TWI223167B - Distributed virtual web cache implemented entirely in software - Google Patents

Distributed virtual web cache implemented entirely in software Download PDF

Info

Publication number
TWI223167B
TWI223167B TW089112368A TW89112368A TWI223167B TW I223167 B TWI223167 B TW I223167B TW 089112368 A TW089112368 A TW 089112368A TW 89112368 A TW89112368 A TW 89112368A TW I223167 B TWI223167 B TW I223167B
Authority
TW
Taiwan
Prior art keywords
node
storage
nodes
monitoring
data element
Prior art date
Application number
TW089112368A
Other languages
Chinese (zh)
Inventor
Scott H Davis
Daniel J Dietterich
Paul D Asselin
Gerard R Plouffe
Philip John Romine
Original Assignee
Mangosoft 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 Mangosoft Corp filed Critical Mangosoft Corp
Application granted granted Critical
Publication of TWI223167B publication Critical patent/TWI223167B/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/957Browsing optimisation, e.g. caching or content distillation
    • G06F16/9574Browsing optimisation, e.g. caching or content distillation of access to content, e.g. by caching

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

A method and system are provided for optimizing the local caching of one or more data components available from a server node. Each of plural nodes connected to a local area network is provided with a locally physically present cache. Each of the caches of at least some of the plurality of nodes are linked together into a single virtual cache. A particular one of the nodes is designated as a repository node for persistently storing a particular data component and for providing a copy of the particular data component to other referencing nodes of the plurality of nodes which lack, but which desire to access, the particular data component. Designation of the particular node as the repository node is unchanged solely by providing a copy of the particular data component to one of the referencing nodes that desires to access the particular data component. The repository function of a node for a particular data component is relatively static. For instance, assume that a first referencing node obtains from the repository node a copy of the particular data component and performs a group of one or more access operations on the copy of the data component If another referencing node lacks a copy of, but desires to access, the particular data component, a copy of the particular data component is provided from the repository node to the other referencing node. One of the nodes may be designated as a monitor node for the particular data component. The monitor node responds to requests by each referencing node to identify the repository node of the particular data component by providing the identity of the particular node which serves as the repository node for the particular data component.

Description

1223167 修正 國的臨時申請案序號 「完全以軟體實施的分散式虛擬 ’該案之申請曰為西元一九九九 申清案之内容在此併入作為參 美國專利案第6, 〇26, 474是有關 使用全域可定址之記憶體以分 」’該案之申請曰為西元一九九 述專利案之權利,已由本申請案 之内容與本申請案之主題内容是 入作為參考。 1 號 89112368 五、發明說明(1) 相關申請案: 本申凊案係依據在美 60/140,645、發明名稱為 網站快取記憶體」而申請 年六月二十三日。該臨時 考。 除此之外,本發明與 連的,該案之發明名稱為 旱用戶端網站快取記憶體 六年十一月二十二日。上 申请人取得。前述專利案 相關連、在此且可完全併 '【發明所屬之技術領域】 本發明是關於電腦系統存取網路節點(n〇des 〇f ΐ資T元統5方法者’特別指在電腦系統存取網際網 ^ 4 件$,精由其他節點已擷取之起始資料來加速電 細系統擷取網路節點資料元件之速度者。 、 【先前技術】 來說’"快取記憶體(cache) ”是在電觸系統中一 二有比主圮憶體較快速度的記憶體。 一 經常被用來指具有較高速度之記憶體電路、如/態體隨機^ 2 =憶體(SRAM) ’其被使用於較慢速度記憶體電路、例如 動㈣㈣|記憶體(DRAM)之補助記憶體。然而,"快取 ^23167 修正 曰 五 案號8910368___年 發明說明(2) 一 用己,^也常被用來指任何形式的較高速記憶體,且被# i體區域之硬碟的一部❾,若?“統 補助記憶胃、並透過低速通信鏈而;伺服器的 硬碟亦被稱為"快取記憶體"。在此於 後者形式的快取記憶體。 + ^月係有關於 令鱼决5記憶體之操作原理如下:一電腦系統在作程式# local貝it枓之f存取f而顯示其時間性基準現場(ΐειηρ〇Ι^1工S 4 ® . Υ ° ^ erence)之特性;亦即,電腦系 常:ίΞ;】;=指令。快取記憶體藉由儲存某= ^ „:;;; ; ^Fa1 ,fl ^ ^ 因此會具有較小的儲存空’也 ί d、體電路作為該電腦系統之快取記憶:’,、這::ϊ ,、成本疋由對它們存取之電腦系統之數目^日γ 分攤,此種情況是更加實際二在V 汁上,、所以此達成存取時間整體之、-, 系統能夠在高速快取記憶體中之複製;份乃;y電腦 :式指令與資料存取之故,而在較 =二:之 份、僅作少次數存取。(請注意:程式指上體中原始備 份資料存取,其係顯示空間性基準現場^ 〇Callty 〇f referenCe)之特性。以上牿 M:怂 I·· 記憶體位£^之取記憶1223167 Provisional application serial number of the amending country "Decentralized virtual implementation entirely in software" The application of this case is called the 1999 1999 settlement. The content of this application is incorporated herein as a reference to US Patent No. 6, 〇26, 474 It is related to the use of the globally addressable memory to divide the "" application in this case is the right of the 1999 patent mentioned in the case. No. 1 89112368 V. Description of the invention (1) Relevant applications: This application was filed on June 23 of the year based on the US 60 / 140,645, the name of the invention is the website cache memory. The provisional exam. In addition, the present invention is related to the name of the invention in this case, which is the cache memory of the client website. Obtained by the applicant. The aforementioned patents are related, and can be fully incorporated here. [Technical field to which the invention belongs] The present invention relates to computer systems accessing network nodes The system accesses the Internet ^ 4 pieces, which uses the starting data retrieved by other nodes to accelerate the speed of the electronic system to retrieve the data elements of the network node. [Previous technology] In terms of '" cache memory "Cache" is a memory that has a faster speed than the main memory in the electric contact system. One is often used to refer to a memory circuit with a higher speed, such as / state random ^ 2 = memory (SRAM) 'It is used in slower speed memory circuits, such as auxiliary memory of dynamic memory | DRAM (DRAM). However, " cache ^ 23167 amended the fifth case number 8910368___ invention description (2 ) Once used, ^ is often used to refer to any form of higher-speed memory, and it is used by a hard disk in the #i body area, if it? "All subsidize the memory and pass the low-speed communication chain; servo The hard drive of the device is also called " cache ". Hereafter + ^ Month is related to the operating principle of the memory of the 5th memory: a computer system in the program # local 贝 it 枓 access f to display its temporal reference scene (ΐειηρ 〇Ι ^ 1 工 S 4 ®. Υ ° ^ erence); that is, the computer system is often: Ξ;]; = instruction. Cache memory by storing some = ^ „: ;;;; ^ Fa1, fl ^ ^ Therefore, it will have a small amount of storage space 'also d, the body circuit as the cache memory of the computer system:' ,, this :: ϊ, cost, the number of computer systems accessed by them ^ days γ sharing, this situation is more practical in V juice, so this achieves the overall access time,-, the system can be copied in high-speed cache memory; share is; y computer: type instructions and data storage For the sake of comparison, only a small number of accesses are required. (Please note: The program refers to the access to the original backup data in the upper body, which is a display of the spatial reference site ^ 〇Callty 〇f referenCe) Characteristics. Above 牿 M: Advice I ·· Memory position £ ^

第8頁 I223l67 曰 案號 89112368 五、發明說明(3) 體亦提供空間性基準現場之存取效率 性持有較低興趣。 一在本案中對此特 在電腦系統與網路上廣泛使用快 許多目的。許多一般性盥特定性夕也%。己^體技術以達成 可同時滿足一般與特定之需求。 匕體知決方式, 快取記憶體的一種特定使用是在對,,全 jWorld Wide Web ) ”或”網路(Web ) ”上存取網°百罔 容。、為了存取網頁,一電腦系統通常執行網覽^ = 網路觀是一種其功能至少能經由例如J = 之廣域網路對訊息封包發出指令的程式,如辨識 者感興趣之内容之網路伺服器,以及由此網路伺服=== 有興趣之内容。網路伺服器之辨識,乃是使用π通 定址器(universal resource 1〇cat〇r)” 位址或即二"、 謂f’URLs”。通用資源定址器位址可被轉換成網路伺服 網際網路協定(IP )位址,再利用擷取内容的其他指^以 達成上述目的。所稱”内容”,其典型的是指”網頁,, 式;或是可由電腦上所執行之網路瀏覽器以視覺與聽^之 方式所顯示者。至於網頁,其典型者是由”超文件標示語 言(hypertext markup language 或 html)” 型式所提供 的、包含文件和袼式化命令,用以根據該電腦系統的能力 對網頁瀏覽器下指令來呈現視覺與聽覺的資訊。網頁也可 能包含崁入(embedded )式命令,以從網路伺服器擁取其 他内容資訊以供顯示。此等其他内容通常由單一的通用資 源定址器(URL )來作連結。於此,組成一可擷取内容資 料的一個單位的資料,稱之為一”小飾物(t r i nke t广,或Page 8 I223l67 Case No. 89112368 V. Description of the Invention (3) The body also provides spatial reference site access efficiency and has a low interest. In this case, it is widely used in computer systems and networks for many purposes. Many general toilets are also sexually specific. Self-contained technology to achieve can meet both general and specific needs. In a dagger-based manner, a specific use of cache memory is to access the Internet on the Internet, the full jWorld Wide Web, or the Web. In order to access web pages, a computer The system usually performs web browsing ^ = network view is a program whose function can at least send instructions to message packets via a wide area network such as J =, such as a web server that identifies content of interest to the viewer, and from this web server === Interested content. The identification of the web server is to use the π universal addresser (universal resource 10cator) address or two ", called f'URLs. Universal resource addresser The address can be converted into a network server Internet Protocol (IP) address, and then other means of extracting the content can be used to achieve the above purpose. The so-called "content" is typically referred to as "web page," Or it can be displayed visually and listeningly by a web browser running on a computer. As for the web page, it is typically provided by the "hypertext markup language (html)" type, which includes files and stylized commands for presenting instructions to the web browser according to the capabilities of the computer system Visual and auditory information. Web pages may also contain embedded commands to grab other content information from a web server for display. These other contents are usually linked by a single universal resource locator (URL). Here, the data constituting a unit capable of retrieving content data is called a "trinket" (or

1223167 年 月1223167 month

案號 8911236S 修正 五、發明說明(4) 資料元件。在此處,資料元件將被認定為網頁内容之最小 的、原子式可存取單元。然而,這僅是為了方便說明,對 於熟悉本發明技術領域之人士,應可了解本發明對於其他 種類的資料兀件、以及其他應用方面的適用性。 ^此種將複數個資料元件之混合體建構為網頁的方式有 許多優點,其中之一是:内容提供者(c〇ntent pr〇vide〇 jr在數個網頁中重復使用該等資料元件,以供應來自網站 :Ϊ!的存:L吾人可以瞭解,如此即呈現了-種利用網 站快取記憶體時的時間性基準現場之特性(pr〇perty 〇f temP〇ral locality 〇f reference)的機會。亦即,大 2商::頁瀏覽器套裝軟體在提供一快取記憶 ί用劉覽器應用軟體之電腦系統之硬碟内、該 示的各網頁之各資料元件( :鮮以供顯 各個被摘取網頁之顯示…並間㈣)。如此可加速 將此資料元件之複L 一次,而是-次 -次某-資料元件形成所存於快取記憶體中。每 諜中之坊次祖- ^ '、’罔頁之一部份時,即擷取硬 下載該資二元:m::中:不是從網路飼服器重新 存取時間總是比從網ς伺服==料元件的複製備份之 件所需時間要快速很多,如此加::際網路下載該資料元 每一資料元件,也彳提供一個來數=網頁即可達成。對 的複製備份之失效曰期。如此::指不該資料兀件 數去判定在快取記憶體内之資料;:劇覽器即可使用該參 -----的複製備份是否已經Case No. 8911236S Amendment 5. Description of the invention (4) Data element. Here, the data element will be identified as the smallest, atomically accessible unit of web content. However, this is only for the convenience of description. For those familiar with the technical field of the present invention, it should be understood that the present invention is applicable to other kinds of data elements and other applications. ^ This method of constructing a mixture of a plurality of data elements into a web page has many advantages. One of them is that a content provider (conntent pr0vide〇jr reuses these data elements in several web pages to Supply from the website: Ϊ! Deposit: I can understand that this presents an opportunity to use the temporal reference site characteristics (pr〇perty 〇f temP〇ral locality 〇f reference) when using the website cache memory . That is, the big 2 merchants :: The page browser package software is in the hard disk of the computer system that provides a cache memory application software. The data elements of the web pages (: fresh for display) The display of each extracted web page ... and time to time). This can speed up the re-l of this data element once, but the -time-time certain-data element formation is stored in the cache memory. Every time in the spy Ancestor-^ ',' part of the title page, it is hard to download the binary data: m :: medium: not re-access time from the network server is always longer than from the server It takes much faster to copy a component back up, so add: Download each data element of the data element on the Internet, and also provide a number = webpage. It can be reached on the date of expiration of the duplicate backup. So :: refers to the number of data elements to determine in the cache memory Contents :: The play browser can use this parameter ----- whether the copy backup has been

第10頁 1223167 “ 戏"六U兄口月、 失效,若失效、則不該用以顯示網頁。若網頁瀏覽器判定 在快取記憶體内之資料元件的複製備份已經失效,則快取 記憶體内之資料元件的複製備份應被丟棄,而由網際網路 下載一份新的資料元件。 在例如區域網路(LAN )之電腦系統網路中,許多電 腦系統經常執行網路瀏覽器。雖然這許多電腦系統能夠支 援個別直接連到網際網路,一般而言、各區域網路具有一 或夕個代理飼服裔(pr〇Xy SerVerS),用以達成在該區域 網路内所有電腦系統的外部資料通信。在各電腦系統上之 ^路剧覽器應用程式被安排為將其所有發給網際網路之訊 息’彳專送給一或多個代理伺服器。接著,代理伺服器將訊 心傳送至網際網路,並且將其由網際網路所接收之訊息分 配給在區域網路上適當之電腦系統。 具有網頁快取記憶體能力的一些商用代理伺服器及代 理伺服為應用軟體包括··微軟代理伺服器(Microsoft Pyoxy ServerTM ),其係由設立於華盛頓地區州雷蒙市之 ,軟a司販售,以及新伺服器(Ne〇serverTM ),其係由 設^於德州休士頓市之康百客(Compaq)公司販售的。這 服器本身保有為操取資料元件的快取記憶體。如 理祠月I ΐ ί路内一特定電腦系統企圖操取資料元件而此代 合將快取j ΐ儲存(未失效)複製備份,則此代理伺服器 ▲是,己憶體的該複製備份傳送至該特定電腦系統。結 U 網站飼服器經由網際網路下載資料元件,電 m ..... σ以車父尚的速度擷取該資料元件。 采用-種稱為快取記憶體陣列隸協定 第11頁 1223167Page 1223167 "Play" six months, invalid, if it is invalid, it should not be used to display the web page. If the web browser determines that the copy backup of the data elements in the cache memory has expired, the cache Duplicate backups of data elements in memory should be discarded and a new data element downloaded from the Internet. In computer system networks such as a local area network (LAN), many computer systems often run web browsers Although these many computer systems can support individual direct connections to the Internet, in general, each local area network has one or more agents (prOxy SerVerS) to achieve all External data communication of computer systems. The ^ Router Browser application on each computer system is arranged to send all of its messages to the Internet '彳 exclusively to one or more proxy servers. Then, the proxy server The router sends the message to the Internet, and distributes the information it receives from the Internet to the appropriate computer system on the local network. Some businesses with web cache memory capabilities Proxy server and proxy server are application software including Microsoft Proxy Server (Microsoft Pyoxy ServerTM), which is established in Raymond, Washington, and sold by the Software Division A, as well as a new server (NeOserverTM). , Which is sold by Compaq Company, Houston, Texas. The server itself holds cache memory for accessing data components. A specific computer system attempts to access data components and this generation will cache j ΐ store (not expired) a duplicate backup, then this proxy server ▲ Yes, the duplicate backup of the self-memory body is transferred to that particular computer system. End U Website The feeder downloads the data element through the Internet, and the electric m ..... σ retrieves the data element at the speed of the car driver. Adopt-a kind of cache memory array slave agreement, page 11 1223167

Reached array routing pr〇t〇c〇1 ;CARp)的技術。 :=記憶體陣列通路協定(CARp),當電腦系統二 。':要擷取由通用貢源定址器(URL)所辨識之特定資 卡=URL被雜湊(hashed)起來,依據此雜湊函數該= 服Γ求θ被傳送至複數個代理词服器中的某一特定代理飼 =奴如A,可將快取記憶體之負載,分配到區域網 之禝數個代理伺服器。 ΠReached array routing pr0c0; CARp). : = Memory Array Access Protocol (CARp), when computer system two. ': To retrieve a specific card identified by the universal tribute source addresser (URL) = URL is hashed, and according to this hash function, = = Γ to find θ is transmitted to a plurality of proxy server A certain proxy feed = slave A can distribute the load of cache memory to several proxy servers in the local area network. Π

_同樣地’在區域網路的外面設有數個網站”快取記憶 體^擎(cache engine),,亦可用於加速傳送資料元件。〜考 慮每個區域網路或電腦系統,彼等通常係經由網際網路服 ,提,者(Internet Service Provider,Isp)及/ 或網路 存取提供者(Network Access Provider,NAP)之設備連 接到網際網路。這些丨sp與NAp擁有伺服器相關設備,使得 彼等能夠發出訊息、甚至包含載有資料元件之訊息,而這 些汛息可在電腦系統和區域網路之間、以及和網站伺服器 之間傳送。這些ISP與NAP的伺服器也可以連接或裝設,,快 取記憶體引擎”、亦即快取記憶體,用於儲存經常被擷取 的貧料7L件。這使得丨Sp與NAp能夠傳送資料元件到執行 網路/劉覽器的電腦系統,因而使得電腦系統每次存取這些 資料元件時、無需從網站伺服器擷取該等資料元件。這些 集中式快取記憶體伺服器與伺服器軟體包括:_Similarly, "there are several websites outside the local area network" cache engine, which can also be used to accelerate the transfer of data components. ~ Consider each local area network or computer system, which is usually Connect to the Internet via Internet Service Provider (Isp) and / or Network Access Provider (NAP) devices. These SPs and NAp own server-related devices So that they can send messages, even messages containing data elements, and these floods can be transmitted between computer systems and local networks, and between web servers. These ISP and NAP servers can also Connect or install, "cache memory engine", that is, cache memory, used to store 7L pieces of lean materials that are often retrieved. This allows Sp and NAp to send data elements to the computer system running the network / Liu Lan, so that the computer system does not need to retrieve these data elements from the web server each time they access them. These centralized cache servers and server software include:

CacheflowTM ’其係由設立於美國加州sunnyVaie市的 Cacheflow公司所販售;Traffic ServerTM,其係由設立 於加州Foster市的Inktomi公司所販售;DynaCacheTM,其 係由設立於美國麻州Wal thanl市的Inf〇1 ibria公司所販CacheflowTM is sold by Cacheflow, Inc., SunnyVaie, California, USA; Traffic ServerTM, is sold by Inktomi, Inc., Foster City, California; DynaCacheTM, by Walthanl, Mass., USA Inf〇1 sold by ibria

第12頁 1223167 曰 案號 89112368 五、發明說明(7) 售;NetcacheTM,其係由設立於加州Sunnyvale市的 Network Appliance 公司所販售;以及 Cisco 5 00 Series Cache EngineTM,其係由設立於加州San Jose市的思科 (Cisco)公司所販售的。這些網際網路服務提供者(ISp)與 網路存取提供者(NAP)的π快取記憶體引擎”可能使用所謂 的”網站快取記憶體控制協定(Web Cache Control Protocol,WCCP ) π ,此協定會使電腦系統由原先向網站伺 服器發出資料元件擷取要求,重新改為向設於ISp與ΝΑΡ設 備中的快取記憶體引擎發出。一般來說,ISP或ΝΑΡ由他們 的快取記憶體引擎來運送資料元件(到接受他們服務的區 域網路與電腦系統),比起他們從設於任意網路的網站伺 服器運送資料元件,來得更快速。簡單地說,在後者的情 況、資料元件將引起較多延遲,此乃由於必須經由數個私 人網路來運送之故。 操作I SP/NAP之快取記憶體引擎,尚有其他技術可資 利用。請參見Geoff Huston,Web Caching,The Internet Protocol Journal (http://www· ciSCCKc〇m/warp/public/759/ipj一2-3 一webcaching· html )。這些技術包含Harvest,Squid 與 網際網路快取記憶體協定(lnternet Caching Protocol,I CP )。這些技術採用一個或一個以上之網站代 理飼服器當作快取記憶體引擎。Harvest與Squid提供集中 式快取記憶體的解決方案。關於網際網路快取記憶體協定 (I CP )技術,如果代理伺服器一時沒有電腦系統所要求 的特定資料元件的有效複製備份,則代理伺服器向另一個1223167 on page 12 Case number 89112368 V. Description of the invention (7) for sale; NetcacheTM, which is sold by Network Appliance Company, Sunnyvale, California; and Cisco 5 00 Series Cache EngineTM, which is established in San California, California Sold by Cisco in Jose. These Internet Service Providers (ISp) and Network Access Providers (NAP) 's π cache memory engine may use the so-called "Web Cache Control Protocol (WCCP) π", This agreement will cause the computer system to send data element retrieval requests to the web server, and reissue to the cache memory engine provided in the ISp and NAP devices. Generally, ISPs or NAPs use their cache engine to transport data components (to the local area network and computer systems that they serve), rather than they transport data components from web servers located on any network. Come faster. To put it simply, in the latter case, the data components will cause more delay, because it must be transported through several private networks. There are other technologies available to operate the cache engine of I SP / NAP. See Geoff Huston, Web Caching, The Internet Protocol Journal (http: //www.ciSCCKc0m/warp/public/759/ipj-2-3-webcaching.html). These technologies include Harvest, Squid, and the Internet Caching Protocol (ICP). These technologies use one or more web proxy feeders as cache engines. Harvest and Squid provide a centralized cache memory solution. Regarding Internet Cache Memory Protocol (I CP) technology, if the proxy server does not have a valid copy backup of the specific data components required by the computer system, the proxy server sends to another

第13頁 1223167 __案號 89112368_ 年 月_日攸工 五、發明說明(8) 代理伺服器發出查詢,以決定是否該另一代理伺服具有 該特定資料元件的複製備份。一般來說,各代理祠服器之 間具有高速溝通通道,因而採用此技術運送資料元件仍然 比從網站伺服器獲得該資料元件,更加快速。 快取記憶體裝置(cache appl i ances)甚至可使用在網 站伺服器’以加速他們的網頁之存取。這種產品的一個例 子是Novell ICSTM,其係由設於加州San J〇se市的N〇veiiPage 13 1223167 __Case No. 89112368_ Year Month_Day Yougong 5. Description of the invention (8) The proxy server issues a query to determine whether the other proxy server has a duplicate backup of the specific data element. Generally speaking, there is a high-speed communication channel between the agents and servers, so using this technology to transport data components is still faster than obtaining the data components from a web server. Cache devices can even be used on a web server ’to speed up access to their web pages. An example of such a product is Novell ICSTM, which is a product of Noveoi located in San Jose, California.

