216 of the computer 2 accepts the transaction request of the initial memory 11 of the main memory formed by the initial program setting program 116 of the computer 1, and sends the program via the transaction program of the computer 1_ 18 The objects accepted and processed by the transaction receiving program 210 of the computer 2 are not shown here. The transaction receiving program 210 of the computer 2 is a part of the transaction holding means. In the initial setting of the resident data of the main memory of its own area, the received transaction request for setting the initial transaction request is mixed, and it is not entered in the computer. The transaction of 2 is implemented in queue 24, and the transaction of computer 2 is maintained in queue 219. After all the initial transaction requirements of the main memory resident data are set, the computer 2's transaction receiving program 210 is protected by the computer 2's transaction (please read the precautions on the back before filling this page) ------ -^ --------- Printed on the paper by a consumer co-operative of the Intellectual Property Bureau of the Ministry of Economic Affairs d- (CNS) Al from the grid 丨: · ::, V) A7 560143 8996pif.doc / 008 V. Description of the invention (force) The holding queue 219 takes out the held transaction request, and enters the transaction execution queue 24 of the computer 2 as in the conventional transaction execution program 26 to process the transaction request. Figure 14 shows the status of processing the transaction request received during the initial period of setting the main memory resident data. The resident data of the main memory of the computer 1 becomes version 20 by processing the transaction request, and the computer 1 sends a transaction request with a transaction number 20 to the computer 2 and the computer η. Computer 2 and computer η receive and execute this transaction request, so that the individual version is 20. When the computer 2 is stopped, the version of the subsequent computer 1 is updated to 21, and with the version update of the main memory resident data, only the main memory resident data of the computer η is updated. The main memory resident data of computer η is updated to become version 21 ° as the main memory resident data of computer 1 is updated. * After the main memory resident data of computer 1 becomes version 21, computer 2 is started and computer 2 is connected to the computer. 1 Send a transaction request to set the initial size of the main memory resident data. By receiving and executing the transaction request of the initial data of the main memory resident data from the computer 1, the initial data of the main memory resident data of the computer 2 is set to version 21. While computer 2 is executing the transaction request for initial registration of the set master resident data 21 received by computer 1, while computer 1 receives the transaction request with transaction number 22 and the transaction request with transaction number 23, 26 copies Paper, 乂 墁 ⑨ 川 屮 内 w (CNS) A1 1) ____________-------- ^ (Please read the notes on the back before filling in this page)
經濟部智慧財產局員工消費合作社印K 560143 Λ7 B7 8996pif.doc/008 五、發明說明（/ ) 電腦1的交易發送程式18發送至電腦2的交易接收程式 29 〇 此由電腦1的交易發送程式18發送至電腦2的交易 接收程式29之交易要求，亦可能因爲網路的暫時性故障 損害而遺漏的場合。 於交易要求附有表示交易實行順序的交易編號，與 前述第2實施例相同，·交易要求的遺失可藉由編號的遺漏 檢測出。本實施例由電腦1的交易發送程式18發生交易 遺漏，則由電腦2的交易實行程式26檢測出。 電腦2的交易實行程式26 —旦檢測出交易要求的遺 漏，則將此意旨經由電腦2的交易送信程式28以及電腦1 的交易接收程式110通知電腦1的交易實行程式16。 電腦1的交易實行程式16由電腦2的交易實行程式 26接獲交易要求遺漏的通知，則由電腦1的交易結束佇列 Π5中取出遺漏的交易要求以後的交易要求，經由電腦1 的交易發送程式18以及電腦2的交易接收程式210，傳送 至電腦2的交易實行程式26。 此時由於網路的暫時性故障損害時間過長，由於爲 了節制電腦資源，而在具有大小上限的交易結束佇列中產 生覆寫等的原因，使得結束交易要求變成不存在。 當檢測出在交易結束佇列115中已沒有必須再傳送的 交易要求，則通知電腦1的初始値設定程式116形成用以 設定初始値的交易。 電腦1的初始値設定程式116接收由電腦1的交易實 28 (請先閱讀背面之注意事項再填寫本頁) 訂---------線·· 經濟部智慧財產局員Η消费合作社印於 560143 Λ7 B7 8996pif.doc/008 五、發明說明) 行程式16所傳出的通知，基於本身的區域主記億體常駐 資料11，形成用以設定主記憶體常駐資料之初始値的交易 要求，經由電腦1的交易發送程式18以及電腦2的交易 接收程式210，傳送至電腦2的交易實行程式26。 電腦2的交易實行程式26，藉由實行處理由電腦1 的初始値設定程式116所形成之用以設定主記憶體常駐資 料11之初始値的交易要求·，以形成本身的區域主記憶體 常駐資料21。 依照以上的方法，當發生暫時性的網路故障損害等 、而導致交易要求發生遺失的場合，即使是再實行回溯過去 的交易資料以修補交易要求遺漏的方法所無法回復之經過 長時間的交易遺漏，亦能夠提供回復的手段。 第六實施例 第丨7圖以及第18圖所繪示爲以具有主記憶體常駐 資料的分散電腦系統構成重複系統的方法。 原始的主記憶體常駐資料的所在位置，亦即是切換 更新原始的主記憶體常駐資料之電腦的更新權管理手段 310 ’係由具更新權的管理程式3與系統管理程式4所構 成。 第17圖爲由應用程式的交易要求更新版本，持有原 ^的主§5憶體常駐資料之電腦1停止之前的狀態示意圖， ^ 18&圖所繪示爲持有主記憶體常駐資料之電腦1停止， 斤的^有原始的主記憶體常駐資料之電腦2建立之後狀態 n n n H ϋ ϋ I n n I ϋ n · 1.· n 1· ϋ n l ϋ 一 运-· ϋ ·1 ϋ n aM§ i_i I I (請先閱讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印於 29 560143 8996pif.doc/〇〇8 經濟部智慧財產局員工消費合作社印纪 1^7 五、發明說明（β) 的示意圖。 一在電腦即將停止的狀態中，持有藉由應用程式傳出 的父易要求更新之主記憶體常駐資料的原始値者，係爲電 腦1 °電腦1的主記憶體主記憶體常駐資料11，不跟隨應 用程式112的交易要求，版本編號爲23。 