公司出品的,卻由許多原始設備製造廠商例如c〇mpaq公 司、設於德州奥斯丁市的De 1 1公司、以及設於紐約州 Armonk市的IBM公司來販售的。此種快取記憶體產品使得 網站伺服器擷取資料元件更有效率,因而也可以更快速地 將資料元件運送到所要求的電腦系統。另一種網站快取記 憶體服務稱為FreeflowTM,是由設於麻州康橋市的Akamai Technologies公司所提供的。根據這種服務,形成網頁的 資料元件被移居到由Free f 1 owTM所維護的伺服器網路。網 站飼服器上的網頁,必須重新設計、使得資料元件的通用 資源定址器(URLs)指向FreeflowTM服務的伺服器網路。像 這樣’當執行網頁瀏覽器的電腦系統發出要求到網站伺服 器擷取資料元件時,該等資料元件確實是從Free flowTM所 維護的伺服器中擷取出來的。當一電腦系統要求某一資料 凡件時,Free flowTM服務網路會選取一特定快取記憶體伺 服器去運送資料元件,該特定快取記憶體伺服器是很,,靠 近’’該一要求資料元件之電腦系統的,亦即,唯有如此才 能夠最有效率去運送資料元件。 ___^國專利第6,0 2 6,4 7 4提出另一種關於’’網站快取記憶The company's products are sold by many original equipment manufacturers such as cmpaq, De 1 1 in Austin, Texas, and IBM in Armonk, New York. This cache memory product makes the web server more efficient in retrieving data components, so it can also ship the data components to the required computer system more quickly. Another website cache memory service, called FreeflowTM, is provided by Akamai Technologies, based in Cambridge, Massachusetts. According to this service, the data elements forming the webpage are relocated to the server network maintained by FreefowTM. The web pages on the web server must be redesigned so that the universal resource locators (URLs) of the data components point to the server network of the FreeflowTM service. Like this', when a computer system running a web browser sends a request to a web server to retrieve data components, the data components are indeed retrieved from a server maintained by Free flowTM. When a computer system requests a certain piece of data, the Free flowTM service network will select a specific cache memory server to transport the data components. The specific cache memory server is very close to `` the request The computer system of data components, that is, the only way to transport data components most efficiently. ___ ^ National Patent No. 6,0 2 6, 4 7 4 proposes another kind of cache memory

第14頁 1223167 修正 曰 案號 89112368 五、發明說明(9) 體的解決方案。根據該專利文件,複數電腦系統中的每 一個電腦系統、或即一節點(n 0 d e )的儲存空間之一部分 (磁碟,記憶體電路··等),被安排用來當作全域位址共 旱記憶體空間(globally addressable shared memory space)使用。而每一節點,被提供有共享記憶體子系統控 制程式(shared memory subsystem control program), 該程式使得各節點能夠存取共享記憶體空間。共享記憶體 空間被區分成為許多原子單元,例如每頁4k位元組的原子 早凡。一特定節點被指定恆久地儲存各頁,亦即該特定 點將各頁儲存在該特定節點的實體記憶 全域位址共享記憶體空間的那一部份内。另有一"負女 點被指定去追蹤各頁的恆久性儲存節點的一致性、ρ (Uentity)。各節點維持有一目錄結構、以 =的:域位址(gl〇bal add⑽)的一頁,鏗定該= 予P 4 ΐί ’而該負責節點乃在追縱怪久地儲存有該被賦 域位址的-頁的節點之-致性的。“ΙΟ 點所周知的一網路位置, j硌上各即 播案映射到組成該檔案的各頁二入:錄、、Ό構’用以將每- -檔案’該節點首先存取點:月立址。因此’要存取 點。然後,該節點點的負責節 送到該-怪久性儲存;:致性摇該節點就會將要求信息傳 以獲取該某一已知頁的信鱼。Page 14 1223167 Amendment Case No. 89112368 V. Explanation of the invention (9) system. According to the patent document, each computer system in a plurality of computer systems, or a part of the storage space of a node (n 0 de) (disk, memory circuit, etc.), is arranged as a global address Use of globally addressable shared memory space. Each node is provided with a shared memory subsystem control program, which enables each node to access the shared memory space. Shared memory space is divided into many atomic units, such as 4K bytes per page. A specific node is designated to store pages permanently, that is, the specific point stores pages in the physical memory global address shared memory portion of the specific node. Another " negative female node is designated to track the consistency, ρ (Uentity) of the persistent storage nodes of each page. Each node maintains a page with a directory structure with =: domain address (gl0bal add⑽), which determines that = = P 4 ΐί 'and the responsible node is storing the assigned domain for a long time. Address-of-page nodes. "I0 is a well-known Internet location. Each instant broadcast on j 硌 is mapped to the pages that make up the file. The entry, record, and structure are used to firstly access the file for each node: Monthly address. So 'require the access point. Then, the node's responsible section is sent to the-strange long-term storage ;: The node will send the required information to obtain the letter of a certain known page fish.

第15頁 性儲存節點的1性二;;;::定=定ί頁 MM 8911^RR 五、發明說明(10) 曰 修正 美國專利第6,026,474更進一牛担 記憶體空間,使得各節點能夠%=及:使用此種共享 劉覽器應用程式的快取記憶體了子在各卽點上裝設的網頁 不像先前描述的各種技術, θ 站伺服器側執行快取記憶體,此一二:fISP/NAP側或網 在”客戶端”執行的,亦即,在取記憶體技術是 ⑽取記憶體的。比起在 乂端快取記憶體技術二7^ 二亍快的操作者具有較高自主性去 飼服器側的;望而:需要依賴1s崎側及/或網站 二)度τ ί ί ί網在區域網路中客戶端節點之間資料傳輸 速度更快二區域網路或客戶端節點的 端節點時具有速度上的優點。The first and second of the sexual storage node on page 15 ;; :: set = fixed 页 page MM 8911 ^ RR V. Description of the invention (10) Amend the US patent No. 6,026,474 to enter the memory space of the cattle, so that each node can% = And: The cache memory of this shared browser application is used. The web pages installed on each site are not like the various technologies described previously. The cache memory is implemented on the server side of the θ station. : The fISP / NAP side or network is executed on the "client", that is, the memory retrieval technology is to grab memory. Compared with the second-stage cache memory technology, the 7 ^ 2 operator has a higher degree of autonomy to go to the feeder side; hope: you need to rely on the 1s saki side and / or the website 2) degree τ ί ί ί ί The data transmission speed between the client nodes in the local network is faster than that of the local network or the end nodes of the client nodes.

#1 A/ ^ m 戶端的快取記憶體解決方案對ISP/NAP ;:; 階声* _ 各種的解決方案可以架構成為一種 ISP曰/Ap伽/安思體系統。事實上,在各網站伺服器端、 起來使用而σ接徂卢端的快取記憶體解決方案,皆可被連結 吏用而k供-種最佳化的解決方案;目為、 解決方案’其平均資料元件操取時間要快 ^ ’美y利案第6,〇26, 474號所描述的技術卻也 第16頁 1223167 ___篆!虎 89112368 _年月日_修正__ 五、發明說明(11) 有某些缺點。最明顯的是,第6, 02 6, 474號專利的系統, 其在每一次存取一檔案時,必須將該一檔案的恆久地儲 存,從一節點”移居”到另一節點。換言之,不僅只是從現 在節點L被當作一檔案的恆久性儲存,去提供該一檔案的 複製備份而送到要求的節點,還要將當作該一檔案的恆久 性儲存的責任、委由要求節點來負責。這就不是時常所要 的結果。特別是考慮到有一負責節點已被指定用於追蹤當 作一檔案的恆久性儲存的節點,因此每一 要改變恒久性儲存節點,一般而言、是沒有效率:案時都 本發明之目的,乃在克服上述習知技術的缺失。 目的與其他 一種方法及 個資料元件 一個連接到 取記憶體。 結而成為一 作為一儲存 特定資料元 件的複數節 點的指定, 給想要存取 變。 元件、其— 目的可 系統, 之區域 區域網 至少在 υα 早一的 節點、 件的複 點中的 不會藉 該特定 【發明内容】 根據本發明,此— 一實施例,本發明提供 節點的可使用的一或多 佳化。複數節點中的每 供有一區域性、實體快 的每一個快取記憶被鏈 體。一特定節點被指定 特定資料元件,且提供 取卻缺少該特定資料元 把特定節點當作儲存節 的資料元件的複製備份 參考節點而來作單獨改 亦即,對於一資料 獲致達成。根4 其對來自伺服^ 快取記憶體作▲ 路的節點,都去 一些複數個節黑 虛擬快取記憶 以恆久地儲存一 製備份給希望存 其他參考節點。 由提供一個特定 資料元件的一個 郎點的儲存功能是相 1223167 案號 89112368 五、發明說明(12) ___ 地靜態的。舉例來說,假設一第一參考節點 得特定資料元件的一複製備份、且對資料元的:即點獲 執行一組包含一次或多次的存取操作。如黑设製備份 ;’其缺少、但希望存取該特定資料元件,= 節 f供該特定資料元件的一複製備份給該另—夫二存即點 這些當中’這會降低追蹤每一資 ς二點。在 性⑴响)的負擔。考慮一典型節點2 =之-致 貧源之節省是相當可觀的。 卞個貝枓凡件,處理 一根據另貫施例,其中一個節點被指定Α 凡件的一個監視節點。詨g — 疋$作特定資料 被當作特定資料-从以""視點’稭由提供特定節點1 節點所發出之一致性,而來考 做出回應。識別特疋資料元件的儲存庫節點之要求 依貫例之句gg 特定資料元件,就,;參考節點想要存取-個其所缺少的 ㈣cess)。該參ί =行::階層過程(hWchlCal 向參考節點指出,攸二適當的監視節點(被指定為 件之-致性)、獲;;指定去餘存想要的資料元 致性。該*考節點, 件之適當儲存節點的一 件的一複製備份。…、後攸儲存卽點去獲得想要的資料元 依實例之說明,炎 網路偵測—或多個訊ί ^ =點可能逐步增加地儲存從區域 監視節點、該等監視^仔到之資訊’及識別各種各樣的 儲存節點再被指定 二又被指定去識別儲存節點(該等 應的資料元件)。除外、或是 第18頁 1223167# 1 A / ^ m The client's cache memory solution is for ISP / NAP;:; order sound * _ Various solutions can be structured as an ISP / Apga / Ansi system. In fact, the cache memory solutions that can be used on the server side of the website and are connected to sigma can be used by the linker to provide an optimized solution; the purpose is, the solution 'its Faster average data component access time ^ 'US yli case No. 6, 〇26, 474 but the technology described on page 16 1223167 ___ 篆! Tiger 89112368 _ year month day _ correction __ V. Description of the invention (11) There are certain disadvantages. The most obvious is that the system of Patent No. 6, 02 6, 474, each time a file is accessed, the file must be permanently stored and "migrated" from one node to another. In other words, not only is node L being used as a permanent storage of a file, to provide a copy of the file and sent to the requesting node, but it is also necessary to delegate the responsibility and responsibility for the permanent storage of the file. Ask the node to take responsibility. This is not always the desired result. Especially considering that a responsible node has been designated to track a persistent storage node as a file, therefore, each time a permanent storage node is changed, it is generally inefficient: the purpose of the present invention at the time of the case, It is to overcome the lack of the above-mentioned conventional techniques. The purpose is connected to the fetch memory with one other method and one data element. It becomes a designation of a plurality of nodes as a storage of a specific data element, and it is changed for a desired access. Component, its — purpose can be systematic, the regional area network is at least one node earlier than υα, the complex point of the component will not borrow this particular [Summary of the Invention] According to the present invention, this-an embodiment, the present invention provides a node One or more optimizations that can be used. Each of the plurality of nodes is provided with a regional, physical cache, and each cache memory is chained. A specific node is designated with a specific data element and provides a lack of the specific data element. The specific node is regarded as a duplicate backup of the data element of the storage section, and the reference node is modified separately, that is, a data is achieved. Root 4 For nodes that serve as a path from the servo ^ cache, all of them go to a number of sections. The virtual cache memory stores the system backup for a long time to other reference nodes that wish to store it. The storage function of a point provided by a specific data element is phase 1223167 case number 89112368 V. Description of the invention (12) ___ It is static. For example, suppose a first reference node obtains a duplicate backup of a specific data element and performs a set of access operations that include one or more times on data elements: click-to-get. Such as black backup; 'It is missing, but wants to access the specific data element, = Section f for a copy of the specific data element to the other-husband and two save and click these' This will reduce the tracking of each resource. two points. The burden of sexual noise). Considering a typical node 2 =-the savings caused by poor sources are considerable. For each piece of processing, according to another embodiment, one of the nodes is designated as a monitoring node of the component.詨 g — 疋 $ is used as specific data. It is treated as specific data-from the "view" point of view, to provide a response from the specific node 1 node. Repository node requirements to identify special data elements According to the usual sentence gg specific data elements, just; reference nodes want to access-a missing cess). This parameter = line :: Hierarchical process (hWchlCal points out to the reference node, the second appropriate monitoring node (designated as the consistency of the pieces), acquisition ;; specifies the consistency of the desired data element. The * Test node, a copy of one of the appropriate storage nodes of a piece of file ...., and then save the point to get the desired data element. According to the description of the example, Yan Network Detection—or multiple messages. ^ = Point may Gradually increase the storage of information from regional surveillance nodes, such surveillance information, and identify various storage nodes, which are then designated and identified to identify storage nodes (the corresponding data elements). Except, or Chapter 1223167

另一種選擇 ’ν · 7χ ® 机恩指向一々且令古 多個節點(例如,多重播放(multicast)或廣播有—或 (broadcast)訊息),去要求確認適當監視節^的— 性、該適當監視節點其可以識別用於儲存想要欠_ 件、或一個資料兀件群組其包含想要 &艸凡 點。 文扪貝枓兀*件的儲存節 【實施方式】 本發明技術供使用於網站存取網際 中連接到區域網路的複數個人電腦李 ' —心叙體’其 網頁劉覽器應用軟體。再者,點丄皆能夠執行 覽器應用軟體時’每個節點企圖呈現從—在丁網” 的網站伺服器所存取得來的一網頁 二域、,罔路遠端 或網際網路。’ 一此種網頁假設是由資料元:::域網路 即,包含文件、聲音資料、靜態影像、電影、圖^ = 應用軟體等等。然m,熟悉本項技藝人’ ς =、 :可他應用軟體上’以使複數節點;以:;月 取,可達到最佳化。 彳只丨叮化1千存 第一圖係顯示本發明實施例的網路丨。 網路3,廣域網路5與網站伺服器網路7聿;:企業 飼服器網路7二者可以是乙太區域網路。二:::與網站 是網際網路。…斤示,網站伺服器網路例 一伺 娜,以及^線二或Another option, 'ν · 7χ ® machine, points to a network with multiple nodes (for example, a multicast or broadcast message) to request confirmation of the proper monitoring of the festival, the appropriateness The monitoring node can identify a component for storing the missing items, or a group of data pieces which contain the & ordinary points. Storage section of text files [Embodiment] The technology of the present invention is used for a plurality of personal computers connected to a local area network in a website accessing the Internet. In addition, when the browser is able to execute the browser application software, 'Each node attempts to present a webpage, two domains, Kushiro remote, or the Internet, which are stored from a web server in Dingwang.' This web page is assumed to be composed of data elements ::: domain network, that is, files, sound data, still images, movies, pictures, ^ = application software, etc. However, those who are familiar with this artist 'ς =,: 可 他The application software is used to enable the plural nodes; to use :; monthly acquisition can be optimized. 彳 only 丨 Dinghua 1000 thousand first picture shows the network of the embodiment of the present invention 丨 network 3, wide area network 5 And the web server network 7 聿;: the corporate feeder network 7 can both be an Ethernet LAN. 2 :: and the web site is the Internet .... for example, a web server network example Na, and ^ line two or

第19頁 服器。企業網路3包含一路由器::,疋华 1223167 案號 89112368 五、發明說明(14) t 月日 倏正 nl2、nl4、nl6、nl8、n20、n22、…、心6、…、n3〇、 n32、n34、n36、n38、n4G、n42 與 n44,其可以是個人電 腦網路及/或伺服器。路由器“與以在對送往或來自網際 網路5的路由器封包作服務。集線器或開關h2、^與“其 作用是在節點n4至π44之間作重複資料溝通,因而在節點 π4至π44之間達成了虛擬仿匯流排環境(virtual bus_Uke environment)。節點n2sn44可執行各種處理功能,例如 產生與傳送封包到其他節點,接收與處理從其他節點傳送 來的封包’與處理接收封包的資料。P. 19 server. The enterprise network 3 includes a router ::, Huahua 1223167, case number 89112368 5. Description of the invention (14) t month day is nl2, nl4, nl6, nl8, n20, n22, ..., heart 6, ..., n3〇, n32, n34, n36, n38, n4G, n42 and n44, which can be personal computer networks and / or servers. The router "is used to serve router packets sent to or from the Internet 5. The hub or switch h2, ^ and" its role is to make repeated data communication between nodes n4 to π44, and A virtual bus_Uke environment was reached. The node n2sn44 can perform various processing functions, such as generating and transmitting packets to other nodes, receiving and processing packets' transmitted from other nodes, and processing data of received packets.

一般來說,網際網路5是將由網路存取提供者() 和網際網路服務提供者(ISP)所維護的複數個私人網路 互連起來,其乃是藉由各種高容量(即,n 、T3 、T4、 0C_3、0C-48等)的專用合約電話線網路線路來達成網路 的互連。網際網路内的通訊,乃是使用一階層式通訊協定 而達成的,這些協定包含:網際網路協定(Internet protocol,IP)、傳輸控制協定 protocol, TCP)、與超文件傳輪協定(HypertextGenerally speaking, the Internet 5 interconnects a plurality of private networks maintained by an Internet access provider (ISP) and an Internet service provider (ISP). , N, T3, T4, 0C_3, 0C-48, etc.) dedicated contract telephone line network lines to achieve network interconnection. Communication within the Internet is achieved using a layered communication protocol. These protocols include: Internet protocol (IP), Transmission Control Protocol (TCP), and Hypertext Transfer Protocol (Hypertext)

transfer protocol,http)。值得一提的是,網際網路5 可以傳送請求資訊的訊息(以封包方式),並從來源節點 (source node)傳送被請求的資訊到適當的目的節點 (destination node)去。舉例來說,節點可以”拜訪網 站n ,並且對特定網站伺服器發出查詢、取得網頁資訊 後i將該’m現出來。網站伺服器以傳送被請求的網 頁貝料至清求節點作為回應。由於網際網路的架構盥運作 6、經是么V &的’其細節部分不再多作描述。要言之Ϊ透過 1223167 案號 89112368 五、發明說明(15) 網際網路5傳送資料、例如從節點n 2到節點n 2 〇,一般是比 在區域網路3或7内傳送資料、例如從節點nl〇到節點n28, 其傳送效率要低很多(花更多時間,使用較多資源、 寬,等等)。 ^transfer protocol, http). It is worth mentioning that the Internet 5 can send messages (in the form of packets) requesting information, and send the requested information from the source node to the appropriate destination node. For example, a node may "visit website n" and issue a query to a specific web server to obtain the web page information i. The web server then sends the requested web page to the requesting node as a response. Because of the structure of the Internet, the details of the V & amp's are no longer described in detail. In other words, through the 1223167 case number 89112368 V. Description of the invention (15) The Internet 5 transmits data, For example, from node n 2 to node n 2 0, it is generally more efficient than transmitting data in local network 3 or 7, such as from node n10 to node n28, which has a much lower transmission efficiency (takes more time and uses more resources) , Width, etc.) ^

第一圖顯示節點n2到節點n44的更詳細内部架構。注 意第二圖中所示的結構實例為了方便討論,已經作適去 化。例如,該節點具有單一匯流排(singU 中央處理器(cpu)或處理器10、一高速記憶體22、一主 記,體24、-磁碟記憶體26、—輸人/輸出裝置或網路介 面卡(例如乙太網路介面卡)18、一輸入裝置(例如 與例如滑鼠、追蹤板、追蹤球或搖桿等等的座 16、以及輸出裝置(例如圖畫加速器與監立 & 擴音器等)14。高速記憶體22包含用於;供二般:速版 二=靜態隨機存取記憶體(SRAM) t路,亦二 於卽點之局部快取記憶體之型式( 種用 幸咋愔驴9/|甘π a 飞、相對於主記憶體24 )。 主圯It體24,其可以疋動態隨機存取記 序列動態隨機存取計憶體(SDRAM ), '板 )、或 式與資料之正常揮發性工作或儲存 疋供該即點作程 節點-般的恆久性儲存。處理器 磁碟記憶體以提供該 作。例如’處理器10執行根據本發C:令與資料運 等指令可被儲存在記憶體22至26中 /之私式指令,此The first figure shows a more detailed internal architecture of nodes n2 to n44. Note that the structure example shown in the second figure has been adapted for the sake of discussion. For example, the node has a single bus (singU central processing unit (cpu) or processor 10, a high-speed memory 22, a master, a body 24, a disk memory 26, an input / output device or a network Interface cards (such as Ethernet interface cards) 18, an input device (such as a dock 16 such as a mouse, trackpad, trackball or joystick, etc.), and an output device (such as a graphics accelerator and monitor & expansion) Musical instruments, etc.) 14. The high-speed memory 22 contains the following: for the second version: the speed version two = static random access memory (SRAM) t, and also the type of local cache memory at the point Fortunately, the donkey 9 / | gan π a fly, relative to the main memory 24). The main it body 24, which can be a dynamic random access memory sequence dynamic random access memory (SDRAM), 'board', OR and the normal volatile work or storage of data, for the point-of-work node-like permanent storage. Processor disk memory to provide the work. For example, 'processor 10 executes according to this C: order and Instructions such as data movement can be stored in the private instructions of the memory 22 to 26.

節點與連接該節點之區域網路之;,介面卡18,其在 與資料。值得一提的是,網路介面/封包方式傳送訊息 層位址、指向本節點之問題封包、 可以區別具有MACThe node is connected to the local area network; the interface card 18, which is connected to the data. It is worth mentioning that the network interface / packet mode sends the message layer address, the problem packet pointing to this node, and it can be distinguished that it has MAC

第21頁 輸入裝置1 6從操作人員接收輪入此人及非此種之封包。 __窗舰腸. _Z:---例如,去拜訪各種 1223167 j號 891123fiS 五、發明說明(16) 網^的命令。輸出裝置1 4用於呈現視覺與聽覺資訊給操作 人員,亦即,呈現網頁之視訊與音訊效果。 、μ ^第二圖所示,根據本發明,記憶體22至26之一部分 被叹疋為快取記憶體丨〇 〇。作為一實例,(處理器丨〇,其 根據本發明執行合適的軟體)節點^到節點中、至少 f二些節點,將其磁碟記憶體26之至少一部份指定給快取 j =體100(根據本發明,記憶體22與24可視需要、一次又 二1田作快取纪憶體使用)。此快取記憶體丨⑽ dlrect〇ry)^^(partltl〇n)^ =:情況下*網路㈣器程式設定,用於 存、 ;;'近存取之資料元件。另-種做法是,此快取= ::7個別的磁碟目錄或分區…於儲存資料元件Ϊ 憶體,作比起正當在跑-ί隐 谷1受限制之儲存記 高的户t ΐ ΐ 存資料元件之,,主記憶體",1且有較 间的存取速率;在此情形中," /有車乂 器網路7聚集而形成。 记隐體乃係由網站伺服 此等節點之快取記憶體100”連 :由所有節點存取的單一虛擬快取在成-個 :點可以對形成此共同虛擬快體n,各 即點中所儲存之資料元件作存取。4份之其他各 數節點中快取記憶體之”連处 :據本毛曰月,此種在複 f巧之查詢通信程序所實連施。心料 ·,、、占之處理器1 〇執行軟體程式且次 又明、此等谛 取記情許夕、i έ士 ν , 、 σ蔓貝料結構’其合達成体 作更詳細說明。 °己匕體,此點將在以下 1223167 修正P.21 The input device 16 receives packets from the operator and those who do not. __Window ship bowel. _Z: --- For example, to visit various 1223167 j 891123fiS V. Description of the invention (16) Order of the Internet ^. The output device 14 is used to present visual and auditory information to an operator, that is, to present video and audio effects of a web page. As shown in the second figure, according to the present invention, a part of the memories 22 to 26 is sighed as a cache memory. As an example, (processor 丨 0, which executes suitable software according to the present invention) nodes ^ into nodes, at least f two nodes, and assign at least a part of its disk memory 26 to the cache j = body 100 (According to the present invention, the memories 22 and 24 can be used as cache memory memory once and twice as needed). This cache memory 丨 dl dlrect〇ry) ^^ (partltl〇n) ^ =: In case * network device program settings, used to store, ;; 'Recently accessed data components. Another way is to use this cache = :: 7 individual disk directories or partitions ... to store data elements Ϊ memory, which is higher than the legitimate storage of the limited storage of the hidden valley 1 t t ΐ资料 Among the data components, the main memory is "1" and has a relatively fast access rate; in this case, the "/ vehicle network 7" is formed by aggregation. Cryptography is the cache memory 100 ”of these nodes that are served by the website: a single virtual cache accessed by all nodes is formed into a set: points can be paired to form this common virtual cache n, each in a point The stored data elements are used for access. The "connection" of the cache memory in the other four nodes: According to this book, this type of connection is implemented by the complex communication program. I expect that the processor of the processor 10 executes the software program and it is clear that the following are the details of the relationship between Xu Xi, I 士, ν, and σ vine shell material ’. ° Self-dagger body, this point will be corrected below 1223167

案號 89112368 、發明說明(17) 假設各節點之處理器ίο可以執行多樣的軟體程式以 賢ϊίί定t㉟。例如,一節點之處理器10可執行網頁瀏 。應用程式’使得操作人員可查訪各種網頁,擷取 ::將網頁顯不《。有此操作能力的節點,可被視為一 即點(browsing node)。根據本發明,一節點之處理 可以執行軟體程式而達成以下三種功能之一,即:: 參考、或獲得資料元件之複製備份以提供給瀏覽節點&, (b )當作資料元件之儲存器或恆久性儲存器,以及 監視那些當作資料元件之儲存器之節點的一致性。=各 作上述三功能之一的各節點,其分別被當作是參考節點田 (referencing node)、儲存節點(rep〇sit〇ry 、”’與 監視節點(monitor node)。注意,一節點可能呈有上述、 -種、二種或全部角色,亦即、—節點可以是參考節點、 健存節點、與監視節點之任何組合。一節點的角色, 是跟據其特定時點執行的程序、以及特定時點其 之型式,而隨時變化。為簡明計,下文中將不會談論 點之處理器1 0執行何種處理,而會以「這節點執: 能」來替代。 77 此外,在以下的2中,這些節點將被描述為傳遞盘 接收封包,其包含讯心與/或資料元件之封包。而、 中,此種達成該特定機制的描述,將大大被略去。孰来 技藝的人士,將可了解運用許多電路和技…j二 的。以實例而言,各節點使用各自的产@ 1 至1在目 π合目的處理器10、印愔駚99 至26、輸入/輸出設備或網路介面卡18與使用公知丫合適 6^冓it m 包。再者,根據本發明‘明, 1223167 —------篆號 89112撤___年月日 修正 ^發明說明^ 各節點乃使用傳輸控制協定/網際網路協定(TCP/IP )來 漢通訊息與含有封包的資料元件。然而,這僅僅是以實例 說明而已。 如上文所述,一參考節點當作參考、或獲得一特定 被存取資料元件之一複製備份之用。根據本發明,一參考 卽點使用一階層式存取過程(hierarchical access process)達成前述功能,該過程包括以下步驟: (1 )識別一資料元件的監視節點; (2 )如果可能,查詢監視節點以識別一資料元件的儲存 節點;與 (3 )如果可能,從適當的儲存節點擷取該資料元件。 唯有在參考節點在此階層式存取過程失敗時,參考節點才 會以常規的方式從網際網路去擷取資料元件。此點將會在 下文作更詳盡描述。 為協助一參考節點,該參考節點最好以漸增式建構並 維護一監視表(moni tor table),以方便識別一些特定資 料元件的監視節點。表一提供這種監視表的一個範例。 表一 URL Group 監視位置 www. uspto. gov. tw/web/menu/pats. htm; Pepper (胡椒) www. somehost. com/linkX Salt (鹽) — www. mango, com/help/genhelp. html Snort (喷鼻息) www. somehost. com Garlic (蒜)Case No. 89112368, Description of the Invention (17) It is assumed that the processor of each node can execute various software programs to determine t㉟. For example, a node's processor 10 can perform web browsing. Application ’allows operators to access a variety of web pages, to retrieve :: display web pages. A node with this operation capability can be regarded as a browsing node. According to the present invention, the processing of a node can execute a software program to achieve one of the following three functions, namely: reference, or obtain a duplicate backup of the data element to provide to the browsing node & (b) as a storage of the data element Or persistent storage, as well as monitoring the consistency of those nodes as storage of data elements. = Each node that performs one of the three functions described above is considered as a reference node field (referencing node), a storage node (rep0sitory, "', and a monitor node. Note that a node may be It assumes the above-mentioned, one, two or all roles, that is, a node can be any combination of a reference node, a healthy node, and a monitoring node. The role of a node is a program that is executed according to its specific point in time, and At a specific point in time, its type changes at any time. For the sake of brevity, the following will not talk about what kind of processing the point processor 10 performs, but will replace it with "this node performs: can" 77 In addition, in the following 2 In these, these nodes will be described as delivery disk receiving packets, which contain packets of heart and / or data elements. However, such descriptions of achieving this specific mechanism will be largely omitted. People who come to the arts, You will learn how to use many circuits and techniques ... j. For example, each node uses its own product @ 1 to 1 in the destination processor 10, India 99 to 26, input / output device or network. Road Interface Card 1 8 is compatible with the use of the well-known 6 ^ 冓 it m package. In addition, according to the invention 'Ming, 1223167 ---------- 篆 89891 withdrawal ___ year month day amendment ^ invention description ^ each node is using transmission Control Protocol / Internet Protocol (TCP / IP) communication and data elements containing packets. However, this is just an example. As mentioned above, a reference node is used as a reference, or a specific protocol is obtained. Access to one of the data elements for copying and backing up. According to the present invention, a reference point uses a hierarchical access process to achieve the aforementioned functions. The process includes the following steps: (1) identifying a data element for monitoring (2) if possible, query the monitoring node to identify the storage node of a data element; and (3) if possible, retrieve the data element from the appropriate storage node. Only at the reference node in this hierarchical access process When a failure occurs, the reference node will retrieve data components from the Internet in a conventional manner. This point will be described in more detail below. To assist a reference node, it is best to increase it gradually. Build and maintain a monitoring table (moni tor table) to facilitate identification of monitoring nodes for specific data elements. Table 1 provides an example of such a monitoring table. Table 1 URL Group monitoring location www.uspto.gov.tw/web / menu / pats. htm; Pepper www. somehost. com / linkX Salt — www. mango, com / help / genhelp. html Snort www. somehost. com Garlic

第24頁 1223167 修正 曰Page 24 1223167 Amendment

_A 案號89112郎8 五、發明說明(19) ,一的監視表係由一系列的記入項目(entries)所組成, 每個記入項目至少包含一對形態(type)各異的元素。形態 『URL Group』、亦即第一元素,是資料元件群組的一個 識別子。參考節點可以使用這個元素當作索引,亦即,參 考節點可以比較表格記入項目之『URL Group』元素與資 料元件的通用資源定址器(URL),去鑑定對應的比對表記 入項目。形態『監視位置』、亦即第二元素,用於識別一 ,,考節點最後知道的特殊監視節點,該特殊監視節點被 指定去識別、其URL和同記入項目的URL元素比對成功的對 應貧料凡件之儲存節點。例如,儲存表一的參考節點,一 如儲存在其監視表的資訊顯示節點” s n 〇 r ^ (噴鼻息),,是 以『URL Group』元 素 www.mango.com/help/genhelpehtmln 所識別出來的一 資料元件的監視節點;亦即,在此指出監視節點,,sn〇rt (喷鼻息)π被指定為識別具有『URL Group』元 素 www.mango.com/help/genhelp.html” 的資料元件的儲_A Case No. 89112 Lang 8 V. Description of the Invention (19), a watch list is composed of a series of entries, each entry contains at least a pair of elements of different types. The form "URL Group", that is, the first element, is an identifier of a data element group. The reference node can use this element as an index, that is, the reference node can compare the "URL Group" element of the table entry with the universal resource locator (URL) of the data element to identify the corresponding comparison table entry. The form "monitoring position", that is, the second element, is used to identify a special monitoring node that the test node knows last. The special monitoring node is designated to identify, and its URL and the URL element of the registered item are successfully matched. The storage node of all the poor materials. For example, the reference node storing table 1, like the information display node "sn 〇r ^ (sniffing)" stored in its monitoring table, is identified by the "URL Group" element www.mango.com/help/genhelpehtmln A monitoring node of a data element; that is, a monitoring node is pointed out here, snort (sniffing) π is designated to identify a material having the "URL Group" element www.mango.com/help/genhelp.html " Component storage

• somehost· com/1 ink』對照到所有具有URL資訊 • somehost· com/link』的所有資料元件,例如URL 存卽點。然而,『U R L G r 〇 u p』元素不需要對照特定的每 一個別資料元件,而可以對照整群組之資料元件來替代 之。區分貧料το件成為群組的一個方法是,根據通用資源 定址為(URL)來區分。就如所看到的,每個『url Group』 記入項目所包含的資訊要比資料定址表(data 1〇cat()r table)之URL記入項目所包含的為少;因此,會對照到具 有相同一般URL資訊的所有資料元件。亦即,第一記入項"Somehost · com / 1 ink" compares to all the data elements with URL information • somehost · com / link ", such as URL storage points. However, the "U R L G r 〇 u p" element does not need to refer to each specific data element, but can replace it with the entire group of data elements. One way to distinguish the poor materials into groups is to distinguish them based on the universal resource address (URL). As you can see, each "url Group" entry contains less information than the URL entry of the data addressing table (data 10cat () r table); therefore, it will be compared to All data components of the same general URL information. That is, the first entry

WWWWWW

WWWWWW

Η 第25頁 1223167Η Page 25 1223167

^www.s〇mehost.c〇m/linkX/ad.glf , „ t # , URL ㈣w.somehost.c〇1D/linkY/ad gif』的資料元件。當 二n的分組也還有其他的方法。對資料元件選取 頁為依歸’ ®為這樣會導致在相同的 儲存卽點找到該等資料。監視表的使 更作詳盡描述。 /、建構,下文中將 、此外,要特別一提的是:監視表可為”區域 或”遠端的”監視表。一區域監視表包含:其中參: 當作 越監攝 是監視節點的所有URL群組。一遠端監視表包含$一節點即 監視節點的所有已知節點,而對於每一監視節點、^ www.s〇mehost.c〇m / linkX / ad.glf, „t #, URL ㈣w.somehost.c〇1D / linkY / ad gif”. There are other methods for grouping n . The selection page of the data components is based on the "®". This will lead to the finding of such data at the same storage point. The monitoring table will be described in more detail. /, The construction, which will be described below, in addition, it is particularly important to mention : Watch list can be "area" or "remote" watch list. An area monitoring table contains: Among them: as: Surveillance camera is to monitor all URL groups of the node. A remote monitoring table contains $ a node, that is, all known nodes of the monitoring node, and for each monitoring node,

表包含一串其所監視的所有已知URL群組。 _ 再者,每一監視節點漸增地建構並維護資訊於〜〜 定址表(data locator table)R,而該監視節點用貪举 特殊資料元件的儲存節點。表二提供資料定址;識为 例。 、〜個||The table contains a list of all known URL groups that it monitors. _ Furthermore, each monitoring node gradually builds and maintains information in a ~~ data locator table R, and the monitoring node uses a storage node that greets special data elements. Table 2 provides information for addressing; identification is an example. , ~ A ||

ϋΕ 儲存位置 ^ www. somehost. com/linkX//ad. gif Salt (鹽) 〜〜^- www. somehost. com/logo, gif Garlic (蒜) 〜 www. mangosoft. com/ Snort (噴鼻息) www. uspto. g〇v Snort (喷鼻息)^^、ϋΕ Storage location ^ www. somehost. com / linkX // ad. gif Salt ~~ ^-www. somehost. com / logo, gif Garlic (garlic) ~ www. mangosoft. com / Snort (nasal breath) www . uspto. g〇v Snort (snorting) ^^,