由電腦1往電腦2以及電腦η的交易要求傳送，由 於網路延遲等理由，尙.未接.收到到版本23爲止的交易要 求，電腦2以及電腦η的版本編號分別停留在21以及18。 而且，藉由應用程式傳出的交易要求，表示爲最初 更新之原始的主記憶體常駐資料之所在位置的系統管理資 料之原始値者係爲電腦m。·電腦㈤以外的電腦則爲系統管 理資料的複製。此系統管理資料的更新係由電腦m的系統 管理程式進行。 此狀態電腦1停止的話，檢測出電腦丨停止之電腦 m的系統管理程式3，使用更新權限以更新本身的區域系 統·管理資料’藉由應用程式傳出的交易要求進行最初的更 新，以電腦2作爲具有原始的主記憶體常駐資料的電腦。 以電腦m的系統管理程式，主記憶體常駐資料以及 系統管理資料之原始値更新的話，此更新內容作爲交易要 求由電腦m發送至電腦2以及電腦n ° 尙且，電腦m的系統管理程式3，係以定週期輪詢 (Polling)等的方法’而能夠檢測出電腦1的停止。 電腦2對其後電腦11所發送的交易要求所附編號爲 22，電腦ϋ檢測出此父易要求並不連纟買。檢測出交易不連 30 --------------------訂---------線· (請先閱讀背面之注意事項再填寫本頁) 560143 A7 H7 8996pif.doc/008 五、發明說明（4) 續時的交易回復方法則如同前述第二實施例或是第五實施 例所述的方法。 依此發明，即使具主記憶體常駐資料的電腦網路系 統中，持有原始的主記憶體常駐資料之電腦發生停止，在 此狀態下，由於其他電腦能夠持有新的原始資料，因而能 夠構成重複系統。 第七實施例 前述實施例所說明的發明，能夠藉由儲存在記憶媒 體內的程式讀入電腦內以實現。 發明的效果 依本發明的話，在分散電腦系統中的電腦啓動時， 並不依存磁碟的備份資料，在提供一種設定主記億體初始 値之方法的同時，並提供於分散電腦系統中，由應用程式 隱廠主記憶體常駐資料是否存在於何處的電腦’並於應用 程式，相對於應用程式，並能夠發揮主記憶體之可能限度 的高速存取性。 而且，主記憶體常駐資料所構築的系統能夠構成重 複系統而能夠建構具有高可靠度的系統。 --------------------t· — ϋ ϋ n ft— ϋ Mt I 線-0- (請先閲讀背面之注意事項再填寫本頁> 經濟部智慧財產局員工消費合作社印於 本纸讯K 淘川t阀W孓代？- (〔 NS)A 1从格（丨卜V )Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs K 560143 Λ7 B7 8996pif.doc / 008 V. Description of the Invention (/) The transaction sending program 18 from computer 1 to the transaction receiving program 29 from computer 2 〇 This is the transaction sending program from computer 1 18 The transaction request sent by the transaction receiving program 29 to the computer 2 may also be missed due to temporary network damage. The transaction request is accompanied by a transaction number indicating the order in which the transaction was executed, which is the same as the second embodiment described above. The loss of the transaction request can be detected by missing the number. In this embodiment, if a transaction omission occurs in the transaction sending program 18 of the computer 1, it is detected by the transaction execution program 26 of the computer 2. Computer 2's transaction execution program 26—Once the omission of a transaction request is detected, the intention is to notify computer 1's transaction execution program 16 via computer 2's transaction transmission program 28 and computer 1's transaction receiving program 110. The transaction execution program 16 of computer 1 is notified by the transaction execution program 26 of computer 2 of the omission of the transaction request, and the missing transaction request is retrieved from the transaction end queue 5 of computer 1 and the subsequent transaction requests are sent through the transaction of computer 1. The program 18 and the transaction receiving program 210 of the computer 2 are transmitted to the transaction execution program 26 of the computer 2. At this time, due to the long-term damage caused by the temporary failure of the network, and in order to control computer resources, overwrites were generated in the transaction end queue with an upper limit, which made the transaction termination request non-existent. When it is detected that there is no transaction request that must be retransmitted in the transaction end queue 115, the initial program setting program 116 of the computer 1 is notified to form a transaction for setting the initial program. Computer 1's initial setting program 116 receives transaction data 28 from Computer 1 (please read the precautions on the back before filling out this page) Order --------- Line ·· Intellectual Property Bureau, Ministry of Economic Affairs, Consumer Cooperative Printed on 560143 Λ7 B7 8996pif.doc / 008 V. Description of the invention) The notification issued by program 16 is based on its own regional memory of the resident data of the billion body 11 to form an initial transaction for setting the resident data of the main memory. The request is transmitted to the transaction execution program 26 of the computer 2 via the transaction sending program 18 of the computer 1 and the transaction receiving program 210 of the computer 2. The transaction execution program 26 of the computer 2 is executed by processing the initial request of the main memory resident data 11 formed by the initial program setting program 116 of the computer 1 to form its own area. The main memory is resident Information 21. In accordance with the above method, when temporary network failure damage occurs and the transaction request is lost, even if the method of retrospecting the past transaction data to repair the missing transaction request cannot be restored, the transaction has passed for a long time. Omissions can also provide a means of reply. Sixth Embodiment Figures 7 and 18 show a method of constructing a repeating system