第26頁 1223167 _a. 曰 修」 案號 89112368 五、發明說明(21) 該元素當作資料元件的識別子(identifier)去索引資料定 址表的二個記入項目。請注意,在網路環境使用url去指 定特殊育料7L件、比在網頁中使用相同方式去指定一資料 元件' 比較受歡迎。當然’這可以用任何合適:識別子來 取而代之。此外,表一中使用” S n 〇 r t (噴鼻息),,一 ♦司, 僅是為了說明的緣故。其實,在監視表的記入項目内、最 好是記入節點” snort"之媒體存取控制層位址(mac)盥/ 網際網路協定(IP)位址。 ”〆 n i己3二的第二元素’儲存位置’乃指示著監視節點 所熟知的儲存節點(像MAC位址與IP位址)、其被用 存對應的資料元件,該資料元件是由同表格的記人項目中 被識別出來的。較佳選擇是’監視節點也對每一群組的所 兀件的數目保持計算。f料定址表的使用與建構, 也將在下文中作更詳盡描述。 执一 ί二圖,現在要描述根據本發明的處理流程。假 Γ3:點想要顯示一網頁。在獲取網頁資料的過程 覽印點面臨一個命令:在網頁内、去擷取由一已知 Γ老所鑑定的資料元件。根據本發明,劉覽節點向-特定 劉ί 求去摘取資料元件。以實例而言,那些被 考狄=1 &取的特定參考節點可能是預設的。例如,參 宏I^可能是—被指定的網際網路代理伺服器。此等被指 體,#I f伺服态可被設定於大多數的網頁瀏覽器應用軟 E彳如微軟公司的『網際網路探險家(internet 節點』瀏覽器。另一種選擇是,瀏覽節點與參考 …、j以是相同的節點。____ ΗΐΙΒϋΙ 第27頁 互、發明說明(22) 址過圓中的步驟S2,參考節點執行-快取記憶體定 點決定L/ J i〇catT process),根據該過程、參考節 興趣:Π即:的一?生、該儲存節點乃是被指定儲存感 I::::,:;::;;- :子:點的位置鐘定失敗可能會發 二、;中, ί;:,)。然而這也可:發=本源 :未第一次存取資料元件時。在這種情ΐΐ新 點去鑑定儲存二^ ί枓兀件,故也尚未指定監# > 々々疋儲存即點。在步驟33,如果參 =皿硯郎 即』已經被鑑定過,則參 ς ·,决弋出儲存 參考節點本身被鑑定為是儲Cf :SB4:以決定是否 S8、參考節點決定是否這樣’到步驟 料已經過期,因此不能使用而:ΪΓ 行刪除資料W :記憶錯誤’在步驟S10 ’參考節畔勃 取記則:;:S,到步驟S11。如果Κί 體讀取資料元件:如;參考節點=部的快取記饿 存節點將資料送回表考:二中所標不’在二驟”執行:儲 後的情況,則儲,二即”,。執行步驟s7若是在步騍S9 炎沒有執行右立存即點也是參考節點,故實際上在牛顿Q 二=订有意義的動作。 仕步騍S7 想要去存取二:S4杜參考節點已決定出儲存節點是不同於 貝枓π件之複製備份的參考節點(參 门於 1223167 皇號89】1塗 五、發明說明(23) Λ 曰 修正 乏該:料元件)。如果是這樣, 嘗试攸被鑑定過的儲存節點去獲取次、/驟S5,苓考節點 節點發出一封包之訊息、貝料凡件。此可由參考 複製備份的該參考節點、而達^ 同路中要求資料元件之 決定是否錯誤已經發生。原來,如在步驟S6,參考節點 包的回應訊息給參考節點,指 ,存節點在發出一封 要求的資料元件,則在步驟^會=即點現在沒有儲存所 種理由會發生這種情況,包括;法二2錯誤發生。有好幾 製備份過期。如果沒有錯誤 :=節點的資料元件之複 提供被要求之資料元件的複“則: 的執行是在步驟S6之後,此時儲存J此;,步㈣ 的節點。於是,儲存節點以一個; :、另向其傳遞被要求的資料元件之複製二ί:ίΓ :參=:著參考節點偵測到錯誤發:, 如上所述,根據下列三個理由中之任一個,程十的袖 仃會到達步驟s 11,亦即: 式的執 =)在步驟S3參考節點不能鑑定資料元件之儲存節點, 考節^鑑定出本身即是儲存節點、但在步驟S8谓 =出=取記憶體錯誤,於是在步驟sl〇要求刪除資料元貝 jc )在步驟S6參考節點鑑定出_個非參考節點當作儲 =。上述任-種情況皆顯示:—個可㈣,亦@ 7的資料元件之複製備份是不存在於虛擬快取記憶體的 一 1即’沒有可用#、非過期的資料元件之複製備份儲存 第29頁 1223167 修正 曰 ^S_89U2368 五 、發明說明(24) 於任何一個其快取記憶體連結起來成為虛擬快取記憶體的 節點内)。若如此,參考節點必須嘗試從網際網路擷取資 料元件。於是,在步驟S11,參考節點經由網際網路發出 一封包訊息,指向網站伺服器(一如使用資料元件之URL 所鑑定者)去請求資料元件之一複製備份。在步驟§12, 芩考節點決定是否在嘗試經由網際網路擷取資料元件時、 有錯誤發生。像這樣的錯誤,其發生原因不一而足,這些 原因包括· ·連接區域網路到網際網路(路由器或其他設 備提=的)失誤,網站伺服器設備之失誤,等等。^果在 參考節點偵測出這樣之錯誤發生,然後在步驟s丨3,夂考 節點執行一個無效資料寫入過程(ab〇rt data訂丨。, process),稍後將更詳細描述此無效資料寫入過程。就如 細描·,無效資料寫入過程係涉及參考節點為了 件而與監視節點作溝通,,亦即:通知監視 朱 :郎點擷取資料元件已經失敗。在步糊 , 點不同於參考節點,則回覆一,,不領受(not 視即 父握(handshake ) n 程序。 假設在步驟S12,在嘗試從網際 參考節點執行—寫入資\生過二^ Process),然後如果需#,則於步㈣ 出&視不正確訊息的處理。 ^订适 區域快取圮愔妒叮-丄、、;^件寫入到參考節點之 系統敕俨决拙—耝, goSof t s MedleyTM的檔宰 可被儲存在擁ΐ 案號 89112368 發明說明(25) 五 曰 修正 全部共同的根資料元件的資料 S〇meh〇st. c〇m之資料元件^ (例如,所有關於 又已如上述,當"儲存節妲如此比較有效率。 步驟S7就會被執行。然而,如^,資料給參考節點時, 點被迫去從網際網路獲得資疋廷樣的腳本,則參考節 對儲存節點的定址失敗,或件之複製備份;這是由於 之備份給參考節點之故。因儲存節點不能提供資料元件 (與監視節點,在其任何—祜情況下,參考節點 指定本身當作資料元件之 a又要此資料兀件時)會 有意義的動作。 子即點。於是,在步驟S7並無 接著看第四圖,在此詳、十、 步驟S20,參考節點決定 ?取記憶體的定址過程。在 體儲存所要的資料元件。以:你;郎點其本地"之快取記憶 地之快取記憶體(例如第_二說明,參考即點將觀察本 象考節i執:所要的資料元件。另-種方案,如果 點mr劉覽器應用軟體(亦即也是劉覽節 S21,如果來^二喷裎時、一併執行之。在步驟 體ΘH ~即 ' '、疋責料元件座落在其本地快取記憶 行普通快取記憶體查閱;:;點以網頁劇覽器應用軟體執 =本設:一内部的變數"位置"去指出參考Page 26 1223167 _a. "Repair" Case No. 89112368 V. Description of the invention (21) This element is used as the identifier of the data element to index the two entries in the data address table. Please note that it is more popular to use URLs to specify 7L special breeding materials in the Internet environment than to use the same method to specify a data element in a web page. Of course, 'this can be done with any suitable: identifiers instead. In addition, Table 1 uses "Snort", which is just for the sake of illustration. In fact, in the entry of the monitoring table, it is best to enter the node "snort" media access Control layer address (mac) / Internet Protocol (IP) address. The second element 'storage location' of 熟知 ni indicates the storage nodes (such as MAC address and IP address) that the monitoring node is familiar with, and its corresponding data elements, which are used by the same It is identified in the form's note item. The better choice is' monitoring node also keeps a count of the number of pieces in each group. The use and construction of the material address table will also be described in more detail below. Let's take two pictures, and now we will describe the processing flow according to the present invention. False Γ3: Click to display a web page. In the process of obtaining web page information, the print face is facing a command: in the web page, to retrieve Know the data elements identified by Γ. According to the present invention, the Liu Lan node asks for a specific Liu to extract the data elements. For example, the specific reference nodes that are taken by Cody = 1 & may be preset. For example, the reference macro I ^ may be-the designated Internet proxy server. For these referees, the #I f servo state can be set in most web browser applications such as Microsoft's "Internet Explorer (inter "net node" browser. Another option is to browse the node and the reference ..., j to be the same node. ____ ΗΐΙΒ 互 Ι page 27, description of the invention (22) Step S2 in the address is too round, refer to node execution-cache The fixed point of the memory determines the L / J catcat process), according to this process, the reference section interest: Π means: the one ?, the storage node is the designated storage sense I ::::,:; :: ;; -: Sub: The position of the point may fail to be set in two,; ,,,,,,). However, this also works: Send = Origin: When the data component is not accessed for the first time. In this case, a new point Go to appraisal and store the two parts, so you have n’t specified a monitor. Click “Save”. In step 33, if the reference is already identified, then you must refer to The storage reference node itself is identified as the storage Cf: SB4: to determine whether it is S8, the reference node determines whether it is' to the step material has expired, so it cannot be used while: ΪΓ line delete data W: memory error 'in step S10' reference section Ban Bo take note ::: S, go to step S11. If Κί body reading information Material components: such as; reference node = Department's cache memory and storage node to send the data back to the test: the second middle school does not perform 'in the second step': the situation after storage, then store, the second is ",. If s7 is in step S9, Yan does not perform a right-storage point, which is also a reference node, so in fact Newton Q2 = order a meaningful action. Shibuya S7 wants to access two: S4 Du reference node has decided to store The node is a reference node that is different from the copy and backup of the Beacon files (see 1223167, Emperor 89) 1 Tu Wu, the description of the invention (23) Λ said that the correction should not be: the material element. If so, try to be The identified storage nodes go to get the time / step S5, and the Lingkao node sends out a package message and everything. This can be done by referring to the reference node where the backup was made, and the required data element in the same path determines whether an error has occurred. It turns out that if the response message of the reference node package is sent to the reference node in step S6, it means that the storage node is sending out a requested data element, then in step ^ = Including; Act II 2 error occurred. Several backups have expired. If there are no errors: = the data element of the node provides the requested data element of the "then:" The execution is performed after step S6, at which point J is stored; the node of step 。. Therefore, the storage node is stored as one; 2. Another copy of the required data element is passed to it: ίΓ: ref =: The reference node detected an erroneous transmission: As described above, according to any of the following three reasons, Cheng Shi ’s armband will Reach step s11, that is, the execution of the formula =) In step S3, the reference node cannot identify the storage node of the data element. The test section ^ identified itself as a storage node, but in step S8 ==== memory fetch error, Therefore, in step sl10, it is required to delete the data element jc.) In step S6, the reference node identifies _ non-reference nodes as storage =. Any of the above-mentioned conditions show that:-a copy of the data element that is also @ 7 A backup is a copy of a 1 that does not exist in the virtual cache memory, that is, “No available”, non-expired data components. Page 29 1223167 Amendment ^ S_89U2368 V. Description of the invention (24) In any of its cache memories Body connection Becomes a node in the virtual cache memory). If so, the reference node must try to retrieve the data element from the Internet. Therefore, in step S11, the reference node sends a packet message via the Internet to the web server (a (Assessor using the URL of the data element) to request a copy of one of the data elements. At step §12, the test node decides whether an error occurred while trying to retrieve the data element via the Internet. An error like this, The reasons for this vary, and these reasons include: · errors in connecting the local area network to the Internet (routers or other devices), errors in web server equipment, etc. ^ detected at the reference node Such an error occurs, and then in step s3, the test node executes an invalid data writing process (abort data ordering, process). The invalid data writing process will be described in more detail later. As detailed The process of writing invalid data involves communication between the reference node and the monitoring node in order to inform the monitoring node. In the step, if the point is different from the reference node, then reply one, and do not accept the (not view is the handshake) n procedure. Assume that in step S12, you are trying to execute from the Internet reference node-write resource \ ^ Process), and if you need #, then step out & treat the incorrect message as a step. ^ Order the area cache 圮 愔 jealous- 丄 ,,; ^ The file is written to the reference node system 节点Decisiveness-Alas, goSof ts MedleyTM files can be stored in the support case No. 89112368 Description of the invention (25) Five means to modify the data of all common root data elements Someh.st.com data element ^ ( For example, everything is as mentioned above, when "storage section" is so efficient. Step S7 is executed. However, for example, when the information is given to the reference node, the point is forced to obtain a resource-like script from the Internet, then the reference section fails to address the storage node, or a copy of the file is copied; this is because it is backed up to the reference The reason for the node. Because the storage node cannot provide the data element (and the monitoring node, in any of these cases, the reference node designates itself as a data element and a data element is required) would be a meaningful action. Child clicks. Therefore, there is no step S7. Then look at the fourth figure. Here, in detail, step S20, the reference node determines? Addressing process of fetching memory. Store the required data components in the body. Take: You; Lang points to its local " cache memory (for example, the second description, the reference point will observe this image test section: the required data components. Another-a plan, if you click mr Liu Lan browser application software (that is, Liu Lan Festival S21, if you come to the second spray, execute it together. In the step body ΘH ~ that is, '', the data element is located in its local cache memory line Ordinary cache memory lookup: Click on the web browser application software to set = this setting: an internal variable " location " to indicate the reference

點回送這個值當:;:;:存節·點。在步驟S22 ’參考J 決定資料元件不是儲存在象:二置。否則’如果參考即 S23執行一監視二存址在二考郎點其本身内’則在步驟 process) (咖心 n〇de 1〇C咖 ^mii 圓圓國_| …II 顯議圓__ 第31頁 1223167 _案號891123防 年月日 佟π:__ 五、發明說明(26) locator process)。這些過程將在下文中作詳細描述。執 行這些過程之後,該内部變數”位置”被設定為去指出被指 定儲存資料元件的一個假設的儲存節點。在步驟S22,這 個值會被送回。 第五圖顯示監視節點定址過程的更詳細說明。在步驟 S30 ’參考節點決定是否其擁有資訊足以指出監視節點的 一致性,該監視節點是被指定去鑑定儲存著所要資料元件 的儲存節點之用的。來到處理後段,參考節點存取其監視 表(monitor table)以決定是否參考節點有一入口 (entry)、其係由所要資料元件的URL群組相同的url元件 所檢索而得來的。如果是這樣,在步驟331,參考節點設 立一内部的變數”監視”使等於索引表入口的監視位置元 件該索引表入口是在步驟S30以URL群組比對而鑑定的。 然後’參考節點執行步驟S32並送回”監視,,這個變|數的 ^如果在步驟S3〇,參考節點決定該被指定針對所要資 件的監視節點缺乏一致性,則到步驟S33執行廣播監 視即點位置查詢過程(broadcast monit()r locati⑽The point to return this value is:;:;: Save section · point. In step S22 ', it is determined by referring to J that the data element is not stored in the image: two sets. Otherwise 'If the reference is S23, a monitoring is performed and the second location is within the Ercolang point itself', then in the step process) (Caxin no 〇de 1〇C Coffee ^ mii 圆圆 国 _ |… II Shows a circle __ Page 31 1223167 _ Case No. 891123 Anti-Year Date: __ V. Description of the invention (26) locator process). These processes are described in detail below. After performing these processes, the internal variable "position" is set to indicate a hypothetical storage node for the designated storage data element. In step S22, this value is returned. The fifth figure shows a more detailed description of the monitoring node addressing process. In step S30 ', the reference node decides whether or not its own information is sufficient to indicate the consistency of the monitoring node, which is designated to identify the storage node storing the desired data element. In the later stage of processing, the reference node accesses its monitor table to determine whether the reference node has an entry, which is retrieved from url elements with the same URL group as the desired data element. If so, in step 331, the reference node sets an internal variable "monitoring" to make it equal to the monitoring position element of the index table entry, which is identified by URL group comparison in step S30. Then the 'reference node executes step S32 and sends it back' to monitor. This variable | If the reference node decides in step S30 that the designated monitoring node lacks consistency, it proceeds to step S33 to perform broadcast monitoring. Click location query process (broadcast monit () r locati⑽

Query process)。在此過程中(稍後將詳細描述此過 二的參m ϊ域網路發出一查詢訊息要求確認監視 Γ。監視節點是針對所要資料元件種L群 timp >卜#考:點啟動一内部計時器(internal 考節^占本ί β ^ 6又時間作計時。接下來,在步驟咖,參 $卽點決疋是否内部言+日聋哭 超過時眼。L里止1二 '、收集反應、對廣播查詢已經 1223167 修正Query process). In this process (the detailed description of these two parameters will be described later.) The local area network sends a query message asking for confirmation of monitoring. The monitoring node is for the desired data element type L group timp > bu # 考: Click to start an internal Timer (internal test section ^ account book β 6) and time counting. Next, in step C, refer to $ 卽 point to determine whether the internal language + day deafness cry more than the eyes. L 里 止 1 二 ', collection Response, query for broadcast has been corrected 1223167

_ 案號 89112368 、發明說明(27) 發出的查詢訊息,是否有接收到回應。如果尚未 參考節點的動作返回到步驟S 3 4。 到’ 如果有接收到回應,則在步驟S36、參考節點 些預疋的標準去決定是否接收到的回應、比先前 二些:虛,地說,參考節點可以使用任何標準去ϊ 取最子的監視即點。例如,預定的標準可以是每一 之資料元件之全部數量的計數、該每一群組乃被指定2 應的回應監視節點。亦即,如果從監視節點接收之一施 指出:此次比先前自同監視節點接收之資料元件的= 為大,則以最近接收之回應值來替代先前接收之回應 否,則’、最近接收之回應被捨棄。另一種選擇是,以,,' =作為#定標準;* 士子的監視節點{回應最 二亦即,在預定時間内第一個回應的監』 好的。在此情況下,一曰回雍妯拉l/r奋本— 敢 步驟S32至S36的迴路中離開口應被接收,參考郎點即可從 行牛H ^時牛益在步驟S34計時終止,參考節點立即執 ;::點:庙驟837,參考節點決定是否完全沒有任 'J 視卽點回應。士a θ、丄 點設定其内部變數"於視::^則到步驟⑽,參考節 (注意,▲步驟出傳遞保留回應的監視節點 職送回這個值。二參最者好,,監視節點)’並在步 最好的監視節點當作是表增加-入口,且以 element)。 疋皿視位置兀素(mon x tor 1 oca t i on 參考即點缺乏資訊以供鑑定所要的資料元件_ Case No. 89112368, Inventory Note (27), whether there is any response received. If the action of the reference node has not been returned to step S 3 4. Go to 'If a response is received, then in step S36, the reference node decides whether or not the response is received, compared to the previous two: virtual, in other words, the reference node can use any standard to select the most recent Monitoring is on the spot. For example, the predetermined criterion may be a count of the total number of each data element, and each group is designated as a response monitoring node. That is, if one of the receiving nodes from the monitoring node indicates that the value of this time is greater than that of the data element previously received from the same monitoring node, the response value received recently will be used to replace the previously received response. The response was discarded. Another option is to use, '= as the # setting standard; * Shizi's monitoring node {response is the second, that is, the first response within the predetermined time. In this case, I will return to Yonglila 1 / r Fenben — dare to leave in the loop of steps S32 to S36 should be accepted, you can refer to the Lang point from the line of cattle H ^ when Niu Yi timed out at step S34, The reference node immediately executes; :: point: miaoju 837, and the reference node decides whether there is no response at all.士 a θ, 丄 set its internal variables " Yu :: ^ then go to step ⑽, refer to the section (note that ▲ step out to send a reserved response to the monitoring node to send this value back. The best of the two parameters, monitor Node) 'and the best monitoring node in the step is regarded as a table addition-entry and with element). The apparent position of the dish (mon x tor 1 oca t i on reference point is lack of information for identification of the required data element

第33頁 1223167Page 1223167

之監視節點時,則參考節點在區域網路上發出一廣播或多 重播放訊息(broadcast or multicast message)|_ 求確認 此一監視節點的一致性。依據接收之回應,參考節點在其 監視表加上一入口,以使參考節點有此資訊可供將來使 用。在本特定實施例中,各參考節點在其監視表只加上一 口丄以回應單獨由參考節點送出的請求時使用。也就 Ξ:所:個:應訊息是送至參考節點#(發出請求訊息、 以對所要育料元件的監視節點作鑑定 應該訊息時修改監視表。然而在另即 監視itnt聽(ο,5—)回應訊息並修改他們的 ί該二L S二1 =人〇 ’即使其他參考節點現在並不想存 發心包人卢::::。處理來到後段’回應的監視節點會 ΐϊ:; ϊί 放位址之回應封包,以使所有的參 号即點可以接收此回應。 則Λ果參考節點判定接收回應皆告失敗, 二所ϊ貝枓兀件的儲存節點。已如上述,這是可能發生 :識的監視節點或還沒有監視節點被指定用 設定戋韌#於^卡牛之欠(因為自從處理過程最近重新 :i=rv所要的資料元件尚未被存取)。像這 監視ί::此/節點指定本身成為所要資料元件之 ::即:二參考/監視節點在其監視表產生一新的 為所要資料元件之監視節點彳瓜視即點本身作 件之通用次馮宏u〜 ·’、坆樣的入口包含所要資料元When monitoring a node, the reference node sends a broadcast or multicast message on the local area network. It is necessary to confirm the consistency of this monitoring node. According to the received response, the reference node adds an entry to its watch list so that the reference node has this information for future use. In this particular embodiment, each reference node is used when its monitoring table adds only one bite in response to a request sent by the reference node alone. That is to say: all: the response message is sent to the reference node # (a request message is sent to identify the monitoring node of the desired breeding element, the monitoring table is modified when the response message is sent. However, itnt listens to it (ο, 5 —) Respond to the message and modify them. The two LS two 1 = person 0 'even though other reference nodes do not want to save the pericardium now. Address response packet so that all parameters can receive this response. Then the reference node determines that receiving the response has failed, and the two storage nodes are the storage nodes. As already mentioned above, this may happen. : Identified monitoring nodes or no monitoring nodes have been designated with the setting 戋 ura # 于 ^ 卡 牛 之 欠 (because the data elements required by the: i = rv have not been accessed since the process was recently restarted). Like this monitoring :: This / node designation itself becomes the required data element :: that is, the two reference / monitoring nodes generate a new monitoring node for the desired data element in its monitoring table, which is a general order Feng Hong u ~ ', Like the entry Port contains the required data element

--一》-URL)群組當做通用資源定址器(URIJ 第34頁 1223167 曰--- "URL) group as a universal resource addresser (URIJ page 34 1223167

_案號891〗23fiS 五、發明說明(29) 1~ 群組7G素’且指示參考/監視節點作為監視節點。然後在 步驟S40參考/監視節點設立一内部變數”監視,,,以指出參 考/監視節點本身作為所要資料元件之監視節點;接著在 步驟S32,送回這個值。 第六圖在說明資料定址過程(data 1〇cat〇r prc)cess) 的二驟在步驟S 5 Ο ’參考節點決定是否監視節點(在監 視節點的定址過程作鑑定)即是參考節點本身。如果是, =步^1,參考節點(當作一潛在監視節點)決定是 有 > 汛去鑑定所要資料元件之儲存節點。到處理後 =,參考/監視節點決定是否其資料定址表有一入口、其_ Case No. 891〗 23fiS V. Description of the invention (29) 1 ~ Group 7G element 'and instructs the reference / monitoring node as the monitoring node. Then in step S40, the reference / monitoring node sets up an internal variable "monitoring" to indicate that the reference / monitoring node itself is used as the monitoring node for the desired data element; then, in step S32, this value is returned. The sixth figure illustrates the data addressing process (Data 1〇cat〇r prc) cess) The second step in step S 5 ′ 'The reference node decides whether to monitor the node (identified during the monitoring node's addressing process) is the reference node itself. If yes, = step ^ 1, The reference node (as a potential monitoring node) decides to have> storage nodes to identify the required data elements. After processing =, the reference / monitor node decides whether its data addressing table has an entry, its

=,通用=源定址器(URL)群組元素當作索引的,而此URL =兀素疋和所要貢料元件之通用資源定址器(um)相匹 變,則在步祕2,參考/監視節點設立内部的 5二定址矣等於儲存位置元素的-指示、而該元素是和 點送回,口相匹配的。然後在步驟s53,參考/監視節 郝迗回廷個位置值。 ^ 7 如果在步驟S51,參考節馱妯$次 元件之儲存節·點(即,資料定址缺讯去鑑定所要資料 址器⑽L)元素相匹配的入口疋止表缺乏一個和通用資源定 示是不正確的。發生此種情尤表示監視節點的指 ::點清除其資料定址表(重新;J因!;’包括:參考 點)1這樣,參考節點指;,或初始化參考 &視節點與儲存庫節點二者。2 作為所要資料元件之 監視/儲存節點會為 此杜則在步驟S54 ’參=, General = The source locator (URL) group element is used as an index, and this URL = Wu Susong and the universal resource locator (um) of the desired component change in phase, then in step 2, refer to / The monitoring node establishes an internal 5-2 address, which is equal to the -indication of the storage location element, and the element is sent back to the point, which matches the mouth. Then in step s53, the reference / monitoring section Hao Hui returns the position values. ^ 7 If in step S51, the reference node (the storage node and point of the secondary component) (that is, the data address is missing to identify the required data address device), the entry matching table lacks a general resource specification. incorrect. This happens especially if the monitoring node's finger :: click to clear its data addressing table (re; J due !; 'includes: reference point) 1 In this way, the reference node refers to; or initialize the reference & view node and repository node both. 2 The monitoring / storage node as the required data element will be

第35頁 料定址表入口。也就是說,參二-件,而產生-新的資 ------- ^ 孤視’儲存節點將所要資 修正 曰 —案號 89112368 五、發明說明(3〇) ___ ::::通用資源定址器⑽L),插入至新的 的通用育源定址器(URL)元素。 ^ ^止表 為本身、插入一識別器當作儲存位置考視/館存節點 儲存節點為對應到新資料定:考/監視/ ,總數的計數。然後在步;二群::以 ·、,,占故立一内部變數”位置” 視/儲存郎 驟S53送回這個冑。 去值『未找到』’且在步 假設現在在步驟35〇,參考 點。果如此,則在步驟S56 ^、、, ^ •疋 不是監視節 點過程(Sen“ata 1〇=r。达出資料定址查詢^ ^ 即被執行。在步驟S56,灸考Γ=:ΐ〇」〇ηΐΐ〇Γ Process) 訊息指向監視節點(由監視節出-封包㈣ 識別所要資料元件之儲存節要求 間之計時已到。如果未 /考即點決疋疋否該預定時 否已從監視節點收到回應J::c,參广節點決定是 S57。如果已接收到回應又再-人執订步驟 置"表示在返回回應時所=二,故立一内部變數"位 S53,送回該位置值。不的儲存節點。幻4在步驟 假設在步驟S57,參考節 止。這表示對於參考節點要求鑑冲日1裔之計時已經終 珣訊息,沒有接收到監韻 i疋:存節點的一致性之查 ^ A即黑占的回雁n丄 一 考郎點判定現在沒有任何節點 ^在這樣的狀況,參 節驟S60w:"元件的監視 __隱隱_ ---〜^考卽點指定自己 第36頁 1223167 1223167 _案號89ΠΜΜ 五、發明說明(31) 曰 修正 月 作為資料元件之監視節點。果如此, — 監視表插入一入口,声—心 、 >考/皿視節點於其 料元件的監視節點。該:二乂監人視節點乂身即是所要資 源定址器(URL)群组、其火已'要貧料元件之通用資 素,以及指明要把參考;、定址器咖群叙元 來,在步驟S54,參考/ κ ”、、皿視即點。接下 件的錯存節點。果如此點指定本身作為所要資料元 :為所要的資料元件,產=步=?參考/監視/儲存離 考/監視/儲存節點將所要新的貧料定址表入口。參 (URL),插入至*的資料定貝;' 兀件之通用資源定址器 (URL)元素。參考/監視/ 入口之通用資源定址器 別子當作儲存位置元素見/。錯更存郎點為本身著想,插入一識 應於新的資料定址表入口 >考/監硯/儲存節點為對 的計數。然後,在步驟S55,表u續增加資料元件總數 内部變數"位置"去表示一值/考/ i視/儲存節點設立一 回這個值。 找到』,且在步驟S53送 .第八圖是在說明無效資料宫A, wA;e process)。在步驟s7〇,參考data ”,即,將收到的、錯誤的考即除快取記憶體 步驟S71,參考節點決定、目^牛的賁料予以丟棄。 元件之監視節點。如果這樣,否力目半别也指定本身作為資料 =料定址表入口,資料定址 匕驟以3,參考節點刪除 :點’且遞減在參考節點可存:^考節點本身也是儲存 盗。計數器遞減是因為參考/ κ 育料元件數目的計數 失敗,而該計數器在第丄圄儲存節點儲存所要資料 _ 是持續遞 第37頁 1223167Page 35 Material Addressing Table Entry. That is to say, refer to two-pieces, and generate new capital --------- ^ Solitary 'storage node will amend the required capital—Case No. 89112368 5. Description of the invention (3〇) ___ :::: Universal resource locator (L), inserted into a new universal resource locator (URL) element. ^ ^ Stop list For itself, insert an identifier as the storage location for the examination / storage node. The storage node is set to correspond to the new data: examination / monitoring / and the total count. Then in the step; two groups :: ·· ,, occupy the internal variable "position" of the internal variable as the view / storage step S53 returns to this unit. Go to the value "not found" 'and at step Suppose now at step 35, the reference point. If so, in step S56 ^ ,,, ^ • 疋 is not a monitoring node process (Sen "ata 10 = r. Query data addressing ^ ^ is executed. In step S56, moxibustion test Γ =: ΐ〇" 〇ηΐΐ〇Γ Process) The message points to the monitoring node (from the monitoring node out-packet. The time between the storage node request for identifying the required data element has expired. If not, then click on it. If it is not scheduled, you have not downloaded from the monitoring node. After receiving the response J :: c, the reference node decides that it is S57. If the response has been received, then the -person ordering step is set to "=" when the response is returned, so an internal variable "bit S53 is set and sent Return the position value. Do not store the node. The magic step 4 is assumed in step S57, the reference section is terminated. This means that the reference node has requested that the timing of the first day of the red flag has been terminated, and no monitoring rhyme has been received: save Checking the consistency of nodes ^ A is Hei Zhan's Huiyan n 丄 One test point is determined that there are no nodes now ^ In such a situation, see step S60w: " Monitoring of components __ 隐隐 _ --- ~ ^ Test site designation page 36 1223167 1223167 _ case number 89ΠMM V. Invention Explanation (31) refers to the revised month as the monitoring node of the data element. If so, the monitoring table inserts an entry, the sound-heart, > test / disc node is the monitoring node of its material element. The node body is the required resource locator (URL) group, the general resources of the components that it needs to be lean, and the reference to the reference; the locator group, and in step S54, the reference / κ ", The point of view is the point. Then the wrong storage node of the next piece. If so, designate itself as the required data element: for the desired data element, the production = step =? Reference / monitoring / storage away from the test / monitoring / storage node will The new lean material address table entry is required. Refer to (URL), insert the data to *; 'Universal resource locator (URL) element of the element. Reference / monitoring / entry general resource locator as storage See the position element for error. For the sake of its own sake, insert a reference to the new data addressing table entry > test / monitoring / storage node pair count. Then, in step S55, the table u continues to add data. The total number of internal variables " position " The i-view / storage node sets this value once. It is found, and sent in step S53. The eighth figure is explaining the invalid data palace A, wA; e process). In step s70, refer to data ", that is, the data will be received. If the test result is incorrect, the cache memory is removed in step S71, the reference node is determined, and the target data is discarded. Component monitoring node. If this is the case, it is also necessary to designate itself as the data = material address table entry, the data addressing step is 3, the reference node is deleted: point ', and it can be stored at the reference node in descending order: ^ The test node itself is also a storage pirate. The counter is decremented because the counting of the number of reference / κ breeding elements failed, and the counter stores the required data in the second storage node. _ Yes Continuously page 37 1223167

增的。此種失敗是由於、例如從網際網路擷取資料元件 時,沒有節點可供使用。 在一替代方案,如果另一節點是監視節點,則監視節 點無效訊息過程(monitor invalidate message process) 在步驟S72被執行。在執行步驟S72或步驟S73之後,於步 驟S 7 5、參考節點停止該過程之執行。 注意,無效資料寫入過程(ab〇rt data writeIncrease. This failure is due to, for example, no nodes being available when retrieving data components from the Internet. In an alternative, if the other node is a monitoring node, a monitor invalidate message process is performed at step S72. After step S72 or step S73 is performed, at step S7 5, the reference node stops the execution of the process. Note that the process of writing invalid data (ab〇rt data write

Process)在步驟S13 (第三圖)執行,乃在下述幾種情況之 後·( a )參考節點定址儲存節點失敗,故被假設沒有節 點可當作資料元件之儲存節點;(b)參考節點指定本身 作為儲存節點,以儲存所想要資料元件,且試圖從網際 路獲得資料元件;& ( c )按照企望的偵測到錯誤。於 ^二步驟S72與S73之目的是:移除任何將參考節點作為儲 存卽點的指定。 第九圖是在說明刪除資料流程過程(delete data process)。在步驟S76,參考節點從其内部的快 憶體刪除所要資料元件的通用資源定址器(URL)。接下 5點在f Γ”7 ’參考郎點決定是否也指定本身作為儲存 :點。如果Μ ’則-些其他監視節點目前會將參考節點 好r η I 參考即點正在丟棄資料元件之複製 力疋第二圖之步驟8偵測到快取記憶體錯誤的結 果)m見節點無效訊息過程(mQnitc)i_ invaii )被執行。否則’參考節點必 (已丢棄^所要的貝枓兀件之複製備份Process) is executed in step S13 (third picture), after the following situations: (a) the reference node failed to address the storage node, so it is assumed that no node can be used as the storage node of the data element; (b) the reference node designation It itself acts as a storage node to store the desired data element and attempts to obtain the data element from the Internet; & (c) An error is detected as desired. The purpose of steps S72 and S73 in step two is to remove any designation that uses the reference node as a storage point. The ninth figure illustrates the delete data process. In step S76, the reference node deletes the universal resource locator (URL) of the desired data element from its internal flash memory. The next 5 points are at f Γ ”7 '. The reference point determines whether to also designate itself as a storage: point. If M ′ then some other monitoring nodes will currently refer to the reference node r η I reference point is discarding the copy of the data element If the result of the cache memory error is detected in step 8 of the second figure), the node invalid message process (mQnitc) i_invaii) is executed. Otherwise, the reference node must be discarded Copy backup

,號 89112368 五、發明說明(33) 點從其資料定址表刪除該一 视節點本身為所要資料 =^係用於識別參考/監 點於刪除(或覆蓋寫人)儲^點的。參考/監視節 減,正因為如此才會被指; = 也將其個數逐-遞 S79或步驟S78後,於步驟為m點。在執行完步驟 行。 1、參考節點停止該過程之執 第十圖是在詳細說明κ Μ location nn. 狐硯即點位置查詢過程(monitor :行二考:: 執仃參考即點在區域網路傳送包含一杳%1自夕私勺 要求識別監視節點的一致性,、二询妣心之封包, 別所要資料元件的儲存節點:^視卽點是被指定去鑑 源定址器(URL)訊息所桿示的—貝料疋件是由其通用資 所有節點所㈣(:二特:多重播放群組之 即由那些從參考節點接受二乂二=i視位置查詢過程 行。 又查5旬訊息封包的各監視節點來執 在步驟S82,監視節點從查詢訊息封包萃取 步二Ϊ ::料7^件的通用資源定址器⑽L)。在 到後段’冑視節點搜查其龄 子即點。處理來 、皿祝表,以才戈到一入η且古4S田 Π定t器(肌)群組元素、是和查詢訊息之封包、萃取 的通用資源定址器(URT、 b 了 尸/Γ早取No. 89112368 V. Description of the invention (33) The point is deleted from its data addressing table. The node itself is regarded as the required data = ^ is used to identify the reference / monitoring point to delete (or overwrite the writer) storage point. The reference / monitoring is reduced, which is why it is referred to; = The number is also transferred one by one-after S79 or step S78, the point is m points. After performing the steps OK. 1. The tenth diagram of the execution of the reference node to stop the process is detailed explanation of κ Μ location nn. The process of querying the location of the foxtail point (monitor: line 2 test ::: the point of reference point transmission in the local network includes one percent) 1 Since Xixi requires identification of the consistency of the monitoring node, and the second query of the packet of the heart, the storage node of the other required data element: The point of view is designated by the source addresser (URL) message— The data file is owned by all the nodes of its common resources (: 2 special: multi-play group, that is, those who receive 2 2 = i depending on the reference node from the reference node to query the position of the query process. Also check the monitoring of the message packet at 50 The node executes step S82, and the monitoring node extracts the second step of the query message packet from the query message packet :: the universal resource locator (L7) of the data. At the later stage, the despise node searches its age and clicks. The table shows the general resource locator (URT, b)

被指定用於所要的資料元件疋互/牛匹二:。如果監視節點 — --件,在步驟S84,監視節點產生 1223161 五、發明說明(34) 少 並傳迗回應封包給芩考節點,表明其一致性與資料元件的 總數之計數,而該等資料元件之監視節點可以識別通用資 源定址為(URL)群組之儲存節點。否則,監視節點僅單純 地停止該過程之執行。 第十一圖疋在S兄明資料定址查詢過程(data 1〇cat〇r quejry process)。此過程在第六圖之步驟s56被執行。參 ::L已Ϊ識別一監視節,點(假裝的)其被指定去鑑定 之儲存節點的一致性,發出-封包形式的查 節點。查詢訊息要求那些作為所要 時以盆、甬用-ΐ Ϊ :點的一致性。戶斤要的資料元件,在查詢 日守以其通用_貝源定址器(URL)來 - 節點從查詢訊息之封包 Ί在步糊5,&視 址器(URL)。在步驟‘卒的貧料凡件之通用資源定 件之儲存節點的-致性。此可如下達:3!:道_貝料疋 資料定址表之一入口、A 達^ · ▲視郎點搜查其 該肌元素是和查詢封由勺“素作索引找到的、而 如果監視節點妯4t 子匕所卒取的URL是互相匹配的 如,皿視即點被指定作為識別 I匹配的。 則在步驟S87,監視節點設立_ fU儲存節點, ,在匹配的資料定址表入口的儲存定址儲:庫·,以表示 存節點。 们储存疋址兀素所指明的儲 因此^視節點必定具有關於儲存節點不τ 因此&視即點必須 ::即點不正確資訊。 點。特別來說,如果於視二;^直到正確地識別餘存 為儲存節點。:步】定動 第40頁 1223167 五、發明説明(35) 一 S88,監視節點設立”儲存 者,在步驟S89,監相枚 木找到』。更有甚 table) 1〇Cat〇r 址哭(URL)元去你4皿視即點δ又疋新入口之通用資源定 資;:之通 置元素,以標器(URL)。監視節點設定儲存位 萃取得到之來源位址^點之用(依所接收之查詢訊息封包 r j 一 貧訊所標示者)。 -回應封:ίϊΓΓ有8以驟589後,監視節點產生並傳送 視節點停止執行=二存址庫查=參考節點。然後,監 -vaHd'lt'; Hpr!ee; Γ^ ^ ^ ^ 下兩種選擇狀況而來,盆C_eSS ρ本〃過程會被執行是由以 二:不能正確地識別儲存點皿視即點不是參考節點,其 料元件之複製備份。於盆一的腳=致於不能提供想要的資 存節點-致性的資訊、: J考節點缺乏任何儲 驟咖)。其結果,在假設參考Y5iVp^太(/十-圖之步 引扠卜監視節點在其資料定址# < 局储存即 圖之步驟S89 )。然而,者來 表開攻一入口 (第十一 ^要的資料元件之複製備V時考 監視節點,以指示監視節點考;傳送—訊息封包給 將參考節點識別為資料元件之::=:定址表内、該一 μ於其二的聊本,監視節點:::=r。 況(第十一圖之步驟S87 ),但在’嘗去:即點-致性的資 旨5式去接觸該錯存節點 1223167Designated for the required data element: Mutual / Bull II :. If the monitoring node is a piece of software, in step S84, the monitoring node generates 1223161. V. Description of the invention (34) The response packet is transmitted to the test node, indicating its consistency and the total number of data elements. The monitoring node of the component can identify the storage node whose general resource address is the (URL) group. Otherwise, the monitoring node simply stops the execution of the process. The eleventh figure is the data query process (data 10cat quejry process) in S Brother Ming data. This process is performed in step s56 of the sixth figure. Reference :: L has identified a monitoring node, points (pretends) the consistency of the storage node that it is designated to authenticate, and sends out a check-node in the form of a packet. The query message requires consistency when used as needed, using -ΐ Ϊ: points. The data components required by the user are in the query. Rishou uses its general-purpose source addresser (URL)-the packet from the query message by the node Ί in step 5, & addresser (URL). Consistency of the storage node of the universal resource component at the step 卒 of the poor material. This can be reached as follows: 3 !: Dao_Beijing, one of the entries in the data addressing table, A reaches ^ · ▲ depending on the search, the muscle element is found by querying the "primary index", and if the monitoring node The URLs extracted by 取 4t sub-daggers are matched with each other. For example, the point of view is designated as the match of identification I. Then, in step S87, the monitoring node sets up a _fU storage node, and stores it at the entry of the matching data addressing table. Address storage: library, to indicate the storage node. We store the storage indicated by the address element. Therefore, the view node must have no information about the storage node. Therefore, the view must be: the point is incorrect information. Point. Special For example, if you look at the second; ^ until the remaining storage nodes are correctly identified .: Step] Fixed page 1223167 on page 40 V. Description of the invention (35) a S88, monitoring node establishment "storage, in step S89, monitoring I found it. What's more, the table) 10 cat URL URL (general resource) is used to determine the general resource of the new entry point δ and the new entry; the general element of the element is marked with a URL. Set the storage location of the monitoring node for the extraction of the source address ^ point (according to the received query message packet r j-marked by poor information). -Response seal: After ϊϊΓΓ has 8 to 589, the monitoring node generates and transmits. The visual node stops execution = two storage address library check = reference node. Then, monitor-vaHd'lt '; Hpr! Ee; Γ ^ ^ ^ ^ ^ The following two choices come from the basin C_eSS ρ This process will be performed because of two reasons: the storage point cannot be correctly identified. It is not a reference node, but a copy of its material components. Yu Yiyi's foot = caused the failure to provide the desired storage node-consistent information: J test node lacks any storage coffee). As a result, it is assumed that the reference Y5iVp is too high (/ step of the ten-graph). The monitoring node is located at its data address # < to store the map in step S89). However, the following table introduces an entry (the eleventh ^ required data element preparation of the V test test node to instruct the monitoring node to test; send-a message packet to the reference node identified as a data element :: =: In the addressing table, the first and second chats, monitor the node ::: = r. The situation (step S87 in the eleventh figure), but in the 'taste it: point-to-response, the 5th formula goes Contact this misplaced node 1223167