using a distributed computer system with resident data in main memory. The location of the original main memory resident data, that is, switching the computer to update the original main memory resident data, means for managing the update right 310 ′ is composed of a management program 3 and a system management program 4 with update rights. Fig. 17 is an updated version of the application's transaction request. The state before the computer 1 holding the original main §5 memory resident data is stopped. ^ 18 & The picture shows the main memory resident data. Computer 1 stops, and the state of computer 2 with the original main memory resident data is established nnn H ϋ ϋ I nn I ϋ n · 1. · n 1 · ϋ nl ϋ 运-· ϋ · 1 ϋ n aM § i_i II (Please read the notes on the back before filling this page) Printed on 29 560143 8996pif.doc / 〇〇8 Consumer Consumption Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs BRIEF DESCRIPTION OF THE INVENTION (β). First, in the state where the computer is about to stop, the original person who holds the resident data of the main memory that is updated by the parent application that is requested by the application is the main memory of the computer 1 and the main memory resident data 11 , Does not follow the transaction request of application 112, the version number is 23. The transaction request from computer 1 to computer 2 and computer η was transmitted. Due to network delay and other reasons, 尙. Missed. After receiving the transaction request up to version 23, the version numbers of computer 2 and computer η stay at 21 and 18, respectively. . In addition, the transaction request from the application program indicates that the original source of the system management information where the original main memory resident data was originally updated is the computer m. • Computers other than PCs are copies of system management data. This system management data is updated by the system management program of the computer m. In this state, if the computer 1 is stopped, the system management program 3 of the computer 丨 stopped computer m is detected, and the update authority is used to update its own regional system and management data. The transaction request from the application is used to perform the initial update. 2 As a computer with raw main memory resident data. If the system management program of the computer m, the resident data of the main memory and the original management system data are updated, the update content is sent as a transaction request from the computer m to the computer 2 and the computer n °, and the system management program of the computer m 3 It is possible to detect the stop of the computer 1 by a method such as polling at a fixed period. Computer 2 attached the transaction request sent by computer 11 to serial number 22, and the computer did not detect this request and did not buy it. Detected the transaction is not connected to 30 -------------------- Order --------- line · (Please read the precautions on the back before filling this page ) 560143 A7 H7 8996pif.doc / 008 V. Description of the Invention (4) The method of replying to the transaction is the same as the method described in the second or fifth embodiment. According to this invention, even in a computer network system with resident data of the main memory, the computer holding the original resident data of the main memory is stopped. In this state, since other computers can hold the new original data, it can Make up a repeating system. Seventh Embodiment The invention described in the foregoing embodiment can be implemented by reading a program stored in a memory medium into a computer. Effects of the Invention According to the present invention, when the computer in the distributed computer system is started, it does not depend on the backup data of the magnetic disk. While providing a method for setting the initial size of the master billion, it is also provided in the distributed computer system. The application hides whether the computer's resident data resides in the computer or not. In the application, compared with the application, it can take advantage of the high-speed accessibility of the main memory. In addition, the system constructed by the resident data of the main memory can constitute a duplicate system and can construct a highly reliable system. -------------------- t · — ϋ ϋ n ft— ϋ Mt I line -0- (Please read the precautions on the back before filling this page > Economy Printed by the Ministry of Intellectual Property Bureau Consumer Cooperatives on this paper Newsletter K Taochuan t valve W ?? ((NS) A 1 Congge (丨 卜 V)