MM89Π2368 五、發明說明(36) 時(第三圖之步驟S5),參考節點偵測到鋩 圖之步驟S6)。參考節點指定本身為儲存第三 網路獲得想要的資料元件(第三圖之步驟8 . 且攸網際 誤發生(第三圖之步驟812)。在此後者狀1 =沒f錯 傳送-訊息封包給監視節點,以指示監視節點去麥考 料定址表、使能將參考節點識別為資料元 j /、一貧 (第三圖之步驟S16 )。 =存節點 在步驟S 9 1,監視節點從接收訊息封包萃取出次 件之通用資源定址器(URL)。在步驟S92,於满##貝心料元 出的貧料元件之通用資源定址器(URL),和i次社a卒取 之任何入口的通用資源定址器(URL)元素作比對。如果= 匹配成功,監視節點停止執行本監視節點無效訊息過程… (monitor in val i date message process)。否則 了在 + S 9 3 ’監視節點決定是否該訊息封包含有一刪除指入。’如^ 果是,則在步驟S95,監視節點將其資料定址表内該一匹 配成功之入口,予以刪除。監視節點也將其所服務的通用 資源定址器(URL)群組之總個數,作” 1 ”的遞減。另一方 面’如果在步驟S 9 3 ’監視節點決定該訊息不包含刪除指 令,則在步驟S 9 4,監視節點以訊息封包來源是參考節點 的一個指標,去對被擷取表格之入口的儲存位置元素,作 覆蓋寫入。 第七圖是根據本發明之另一實施例來對第一圖的企業 網路3作說明,其中企業網路(例如,一區域網路)3被區 分成各含一些節點的幾個分離、不重疊池(pools)或子集 合(subsets)。如圖中所示,區域網路3包含子集合a、B與MM89Π2368 5. In the description of the invention (36) (step S5 of the third graph), the reference node detects step S6 of the 铓 graph). The reference node designates itself to store the third network to obtain the desired data element (step 8 in the third picture. And the Internet error occurs (step 812 in the third picture). Here the latter is like 1 = no error sent-message The packet is sent to the monitoring node to instruct the monitoring node to go to the wheat addressing table and enable the reference node to be identified as the data element j /, a poor (step S16 in the third figure). = The storage node is in step S 91, the monitoring node The universal resource locator (URL) of the secondary component is extracted from the received message packet. In step S92, the universal resource locator (URL) of the lean component output by Yu Man ## 贝 心 料 元, and the second time Compare the Universal Resource Locator (URL) element of any entry. If = matches successfully, the monitoring node stops executing the monitoring node invalid message process ... (monitor in val i date message process). Otherwise, it is at + S 9 3 ' The monitoring node decides whether the message envelope contains a delete entry. 'If YES, then in step S95, the monitoring node deletes the matching entry in the data addressing table and deletes it. The monitoring node also deletes the information it serves. General Resources The total number of URL (URL) groups is decremented by "1". On the other hand, if the monitoring node decides that the message does not contain a delete instruction in step S 9 3, the monitoring node sends a message in step S 9 4 The source of the packet is an indicator of the reference node to write over the storage location element at the entry of the retrieved form. The seventh figure is a diagram of the enterprise network 3 of the first figure according to another embodiment of the present invention. Explanation, in which the enterprise network (for example, a local area network) 3 is divided into several separate, non-overlapping pools or subsets each containing some nodes. As shown in the figure, the local area network 3 Contains sub-sets a, B and

第42頁 1223167Chapter 1223167

案號 89112368 五、發明說明(37) C。子集合 A 包含節點 n4,n6,n8,η18,π20,n22,n24, η26,η28,η30 及 η30。子集合 β 包含節點]116,η34,η3β, η38,ηΜ,η42 及 η44。子集合C 包含節點[11〇,η12&η14。 對每一子集合A、Β與C,各有一虛擬快取記憶體被形 成,即,只有相同子集合之節點之快取記憶體,才能被鏈 結在一起。此種分區與分離鏈結而形成的虛擬快取記憶 體,可用好多方式達成之。然而,其中一個方法是提供每 個子集合一個單一的傳輸控制協定(Tcp )埠(未圖 示)。每個節點只以其所屬子集合的TCP埠號碼,來接受 與處理(在TCP層之上)封包。否則,本發明運作類似於 上述之一子集合内,而以子集合為基礎。此種分池 (poo 11 ng)技術提供一附加優點,亦即將區域網路3之快取 記憶體之負載,由該區域網路3之節點,來作分割。 接著要對上面描述的本發明,舉出一些範例。考慮在 某一時點,有一資料元件從未被區域網路3之節點n4至以斗 之任一節點所存取。假設一瀏覽器節點n8希望去擷取一網 頁’而這又涉及必須存取該資料元件。又假設瀏覽器節點 η 8是參考節點。參考節點n 8執行快取記憶體定址過程 (cache locator process)(第三圖之步驟S2)。參考節 點n8決定資料元件之通用資源定址器(URL)不是儲存在節 點n8之快取記憶體(第四圖之步驟820 )。參考節點n8執 行監視節點定址過程(monitor node locator process) (第四圖之步驟S23 )。參考節點n8決定其監視表缺乏資 料元件之入口 (第五圖之步驟S30 )。參考節點n8執行廣 播監視位置位置查詢(broadcast monitor locationCase No. 89112368 V. Description of Invention (37) C. Subset A contains nodes n4, n6, n8, η18, π20, n22, n24, η26, η28, η30, and η30. Subset β contains nodes] 116, η34, η3β, η38, ηM, η42, and η44. Subset C contains nodes [11〇, η12 & η14. For each sub-set A, B, and C, a virtual cache memory is formed, that is, only the cache memories of the nodes of the same sub-set can be linked together. The virtual cache memory formed by this partitioning and separation link can be achieved in many ways. However, one method is to provide a single Transmission Control Protocol (Tcp) port (not shown) for each subset. Each node only accepts and processes (above the TCP layer) packets based on the TCP port number of its own subset. Otherwise, the present invention operates similarly to one of the sub-sets described above, but on a sub-set basis. This pooling (poo 11 ng) technology provides an additional advantage, that is, the load of the cache memory of the local network 3 is divided by the nodes of the local network 3. Next, the present invention described above is exemplified. It is considered that at a certain point in time, a data element is never accessed by any node n4 to any node of LAN 3. Suppose a browser node n8 wishes to retrieve a web page ', which in turn involves having to access the data element. It is also assumed that the browser node η 8 is a reference node. The reference node n 8 executes a cache locator process (step S2 in the third figure). The reference node n8 determines whether the universal resource locator (URL) of the data element is not stored in the cache memory of the node n8 (step 820 in the fourth figure). The reference node n8 executes a monitor node locator process (step S23 in the fourth figure). The reference node n8 determines the entry of its monitoring table lacking data elements (step S30 in the fifth figure). Reference node n8 performs broadcast monitor location query

第43頁 1223167 修正 案號 89112368 五、發明說明(38)Page 43 1223167 Amendment No. 89112368 V. Description of Invention (38)

Query) j第五圖之步驟S33 )。作用像監視節點的各節點 (即··節點n6,nlO至n44 )接收訊息,並萃取出邦、要 料元件之通用資源定址器(URL)群組(第十圖之步^驟 、 S82 ),但判定那不是資料元件之監視節點(第十圖之步 驟S83)。參考節點〇8等待回應(第五圖之步驟s35) 到計時器計時終止(第五圖之步驟S34),但沒有接收到 = f (第五圖之步驟537)。於是’參考節點n8於其 ς =插:一入口,以指明節點n8是想要的資 糊),且送回一指示:本身、即節點n8, =二監視郎點(第五圖之步驟S32,步驟S40)。接下 ΐίυ料ΐ址過程時’參考/監視節點n8鑑定本身作為 y即。第六圖之步驟S50),並判定:對想要的資料 /、貝0枓定址表内並沒有對應入口 (第六圖之步驟 料元件二是,參考/監視節點118於其m址表,為該資 節點;:::…以指明節點n8本身是資料元件之儲存 後二夂ί 貧料計數增加1 (第六圖之步驟S54)。秋 步驟“、ίίΠΜέη8送回"未找到"位置(第六圖之 顽/啟六& 第四圖之步驟S22)。由於參考/監 =儲存庫節點n8送回”未找到"位置(第三圖之步; 記憶體二T第嘗試:網際網路獲得資料元件且料^ 有"監“圖,步驟su ’步驟S12,步驟S15)。沒 王兄即.,、、占無效訊息”需要被送出。 資料ϊ:假:ii節點ni〇想要存取儲存在節點n8之相同 可避❖奋》、1考即點η1〇也搜尋資料元件之監視節點,無 --達步j^S33、於此參考節點η10發出廣播訊息要 1223167 曰 案號 89112368 五、發明說明(39) ^ 求識別資料元件之監視節點的一致性。然而,這次執行步 驟S83 (第十圖)時,監視節點η8匈定事實上它確有一入 口供參考節點nl 0獲取想要的資料元件。於是,監視節點 n8傳送回應訊息指出:資料元件已有監視節點(第十圖之 步驟S84 )。參考節點nl0收到此回應(第五圖之步驟 S35 )且予以保留(步驟S36 )。當計時器計時終止 圖之步驟S34),參考節點nl〇判定其實際已收到回應(第 五圖之步驟S37),且回送監視節點的識別子、*認定 監視節點n8 (第五圖之步驟S38,步驟S32 )。夂 ^ t # , D ^ ^ # ^ 視即點n8即是貧料元件之監視節點。 ▲土— 判定監視節點η8並非其本身(第六,參考郎點1110 % ό,人… 升,、不牙、弟圖之步驟S50 )並送出 二=;;n8以識別儲存節點(步刪)。監視 7:二!二楚並從中萃取想要的資料元件之通用資 ==)(弟十一圖之步驟385),識別想要的資料 疋址表的人口(第十—圖之步驟S86),並送 回儲存卽點資訊於其中(第十一圖之步驟如,步驟 _!九、在此時儲存節點即節點n8。在計時器計時終止以 圖i步驟S57),參考節點nl°從監視節點n8接收 -:二二丨乂驟以8 )。參考節點111 〇回送資料元件之儲存 Γ=別:息,亦即,指明那是儲存節㈣(第六圖之 已、,二找到儲存節點(第三圖之步驟以 點nlO本身(坌r固+本璁以、^ ^ 不疋參考即 訊自要"Λ第 步),參考節點n10發出-要求 I1H I麵 ^ 節點n8獲得#料元件(第三圖之步驟以Query) j Step S33 of the fifth graph). Each node that acts like a monitoring node (ie, nodes n6, nl0 to n44) receives messages and extracts a general resource addresser (URL) group of states and required components (step ^ of the tenth figure, S82) , But it is determined that it is not the monitoring node of the data element (step S83 of the tenth figure). The reference node 08 waits for a response (step s35 in the fifth figure) until the timer expires (step S34 in the fifth figure), but does not receive = f (step 537 in the fifth figure). Then 'reference node n8 is inserted at its ς = an entry to indicate that node n8 is the desired paste), and an instruction is returned: itself, that is, node n8, = two monitoring lang points (step S32 of the fifth figure) (Step S40). In the following process, the reference / monitoring node n8 identifies itself as y. Step S50 of the sixth figure), and determine that there is no corresponding entry in the desired data / location table (the second component of the sixth figure is the reference / monitoring node 118 in its m-address table, For the resource node; ::: To indicate that the node n8 itself is the data element, and the lean material count is increased by 1 (step S54 in the sixth figure). The autumn step ", ίΠΜέη8 is returned " not found " Position (the sixth picture of the stubborn / open six & the fourth picture of step S22). Since the reference / monitor = repository node n8 sent back "the position was not found" (the third picture of the step; the memory two Tth attempt : The Internet obtains the data element and it is expected that there is a "map, step su 'step S12, step S15). Without Wang Xiong, ... ,, account for invalid information" need to be sent. Data ϊ: Fake: node ii ni〇 want to access the same avoidable stored in node n8 fen ", 1 test point η1〇 also search for the monitoring node of the data element, none-up to step j ^ S33, send a broadcast message at reference node η10 here 1223167 Case number 89112368 V. Description of the invention (39) ^ One of the monitoring nodes for identifying data elements However, this time when step S83 (tenth picture) is performed, the monitoring node η8 is determined that it does have an entry for the reference node nl 0 to obtain the desired data element. Therefore, the monitoring node n8 sends a response message indicating that the data element There is already a monitoring node (step S84 in the tenth graph). The reference node nl0 receives this response (step S35 in the fifth graph) and reserves it (step S36). When the timer expires, step S34 in the graph), the reference node nl 〇 judge that it has actually received a response (step S37 in the fifth figure), and return the identifier of the monitoring node, * identifying the monitoring node n8 (step S38, step S32 in the fifth figure).夂 ^ t #, D ^ ^ # ^ View point n8 is the monitoring node of the lean component. ▲ Tu—determine that the monitoring node η8 is not itself (sixth, with reference to 1110% of the Lang point, human ... Sheng, Buya, Ditu step S50) and send two = ;; n8 to identify the storage node (step delete) . Surveillance 7: Two! Two Chu and extract the general data of the desired data element ==) (step 385 of the eleventh figure), identify the population of the desired data address table (tenth-step S86 of the figure) And return the storage point information to it (the steps in the eleventh figure are, for example, step _! 9. At this time, the node is stored as node n8. When the timer expires, step i in step i in Figure i), refer to node nl ° from The monitoring node n8 receives-: 22 two steps (8). The reference node 111 〇 returns the storage of the data element Γ = other: interest, that is, indicates that it is a storage node (the sixth figure has been found, and the second is to find the storage node (the step in the third figure is to point nlO itself (本身 rsolid + This is based on, ^ ^ No reference is required (refer to step Λ), the reference node n10 is issued-request I1H I plane ^ node n8 to obtain # material components (the step in the third figure is based on

1223167 j 號 89112狀81223167 j 89112 shape 8

五、發明說明(40) __ 假設沒有錯誤發生(第三圖之步驟86) ’ 到參考節點Π10 (第三圖之步驟”)。然而、,枓凡件被送回 生’:如、儲存節點η8判定資料元件已過期如果錯誤發 u第一圖之廣域網路5、網站 =:點 η2,重新獲取(refetches)該資料元件,=7之飼服器 n2有新版本可以使用,其結果是將資料元在伺服器 點n8。但如果,伺服器心指出儲節點給芩考節 有效的,則儲存節點n8回送資料二^ ,有的版本依然 而不需要從伺服器n2重新攫取資料子元版1 回:=;;=。件之複製 現在假設參考節點nl2想要存取相同的資 :J與上述很類似的步驟順序會被執行。最重/則 貝料兀件提供給參考節點nl2的不是節點η1(), θ疋,以 點η8。簡言之,即使儲存筋 J 疋儲存節 nl。,節點η8仍保留當作是儲二二供貧料元件給參考節點 :㈣元件寫入蓋過’。執行的步以複;==另-=,/執行步㈣時,儲存節點峨ίϊΐ: Γ 當執行步驟86時Λ 之複製備份。於是, 尸夂去松, 參考郎點η1 4偵測到錯誤發生。這合蚀 于多考郎點η 1 4從網際網路獲取資一 印 步驟S12,步驟Sly *獲取/科(第二圖之步驟S!1, 100。假1 4 # I \ u並且將貢料儲存在快取記憶體V. Description of the invention (40) __ Assume that no error occurs (step 86 in the third graph) 'to the reference node Π10 (step in the third graph'). However, the normal pieces are returned to life ': eg, the storage node η8 Judging that the data element has expired. If the WAN of the first picture is sent by mistake, the website =: point η2, and the data element is re-refetched. A new version of the feeder n2 of = 7 is available. The result is The node is at the server point n8. However, if the server indicates that the storage node is valid for the test section, the storage node n8 returns data 2 ^, and some versions still do not need to re-retrieve data from the server n2. Sub-element version 1 Back: = ;; =. Copy of the pieces now assume that the reference node nl2 wants to access the same data: J The sequence of steps very similar to the above will be performed. The heaviest / then the pieces of material provided to the reference node nl2 are not Node η1 (), θ 疋, with point η8. In short, even if the rib J 疋 is stored at node nl, node η8 is still reserved as a storage element for the lean feed element to the reference node: 写入 element is written over '. Steps executed repeatedly; == Another-=, / When executing steps Storage node Eίϊΐ: Γ The copy backup of Λ is performed when step 86 is performed. Therefore, the corpse is loosened, and an error has been detected with reference to the Lang point η1 4. This is corroded at the Dokalang point η 1 4 obtained from the Internet First print step S12, step Sly * get / section (step S! 1, 100 of the second picture. Leave 1 4 # I \ u and store the tribute in the cache memory

時,沒有錯誤發生。參考節點 ΰΒ' __________ !223167No errors occurred. Reference node ΰΒ '__________! 223167

五、發明說明(41) 必須發出監視節點無效訊息(步驟Sl6 )至龄點 n ,因而監視節點n8不再判定節點n8當作儲二反 ,二監視節點〇8會識別儲存節點nl4當作此資卩元〃件之儲 。:是:參考/儲存節點η"發出—替代訊息給監視 :..^8。監視節點以從訊息中萃取出資料元件之⑽匕(第 二圖之步驟S91),從其資料定址表獲得匹配表的入口 ^十二圖之步驟S92),並判定沒有接收到—刪除訊息 二圖之步驟S93)。像這樣,監視節點⑽取代入口 :存位置元素、以指出儲存節點n丨4 (第十二圖之步驟 S 9 4 )。 ” 一現在假設節點nl 〇目前缺乏資料元件,當想要再次存 取^料元件。類似前述的步驟被執行。然而一次,參 考f點nl 0判定其監視表有一入口可識別此資料元件之監 ,即點(第五圖之步驟S30 ),名為監視節點Μ。因此, 參考節點1)10回送節點以當作資料元件之監視節點(第五 圖之步驟S31,步驟S32)。資料元件之摘取則是和以前相 類似(注意:現在監視節點n8識別儲存節點nl4當作資料 凡件之儲存節點,所以資料元件是從儲存節點nl 4 取)。 現在假設節點nl 〇希望去存取相同的資料元件,但是 節點n8,前無法使用(例如:故障,忙碌',未安裝,一或沒 有電源等)。因此,當參考節點nl 〇送出一資料定址查詢 給監視節點π8時(第六圖之步驟S56),在計時器終止以 前(第六圖之步驟S57),沒有回應被接收(第六圖之步 驟358 ) °如此’藉由以下的動作,參考節點η 1 0會指定本V. Explanation of the invention (41) The monitoring node invalid message (step S16) must be sent to the age point n, so the monitoring node n8 no longer determines that the node n8 is regarded as the second storage, and the second monitoring node 08 will identify the storage node nl4 as this. Storage of assets and documents. : Yes: Reference / storage node η " issues-substitute message for monitoring: .. ^ 8. The monitoring node extracts the data element from the message (step S91 in the second figure), obtains the entry of the matching table from its data addressing table ^ step S92 in the twelve figure), and determines that it has not been received—delete message two Step S93 in the figure). As such, the monitoring node ⑽ replaces the entry: a storage location element to indicate the storage node n 丨 4 (step S 9 4 in the twelfth figure). Now assume that node nl0 currently lacks a data element. When you want to access the data element again. Similar steps are performed. However, once, reference point nl 0 determines that its monitoring table has an entry that can identify the data element. That is, the point (step S30 in the fifth diagram) is called the monitoring node M. Therefore, the reference node 1) 10 returns the node as a monitoring node of the data element (step S31, step S32 in the fifth diagram). The extraction is similar to the previous one (note: the monitoring node n8 now recognizes the storage node nl4 as the storage node of the data, so the data element is taken from the storage node nl 4). Now suppose the node nl 〇 wants to access the same Data element, but node n8, cannot be used before (for example: fault, busy ', not installed, one or no power, etc.). Therefore, when the reference node nl0 sends a data address query to the monitoring node π8 (sixth figure (Step S56), before the timer expires (step S57 in the sixth figure), no response is received (step 358 in the sixth figure) ° So 'By following actions, refer to section η 1 0 specifies how the

第47頁 1223167Chapter 1223167

身作為資料元件之監視節點與儲存節點:(1)修改立監 視表^口、以標示參考/監視/儲存節點nlG是資料元;;的 監視郎點(第六圖之步驟S60),與(2)在其資料定址表 加上一入口、以標示參考/監視/儲存節點〇1〇是此資料元 件存節點,並對於把參考/監視/儲存節sni〇當作是 監視即點的全部資料元件之總個數,予以增加丄(第六圖 之γ驟S 6 0 )。參考節點n 1 〇回送該位置"未找到"(步驟 S 5 5 v驟S 5 3 ),因而使得參考節點n 1 〇從網際網路獲得 資料凡件(第三圖之步驟S3,步驟su ’步驟S12,步驟 S5 )、。注意,『沒有節點無效』會被執行(第三圖之步驟 S16),因參考節點nl0並不知道有一監視節點曾經存在 過0 現在假設節點n8再次可被使用。如此,則資料元件有 二個監視節點,即節點⑽與節點nl〇 (也請注意資料元件 有二個儲存節點,即節點nl4與節點nl0)。假設現在節點 想要去存取資料元件。參考節點nl6在其監視表缺乏資 汛以識別資料元件之監視節點。像這樣,節點nl6發出一 廣播訊息要求識別監視節點之一致性(第五圖之步驟 S33)。這一次,參考節點nl6接收到一個以上的回應(第 =圖之步驟S35 ),亦即、從監視節點n8的一個回應'與 從監視節點nl〇的一個回應。參考節點η16使用預定〜之標、 準’由他門的回應要求從η 8與η 1 0中選取最佳之監視節點 (第五圖之步驟S36 )。存取有關的其餘部分,和上述相 類似。Monitoring nodes and storage nodes as data elements: (1) modify the monitoring table to indicate that the reference / monitoring / storage node nlG is a data element; the monitoring point (step S60 in the sixth figure), and ( 2) Add an entry to its data addressing table to indicate the reference / monitoring / storage node. 010 is the storage node of this data component, and for the reference / monitoring / storage section sni〇 as all the data of the monitoring point Increase the total number of components 丄 (γstep S 6 0 in the sixth figure). The reference node n 1 〇 returns the position " not found " (step S 5 5 v step S 5 3), so that the reference node n 1 〇 obtains the data item from the Internet (step S3, step 3 of the third figure) su 'step S12, step S5) ,. Note that "no node is invalid" will be executed (step S16 in the third figure), because the reference node nl0 does not know that a monitoring node ever existed 0. Now suppose node n8 can be used again. In this way, the data element has two monitoring nodes, namely node ⑽ and node nl0 (also note that the data element has two storage nodes, namely node nl4 and node nl0). Suppose now that the node wants to access data components. The reference node n16 lacks resources in its monitoring table to identify the monitoring node of the data element. As such, the node n16 sends a broadcast message requesting to identify the consistency of the monitoring node (step S33 in the fifth figure). This time, the reference node n16 receives more than one response (step S35 in the figure), that is, a response from the monitoring node n8 and a response from the monitoring node n10. The reference node η16 uses a predetermined standard, and the response of the other gate requires that the best monitoring node be selected from η 8 and η 10 (step S36 in the fifth figure). The rest of the access is similar to the above.

第48頁Page 48

第49頁Page 49

1223167 _案號89112368_年月日 倐正_ 圖式簡單說明 第一圖顯示本發明所使用的一網路。 第二圖顯示根據本發明實施例之一節點。 第三圖顯示根據本發明實施例之一處理流程圖。 第四圖顯示根據本發明實施例之快取記憶體定址過程 (cache locator process)之流程圖 〇 第五圖顯示根據本發明實施例之監視節點定址過程 (monitor node locator process)之流程圖。 第六圖顯示根據本發明實施例之資料定址過程(data locator process)之流程圖 ° 第七圖顯示第一圖的網路之一部分,其中根據本發明實施 例網路部分之複數節點被區分成數個節點子集合。 第八圖顯示根據本發明實施例之無效資料寫入過程(ab〇rt data write process)之流程圖。 第九圖顯示根據本發明實施例之刪除資料過程(d e 1 e t e data process)之流程圖。 第十圖顯示根據本發明實施例之監視節點位置查詢過程 (monitor location query process)之流程圖 〇 第十一圖顯示根據本發明實施例之資料定址查詢過程 (data locator query process)之流程圖 〇 第十二圖顯示根據本發明實施例之監視節點無效訊息過程 (monitor invalidate message process)之流程圖 〇 元件符號說明 1 網路1223167 _Case No. 89112368_ Year Month and Day _ Zheng Brief Description of the Drawings The first figure shows a network used in the present invention. The second figure shows a node according to an embodiment of the invention. The third figure shows a processing flowchart according to one embodiment of the present invention. The fourth figure shows a flowchart of the cache locator process according to the embodiment of the present invention. The fifth figure shows the flowchart of the monitor node locator process according to the embodiment of the present invention. The sixth diagram shows a flowchart of the data locator process according to the embodiment of the present invention. The seventh diagram shows a part of the network of the first diagram, in which the plural nodes of the network part are divided into numbers according to the embodiment of the present invention. Node subsets. FIG. 8 is a flowchart of an abort data write process according to an embodiment of the present invention. The ninth figure shows a flowchart of a data deletion process (d e 1 e t e data process) according to an embodiment of the present invention. The tenth figure shows a flowchart of a monitor location query process according to an embodiment of the present invention. The eleventh figure shows a flowchart of a data locator query process according to an embodiment of the present invention. The twelfth figure shows a flowchart of a monitor invalidate message process according to an embodiment of the present invention. Symbol description of components 1 Network

第50頁 1223167 _案號89112368_年月日 修正 圖式簡單說明 3 企業網路 5 網際網路 7 網站伺服器網路 10 中央處理器 12 單一匯流排 14 輸出裝置 16 輸入裝置 18 網路介面卡 22 高速記憶體 24 主記憶體 26 磁碟記憶體 100 快取記憶體 A、B、C 子集合 r 2、r 4 路由器 n8 瀏覽節點 n2 、 n4 ' n6 、n8 、nl0 、nl2 、nl4 、nl6 、nl8 、n20 、 n22、n24、 n26、n28、n30、n32、n34、n36、n38、n40、 n42、n44 節點 h2、h4、h6 集線器 ,Page 50 1223167 _Case No. 89112368_ Simple Modified Illustration 3 Enterprise Network 5 Internet 7 Web Server Network 10 Central Processing Unit 12 Single Bus 14 Output Device 16 Input Device 18 Network Interface Card 22 High-speed memory 24 Main memory 26 Disk memory 100 Cache memory A, B, C Sub-collection r 2, r 4 Router n8 Browse nodes n2, n4 'n6, n8, nl0, nl2, nl4, nl6, nl8, n20, n22, n24, n26, n28, n30, n32, n34, n36, n38, n40, n42, n44 nodes h2, h4, h6 hubs,

第51頁Page 51

Claims (1)

1223167 Λ_ 修正 曰 連1 891123R8 六、申請專利範圍 1、一種將來自伺服器節點 記憶體作最適化之方土 τ供使用貝枓70件之區域快取 2週化之方法,包括下列步驟: 私 π】^連接至區域網路之複數節點的每一節點^ 一 £域性貫體存在之快取記憶體; 卩點,提供 此複數節點之至少某些節點之各快取記情h 接起來而成為單一虛擬快取記憶體;以及 己隐體連 C扣定此等節點之特定一個^ ^ ^ 於恆久地儲存一特定眘钮-从 F々减评即點以用 之禝I備份給複數節點的一 4 7L件 資料元件的其他參考節點;,、缺之仁-要存取該特定 其中指定特定節點作為儲存節點, 資料元件之複製備份提供給此種想要存取特特^ 一個參考節點而改變。 /V+疋件的 2、如申請專利範圍第丨項之方法,再包括下列步 U)如果即將被一個參考節點存取的資料元件沒有 效的複製備份存在於虛擬快取記憶體,則選擇地從伺服 節點獲得該資料元件的複製備份,其中以一個來 ^ 虛擬快取記憶體存取資料元件的複製備份,比:砂# 點存取資料元件的複製備份花費較少的時間。 胀裔即 3、根據申請專利範圍第1項之方法,再句虹_ , 丹a括下列步驟 : (d ) —第一參考節點從儲存節點驊復 ...^ μ *制#於· 于即點獲仵—特定資料元件 的禝製備份,1223167 Λ_ Modified Yuelian 1 891123R8 6. Scope of patent application 1. A method of optimizing the memory from the node of the server node for the use of caching of 70 pieces of shellfish for 2 weeks, including the following steps: π] ^ Each node of a plurality of nodes connected to the local area network ^ 1. Cache memory for the existence of a local entity; Click to provide the cache memory of at least some nodes of this complex node h And become a single virtual cache memory; and the hidden body and C delimit a specific one of these nodes ^ ^ ^ permanently store a specific caution button-downgrade from F々 click to use 即 I back up to the plural A reference node of a 4 7L piece of data element; ,, lack of kernel-to access the specific where a specific node is designated as the storage node, a copy backup of the data element is provided to this particular type of access ^ a reference Nodes. / V + file 2. If the method in the scope of the patent application is applied, it includes the following steps. U) If the data element to be accessed by a reference node is not a valid copy backup exists in virtual cache, select Obtaining a copy backup of the data element from the servo node, one of which uses virtual cache to access the copy backup of the data element, which takes less time than the copy backup of the sand point access data element. Expansion means 3, according to the method of the scope of the patent application, No. 1 and the following steps: (d) —The first reference node is restored from the storage node ... ^ μ * 制 # 于 · 于Click-to-get—a copy of a specific data component, 1223167 _案號89112368_年月曰 修正_ 六、申請專利範圍 (e )在該第一參考節點對該資料元件的複製備份,執 行一組一或多次存取作業;以及 (f )在步驟(d )之後,從該儲存節點提供一特定資 料元件的複製備份給想要存取卻缺少該特定資料元件的一 參考節點。 4、 如申請專利範圍第1項之方法,再包括下列步驟 :1223167 _ Case No. 89112368 _ revised month of the year _ 6. The scope of the patent application (e) copies and backs up the data element at the first reference node to perform a set of one or more access operations; and (f) at step (D) After that, a duplicate backup of a specific data element is provided from the storage node to a reference node that wants to access but lacks the specific data element. 4. If the method of item 1 of the scope of patent application is applied, it further includes the following steps: (d )指定一個節點當作特定資料元件的監視節點,該監 視節點以提供特定節點與特定資料元件的儲存節點的一致 性,而去回應各參考節點所提出、識別特定資料元件的儲 存節點的要求。 5、 如申請專利範圍第4項之方法,再包括下列步驟 : (e )至少在一第一參考節點,逐步增加地儲存從區域 網路偵測一或多個訊息而得到的資訊,其中每個偵測到的 訊息用以識別一特定節點當作監視節點、而該監視節點被 指定去鑑別一個被指定儲存一對應的資料元件的儲存節 點;以及(d) designating a node as a monitoring node of a specific data element, the monitoring node provides consistency between the specific node and the storage node of the specific data element, and responds to the storage node identification of the specific data element proposed by each reference node Claim. 5. If the method of claim 4 of the patent application scope further includes the following steps: (e) at least one first reference node, gradually increasing the information obtained by detecting one or more messages from the local area network, each of which A detected message to identify a particular node as a monitoring node, and the monitoring node is designated to identify a storage node designated to store a corresponding data element; and (f )至少在該第一參考節點,存取該資訊為的是企圖 決定一個被指定儲存一特定資料元件的儲存節點的一致 性、而該特定的資料元件乃該第一參考節點想要去存取 者。 6、 如申請專利範圍第5項之方法,其中至少該第一參考節(f) At least in the first reference node, the information is accessed in an attempt to determine the consistency of a storage node designated to store a specific data element, and the specific data element is the first reference node that wants to go Accessor. 6. The method of claim 5 in the patent application scope, wherein at least the first reference section 第53頁 1223167 案號 89112368Page 53 1223167 Case number 89112368 六、申請專利範圍 __ 點能夠逐步增加地儲存從區域網路 訊,而該等訊息是指向第一夂去 > 祝心偵測而到之資 …點以外的其他節點。 7、如申請專利範圍第5項之方法 步增加地儲存從區域網路作气^ ’其中該第一參考節點逐 訊息是僅指向第一參考節點者w偵测而到之資訊,而該等 如申請專利範圍〜,再 (d )在想要存取一第一次-再包括下列步驟·· 一二六〜貝料元件的一致性的一楚一儲存節點被指定儲 /子節點的一致性;以及 —監視節點,獲得該第 一=μ在該第—參考節點,從兮# 貝;7L件的一複製備份。Μ 一儲存節點、獲得該第 、申請專利範圍第1項之方漆 網路^少在—第一參考節點,’、Λ包括下列步驟: 訊息用多個訊息而得到的資Υ ^ 視節點被;ί二特定節點並將之於:广母個偵剛到的 健存節锂別-個被指定;;=見二:而該監 (,以及 々仔對應的資料元件的 決^ )至少在該第一炎者# 性指定健存-特定?料_,取該資訊為的是企圏 乃4第—參考節點想要去存取 第54頁 1223167 _案號 89112368_年月日__ 六、申請專利範圍 者。 1 〇、如申請專利範圍第9項之方法,其中至少該第一參考 節點能夠逐步增加地儲存從區域網路作訊息偵測而到之資 訊,而該等訊息是指向第一參考節點以外的其他節點。 11、如申請專利範圍第9項之方法,其中該第一參考節點 逐步增加地儲存從區域網路作訊息偵測而到之資訊,而該 等訊息是僅指向第一參考節點者。Sixth, the scope of patent application __ points can gradually increase the storage of messages from the local area network, and these messages are directed to the first node > funds from the heart detection ... other nodes than points. 7. According to the method of the scope of the patent application, the method step 5 stores the information from the local area network ^ 'wherein the first reference node message-by-message is only the information detected by the first reference node w, and these If the scope of the patent application is ~, then (d) when you want to access it for the first time-and then include the following steps ... 126 ~ the consistency of the storage element is designated as the storage / child node consistency And—monitor the node to obtain the first = μ at the first—reference node, a duplicate backup of 7L pieces. Μ a storage node, which obtained the first and the first patent application scope of the paint network ^ less-the first reference node, ', Λ includes the following steps: Information obtained by using multiple messages ^ depending on the node is ; ί two specific nodes and will be: the health festival section that Guangmu has just arrived-is specified; = see two: and the supervisor (and the decision of the corresponding data element of Taipa) at least in The first inflammation of the # sexually designated health-specific? Data _, take this information for the purpose of the enterprise, but the 4th-the reference node wants to access Page 54 1223167 _ case number 89112368 _ year month day __ Sixth, the scope of patent applications. 10. The method according to item 9 of the scope of patent application, wherein at least the first reference node can gradually increase the information obtained from the area network for message detection, and the information is directed outside the first reference node. Other nodes. 11. The method according to item 9 of the scope of patent application, wherein the first reference node gradually stores information obtained by detecting the information from the local area network, and the information is only directed to the first reference node. 1 2、如申請專利範圍第1項之方法,再包括:指定各監視 節點去儲存具有共同通用資源定址器(URL )資訊的一組 資料元件的步驟。 1 3、如申請專利範圍第11項之方法,再包括:依據共同通 用資源定址器(URL)資訊之子部分(subdivision),來 指定複數個監視節點給予一組資料元件的子集合的步驟。1 2. The method according to item 1 of the scope of patent application, further comprising the step of designating each monitoring node to store a group of data elements having a common universal resource addresser (URL) information. 1 3. The method according to item 11 of the patent application scope further includes the step of designating a plurality of monitoring nodes to give a subset of a set of data elements according to a subdivision of a common resource locator (URL) information. 1 4、如申請專利範圍第9項之方法,再包括下列步驟: (f )從第一參考節點發出一要求到已被識別的監視 節點,以識別資料元件的儲存節點。 1 5、如申請專利範圍第1 4項之方法,再包括下列步驟 : (g )從已被識別的監視節點發出一指明資料元件的14. The method according to item 9 of the scope of patent application, further comprising the following steps: (f) sending a request from the first reference node to the identified monitoring node to identify the storage node of the data element. 15. The method according to item 14 of the scope of patent application, further comprising the following steps: (g) sending a specified data element from the identified monitoring node 第55頁 !223167 曰 _案號 891123fiR 六、申請專利範圍 儲存節點之一致性的訊息’到該第—夂 、(h)從該第一參考節點發出對:次考^點; 的要求,到經由訊息已鑑定的儲存節貝點Λ件的複製備 (1 )從已被識別的儲存節 ‘·、,以及 製備份。 該肓料元件的-個複 16 點 性 如申請專利範圍第1 4項之方法, U )在已被識別的監視節點,^ 一包括^下列步驟: 接收一個鑑定資料元件的儲仗一特定的參考節 (h )如果監視節點缺少資料元 :胥永,以及 則該監視節點指定特定的來考凡〜的儲存節點的一致 件的儲存節點。 ^即點、當作被要求之資 17、如申請專利範圍第16項之方法, (i )如果特定的參考節點去獲取=—I列步驟: 份失敗,該特定的參考節點發出一自牛的一複製備 視節點,以取消將該特定的參考二==二破識別的監 節點的指定。 *、,占S作貝料元件的儲存 18、:=專:"'圍第”之方法,恤 點的次π & 第一參考即點只使用儲存在該第一夂考a 參考郎點發出訊息到區域網路中 貝丨k該第 的複數節點的子隼合,要求::中、指向少於全部或" 集口要衣確6忍-個被指定去識別資料天 修正 曰 —89112368 车 月 六、申請專利範^ ^~±一·- 牛的儲存節點的監視節點的一致性 19 > 接到 圍第18項之方法’再包括下列步驟: 節點回應傳送到區域網路、要求確認 夂1±的Λ心,無法正確地識別儲存節點, 區试*於該第一芩考節點、試圖從一經由廣域網路、查 (h )將次把一要的貧料元件; 丨的快取記憶體;以及 考即點之本地貫體 )該第—參考節點指定其本身為特 。 了夂貝抖兀件的儲存 20、, ·申請專利範圍第19項之方法,再 (])如果第一參考節匕括下列步驟·· 確認儲存h的-致性的P訊自任何監視節點發出 則為了對Jm點而無法識 組特定的資料元件、包含且古^其本身作為資料元 L )資訊的資料元件的儲存節點/、同通用資源定址器 ^申凊專利範圍第1 8項之方法, 鹿的於g、)複數個監視節點的各個節點包括下列步驟·· 心、皿硯節點確是以第一參考節點所,-错由表明個別回 _ 厅軚不的資料元件的監 第57頁 1223167 案號 89112368 年 月 曰 修正 六、申請專利範圍 視節點,而對發出訊息的該第一參考節點作回應;以及 (h )該第一參考節點從中選取回應最好的監視節點 當作資料元件的監視節點。 2 2、如申請專利範圍第2 1項之方法,再包括下列步驟: (i )在第一參考節點發出訊息後的特定期間内,於 該第一參考節點收集一回應監視節點所發出的每一訊息; 以及Page 55! 223167 _Case No. 891123fiR Sixth, the message of consistency of the patent application storage node 'to the first-(i), (h) The request from the first reference node to: secondary examination points; to The duplicate preparation (1) of the storage node identified by the message (1) is from the identified storage node, and a backup is made. The 16-point nature of the data element is the method described in item 14 of the scope of patent application. U) At the identified monitoring node, ^ one includes ^ the following steps: receiving a storage of an identification data element a specific Reference section (h) If the monitoring node lacks a data element: Yong Yong, and the monitoring node designates a specific storage node that is consistent with the storage node of Fanfan ~. ^ Points, as the required capital 17, such as the method of applying for the scope of the patent No. 16, (i) if a specific reference node to obtain =-column I steps: If the copy fails, the specific reference node issues a self-promotion The preparation of the view node is repeated to cancel the designation of the specific reference node == the second break identified by the monitor node. * 、, accounting for the storage of S as a shell material component 18: ==: "quote" method, the second point of the shirt point & the first reference point is only stored in the first test a reference Lang Send a message to the child node of the local plural node in the local network, request :: middle, less than all, or "quote", which is designated to identify the data. —89112368 Car month VI. Patent application ^ ^ ~ ± 一 ·-Consistency of monitoring node of cattle storage node 19 > Method of receiving item 18 'and including the following steps: Node response sent to the local network 2. Requires confirmation of the Λ heart of ± 1 ±, unable to correctly identify the storage node, and the district test * attempts to check a primary lean component from the first test node through the wide area network, (h); 丨The cache memory; and the local consistency of the test point) The first reference node designates itself as a special feature. For the storage of the sturdy parts, 20, · The method of applying for the scope of the patent, and then () ) If the first reference section includes the following steps: Confirm the storage of H-consistent P Issued from any monitoring node for the purpose of Jm point and cannot identify a specific data element, a storage node of a data element that contains and is used as a data element L) information, and the same as the general resource locator. In the method of item 8, each of the deer's multiple monitoring nodes includes the following steps: · The heart and dish nodes are indeed based on the first reference node,-the wrong reason indicates that the information returned by each individual Component monitoring Page 57 1223167 Case No. 89112368 Amendment VI. The scope of patent application depends on the node and responds to the first reference node that sends the message; and (h) The first reference node selects the best responding The monitoring node is regarded as the monitoring node of the data element. 2 2. If the method of the scope of patent application No. 21, further includes the following steps: (i) within a specific period after the first reference node sends a message, in the first reference The node collects a response to every message sent by the monitoring node; and (j )根據一預先定義的標準,從所收集的訊息選取 最佳的監視節點。 23、 根據申請專利範圍第22項之方法,其中第一參考節點 所選取最佳的監視節點是能識別最大數目的資料元件、這 些資料元件具有共同通用資源定址器(URL)資訊且是第一 參考節點所想要擷取者。(j) Select the best monitoring node from the collected information according to a predefined criterion. 23. The method according to item 22 of the scope of patent application, wherein the best monitoring node selected by the first reference node is the one that can identify the maximum number of data elements that have common universal resource locator (URL) information and are the first Who the reference node wants to retrieve. 24、 根據申請專利範圍第22項之方法,其中該第一參考節 點所選取最佳的監視節點是,在其他回應的監視節點發出 的其他訊息之前,其訊息被該第一參考節點所接收的監視 節點。 2 5、如申請專利範圍第1項之方法,再包括下列步驟 : (d )指定複數個節點當作複數個資料元件的儲存節 點;以及24. The method according to item 22 of the scope of patent application, wherein the best monitoring node selected by the first reference node is that the information received by the first reference node is received by the first reference node before other messages from other responding monitoring nodes. Monitoring nodes. 25. If the method of item 1 of the patent application scope further includes the following steps: (d) designating a plurality of nodes as storage nodes for a plurality of data elements; and 第58頁 -案號Ml 19?阳 修正 曰 六、申請專利範n S^毛月 (e ) 4日疋複數個節 點,每一監視節點、# : 5作複數個資料元件的監視節 要的資料元件、或一 提供給一特定的、被指定儲存想 的儲存節點之一致性為料元件其包含有想要的資料元件 複數個資料元件中之i而來回應參考節點所發出、去識別 的資料元件的儲存節點2、想要被對應的參寺節點所存取 的要求。 2 6、如申請專利範 ("如果有一第,之方法包括下列步驟: 少該複數個資料元件中沾個參考節點想要去存取、但卻缺 時,隨即決定該被指定用二個特定資料元件的複製備份 儲存節點、目前並沒右=J儲存該特定資料元件的一特定 點、在一段比該特定儲資料元件,然後該第—參考節 過時間更長的時間區間二即點/能夠提供特定資料元件的經 資料元件的複製備;:’即從-伺服器節點擷取該特定 (g )將操取的倍中次φ、丨— 一參考節點;以及 貝;斗70件複製備份,儲存在該第 第_ H複數,視節點中的-個特定監視節點,尸" 參考卽點當作該特定資料元件的儲存節即:。“該 27、如申請專利範圍第1項 (〇將複數個節點區分成各广下列步驟: 無即點的數個子集合; 炎数個非重疊、非空 e )在每一子集合 的母一節點執行步驟(b ) 以形 第59頁 1223167Page 58-Case No. Ml 19? Amendment VI. Patent application n S ^ Maoyue (e) 4th, a plurality of nodes, each monitoring node, #: 5 is a monitoring summary of a plurality of data elements. The consistency of a data element, or a storage node provided to a specific, designated storage node, is that the data element contains the desired i of the plurality of data elements and responds to the reference node's issue and identification. The storage node 2 of the data element is a request to be accessed by the corresponding temple node. 2 6. If there is a patent application method, the method includes the following steps: When there is no reference node in the plurality of data elements, but it is lacking, it is decided to use two. The copy and backup storage node of a specific data element is not currently right = J stores a specific point of the specific data element, in a period longer than the specific storage data element, and then the time period of the first-reference section is longer. / Reproduced by a data element capable of providing a specific data element: 'that is, fetching from the -server node the multiple times φ, 丨 — a reference node that this particular (g) will operate on; and a shell of 70 pieces Copy the backup and store it in the _hth complex number, depending on a specific monitoring node in the node. The reference point is regarded as the storage section of the specific data element: "The 27. The item (0 divides a plurality of nodes into the following steps: a number of sub-sets without a point; a number of non-overlapping, non-empty e) perform step (b) on the parent node of each sub-set to form the 59th Page 1223167 _案號 89112368 六、申請專利範圍 成複數個虛擬快取記憶體, 記憶體、其係由:只有對 子集口匕^ 一個虛擬快取 體,才被鏈結起來而形成^ ·該:集合的即點的快取記憶 (f )指定複數個儲存銘Μ — u — 忠妙六々A 仔即點’母個儲存節點係姑 末儲存各自特定的資料元糸破私定用 ΛΛ- tol te / L » L ' 並*對相同千集会"ίΛ甘/1 即點k供特定的資料元件 的其他參考 ^ ^ X ^ 丁 W複製備份’而母個炎去从 希望去存取但卻缺少該久 4考郎點是 豆由枝六… 各自資料元件的複製備份的r机 其中儲存節點的指定,不合 加姑 即點, 的複製備份給想要存取兮I二~ /、 、疋的貧料元件 個夂去 ^ 取该特疋資料元件的相同子隼人的 個參考即點而來作單獨改變。 j于杲σ的一 2 8、如申請專利範園坌 (彳圍苐27項之方法,再包括下列步驟: 的夂者#點沾、/、寺定的資料元件的一複製備份給一特定 的蒼考即點的過程中,^ % ▲ 域網路發出訊息;,複數個節點中之—個節點、在區 (h )和參考筋κ 6 點、以及和參考節點不Λ子集合不同子集合的每一節 應忽略該訊息。 虛擬快取記憶體的每一節點,都_Case No. 89112368 6. The scope of patent application is a plurality of virtual cache memories. The memory and its reason are: only a subset of the virtual cache can be linked together to form ^. The: collection Cache memory (f) of a given point specifies a plurality of storage names M — u — Zhong Miao Liu 々 A 即 points' parent storage nodes are the storage of specific data elements 糸 Λ- tol te / L »L 'and * for the same thousand rally " ίΛ 甘 / 1 point k for other references of specific data elements ^ ^ X ^ D W copying backups' while the mother goes to want to access but lacks The long 4 test point is a copy of the data backup of the respective data elements in the r machine, which specifies the storage node, which does not add a point, the backup is required to access the I II ~ /,, and the poor Material elements are removed ^ Take the individual reference points of the same child of the special data element to make separate changes.杲 σ 的 2 8. If the method of applying for a patent for a patent (范 27) method, including the following steps: 夂 者 # 点 沾, /, a copy of the data element of the temple to a specific backup During the Cangkao point-of-click process, the ^% ▲ domain network sent a message; one of the plurality of nodes, the zone (h) and the reference rib κ 6 points, and the reference node is not a different set of sub-sets. This message should be ignored for each section of the collection. Every node in the virtual cache 2 9 如申清專利範圚 · (d)為一:割臨Ϊ第項之方法,再包括下列步驟· 的資料n以:即點,參考節點從儲存節點獲得該特 (e ) /1 ]覽卽點執行_ "a免,丨的節點 求擷取包含特定的資料-f覽裔應用耘式,從別的I 貝科疋件的一或多個資料元件。2 9 If the patent patent application is cleared, (d) is one: the method of cutting out the item in item 1, and then includes the following steps. The data n is: point, the reference node obtains the special (e) / 1 from the storage node. At the point of execution, the node seeks to retrieve one or more data elements that contain specific data from a particular application. 第60頁 修正Page 60 Correction SS_^9j_12368 六、申請專利範圍 3 0、一種將來自伺服器 吕己憶體作表適化之系統 被連接到區域網路 一被指定當作儲存節點 該一特 卻缺少 資料元件並提供 點中、想要存取 以及 複數個快取 記憶體 性實體 來而形 定節點 的複製 一個參考節點而來作單 提供的一個區域 記憶體被鍵結起 其中指定特 特定的資料元件 P點可供使用資料元件之區域 ’包括: 的複數個節點,該複數個節點包含 =特疋節點、其恆久地儲存一特定 疋寅料元件的複製備份給複數個 該特定資料元件的其他參考節:即 ^包含複數個節點之每一節點其被 存在的快取記憶體,至少一些快取 成一單一虛擬快取記憶體, 當作儲存節點,不會藉由提供一個 備份給想要存取該特定資料元件的 獨改變。 ' 3上如申請專利範圍第3〇項之系統,其中複數個節點再包 M It Z選擇地從伺服器節點獲取資料元件之-複製備 貧料凡件的複製備份時、其可從伺服器節點獲取該Ϊί :牛的複製備份,雖則從虛擬快取記憶體獲取的時間、:^ 伺服器節點獲取的時間要少。 比從 32、如中請專利範圍第31項之系統,其中複數個節點再包SS_ ^ 9j_12368 VI. Scope of patent application 30. A system that adapts the table from the server Lu Jiyi body to the local area network is designated as a storage node, but it lacks data components and provides a point. , Want to access and a plurality of cache memory entities to form a node to copy a reference node to provide a single area of memory is keyed in which the specified specific data element P point is available for use The area of the data element 'includes: a plurality of nodes, the plurality of nodes including = special nodes, which permanently store a copy of a specific data element backup to a plurality of other reference sections of the specific data element: that is, ^ contains Each node of the plurality of nodes has its existing cache memory, at least some of which are cached into a single virtual cache memory, as storage nodes, and will not provide a backup to those who want to access the specific data element. Independence changes. '3 as in the patent application system No. 30 system, in which a plurality of nodes are repackaged M It Z to selectively obtain data elements from the server node-when copying backup copies of poor materials, it can be downloaded from the server The node obtains the duplicate copy of the bull: Although it takes less time to obtain from the virtual cache memory, the: ^ server node takes less time. Than 32, such as the system of the 31st patent scope, where multiple nodes are repackaged 第61頁 1223167 案號 891123BS 六、申請專利範圍 曰 修正 一第一參考節點,其能你 — 件的複製備份,且能對該特定資;卽:獲得:特定資料元 組包含一個或多個的存取操作·、’以^牛的複製備份執行— 失者-點〜3 :取:缺少該特定資料元件之複制借 製備份,但須在提供特定的:予$供特定資料元件的複 參考節點之後。 ’凡牛的複製備份給該第一 33、 如申請專利範圍第31 括: 項之系統’其中複數個節點再包 一被指定到特定資料元件 每一參考節點所送來識別特定次:_印點,該監視節點對 求,以提供特定節點和特定次二’、凡件的儲存節點的要 來回應之。 貝料元件的儲存節點的一致性 34、 如申請專利範圍第33 : 、之系、、統,其中複數個節點包括 逐步增加地儲存從區域網路 之資訊的一第一參考節點,其、或多個訊息而得到 別-特定節點當作-監視節點的訊息用於識 儲存對應的資料元件的儲存節點Λ?即點被指定去識別 —些資訊、為的是嘗試識別特定的=第:參考節點也存取 而該特定的資料元件乃是該第—貧料元件的儲存節點, 參考節點想要去存取的。 1223167 _紐89112細 六、申請專利範圍 35、如申請專利範圍第34項之系統,其 節點可逐步增加地儲存從區域網路^自;^〜苐〜參考 訊該等訊息是指向第-參考節點二;的::?到的資 』八他郎點。 36、如申請專利範圍第34項之系、紙, 訊 可逐?增加地儲存從區域網路作訊息;貞參考節點 而該等訊息僅指向該第一參考節點。、、于1丨之資 37 括 其中複數個節點包 如申請專利範圍第3 0項之系統 一第一儲存節點,其被指定儲 欠 一第一監視節點,並被指 第一貝料元件; 元件的參考節.點、標示該第節以望2第一資料 -第-參考節點,其想要存取巧付:;性;以及 該第一監視節點獲得該第一儲=第一 _貝料元件,以從 節點也可從該第一儲存節點莽之識別,該第一參考 份。 · X侍該第一資料元件的複製備 2 :如申請專利範圍第30項之系統,其中複數個節點包 〜/ >日^〜吨仔從區域網路 之資訊的一第一參考筋 、、或夕個訊息而續 別一特定節點當作—於损μ ^中每個偵測到的訊息用 |視即點、該監視節點被指定去 1223167Page 61 1223167 Case No. 891123BS Sixth, the scope of the patent application is to amend a first reference node, which can make a copy of your files and back up the specific resources; 卽: Obtain: the specific data tuple contains one or more Access operation ·, 'Implementation of copy backup execution — Lost-point ~ 3: Take: Missing copy backup of the specific data element, but must provide specific: $ for the reference of the specific data element After the node. 'Fan Niu's copy is backed up to the first 33, such as the scope of the patent application, including the 31st item of the system', where a plurality of nodes are then assigned to a specific data element, each reference node sent to identify a specific time: _ 印At this point, the monitoring node asks to provide the specific node and the specific secondary storage node's requirements to respond to it. The consistency of the storage nodes of the component 34, such as the scope of patent application 33: system, system, wherein the plurality of nodes include a first reference node that gradually stores information from the local area network, which, or Multiple messages to get other-specific nodes as-monitoring messages are used to identify the storage node Λ that stores the corresponding data element. That is, the point is designated to identify-some information in order to try to identify a specific ==: reference Nodes also access and the specific data element is the storage node of the first-lean component, the reference node wants to access. 1223167 _New 89112 details VI, patent application scope 35, such as the patent application scope item 34 system, its nodes can be gradually increased from the local area network ^ from; ^ ~ 苐 ~ reference information These messages are directed to the-reference Node two; ::? Arrivals 』Hachitaro. 36. If the application for patent No. 34 is related to paper or paper, can the information be sought? Increasingly stores messages from the local area network; the reference nodes and these messages only point to the first reference node. The capital 37 in 1 丨 includes a plurality of nodes including a system-first storage node such as the 30th in the scope of patent application, which is designated to owe a first monitoring node and is referred to as the first shell material element; The reference section of the component. Point and mark the section to look at the 2nd information-the first-reference node, which wants to access Qiaofu :; and the first monitoring node obtains the first storage = first_be The component can be identified from the first node or the first reference node from the first storage node. · The second preparation of the first data element 2: such as the system of the 30th scope of the patent application, in which a plurality of node packages ~ / > day ^ ~ a first reference bar of information from the local network, , Or even a message, and a specific node is continued as-for each detected message in the loss μ ^, the point of view, the monitoring node is designated to 1223167 修正 案號 89112368 六、申請專利範圍 儲存對應的資 一些資訊、為 而該特定的資 3 9、如申請專 節點可逐步增 訊,而該等訊 4 0、如申請專 可逐步增加地 而該等訊息僅 料元件的儲存節點,誃 的是嘗試識別特定的節點也存取 料元件乃是該 2二3凡件的儲存節點, 參考卽點想要去存取的。 利範圍第38項之糸紅 ^ 加地儲存從E :、 /、中至少該第一參考 息是指向該第一夂I ^心偵測而得到的資 >考卽點以外的其他節點。 利範圍第38項之系統, 儲存從f a _ 共干該第一參考節點 _ & $ : 2、、路作訊息偵測而得到之資訊, 才曰向該第一參考節點。 41、如申請專利範圍第3〇 銘赴 — Μ、日— 、之系統’再包括一或多個監視 即點,母一監視節點被 7似皿祝 料元件。 皮扣疋到具有共同URL·資訊的一組資 I; t it :專利祀圍第4〇項之系統,再包括複數個監視節 等指定m料遍1vlslQn)、而將彼 I3二如t凊專利範圍第38項之系統,#中該第-參考節點 ί向已識別的監視節點發出要求,以識別資料元件的儲存 卽點。 τAmendment No. 89112368 Sixth, the scope of the patent application stores some information about the specific information, and for this specific information 39, if you apply for a special node, you can gradually increase the information, and if such information 40, you can gradually increase the number of The information is only the storage node of the component. What I am trying to identify is that the specific node also accesses the component. This is the storage node of the 2 2 3 pieces. Refer to the point where you want to access. The red reference of item 38 of the scope of interest is added to store at least the first reference information in E :, /, and the data obtained by pointing to the first reference point > other nodes than the test point. The system of item 38 of the profit range stores the information obtained from the f a _ co-operation of the first reference node _ & $: 2. The information obtained from the detection of the road information is sent to the first reference node. 41. If the scope of the patent application is 30th, the system of “M, M, Day,” further includes one or more monitoring points, and the parent-monitoring node is a 7-element blessing element. Leather buckle to a set of resources with a common URL and information; t it: the system of patent enclosing item 40, and then including a plurality of monitoring sections and other designated materials (1vlslQn)), and the other I3 as t 凊The system of item 38 of the patent, in which the # -reference node ί sends a request to the identified monitoring node to identify the storage point of the data element. τ 第64頁 上--一案色 /、、申請專利範圍 ^、如申請專 節點乃在向第 存節點的一致 別的儲存節點 且從該已識別 =、如申請專 節點乃在從— 節點之要求, 郎點的一致性 料元件的儲存On page 64-a case /, the scope of the patent application ^, if the application node is in the same storage node as the storage node and from the identified =, if the application node is in the slave node Requirement, storage of Lang point uniform material components 利範圍一參考 性;又 發出一 的儲存 利範圍 特定的 該已識 ’則也 節點。 f4 3項之系統,其中該已識別的監視 即點發出一訊息以指出資料元件的儲 其中該第一參考節點乃在向由訊息識 f取一資料元件的複製備份的請求, 雖點接收該資料元件的複製備份。 第4 3項之系統,其中該已識別的監視 參考節點接收去識別資料元件的儲存 別的監視節點若缺少資料元件的儲存 指定該特定的參考節點當作所要求資 4 6、如申請專利 點在要獲取資料 視節點發出一訊 料元件的儲存節 47、如申請專利 也於區域網路發 個郎點的子集合 的資訊無法識別 指定去識別資料 範圍第4 5項之系統,其中該特定 元件的複製備份失敗時,乃向已 息’用以取消該特定的參考節點 點的指定。 範圍第38項之系統,其中該第一 出一指向一個包含少於全部、或 的訊息,在第一參考節點僅利用 貧料元件的監視節點時,請求確 义件的儲存節點的一致性。 的參考節 識別的監 被當作資 參考節點 全部複數 其所儲存 認一個被 48 、如申請專 利 範圍第4 7項之系統,其中若第一參考節點The profit range is a reference; the storage of a profit range is issued, and the identified ’is also a node. f4 item 3 system, in which the identified surveillance point sends a message to indicate the storage of the data element, where the first reference node is requesting a copy of a data element from the message identification f, although the point receives the Copy backup of data components. Item 43. The system in which the identified monitoring reference node receives to identify the storage of the data element. If other monitoring nodes lack the storage of the data element, the specific reference node is designated as the required asset. In the storage section 47 of the information element sent by the video node to obtain data, if the patent application is also issued on the local network, the information of the sub-collection of the sub-collection cannot be identified. The system designated to identify the data range of item 4 and 5, in which the specific When the component backup fails, it is used to cancel the designation of the specific reference node point. The system of the scope item 38, wherein the first output points to a message containing less than all or OR, and when the first reference node only uses the monitoring node of the lean component, the consistency of the storage node of the requested component is requested. The reference section identified is used as the reference node. All the nodes are stored. It is recognized as a 48, as in the application of the patent scope of item 47. Among them, if the first reference node 第65頁 Λ 日 --紐89m气 六、申請專利範圍 在回應由區域網路傳 ' ^~ ^:域,法正確識別儲存節點,;:5:節點的一致性的訊 由廣域網路從一連接到區域網=j弟一參考節點試圖經 料兀件,接著將該資料件5服節點擷取想要的資 件的儲存節點。 ‘體立4日疋其本身為該特定資料元 4 9如申凊專利範圍第4 8頂之糸纪 由於没有任何監插# 、 二、、、,其中若第一參考節點 自 1、# 皿即點傳來指出儲存節點Μ ^ . ,· 息,而導致識別儲存節點 ^即2的一致性的訊 訊息、要求確認監视節點的一致性::二從:J網路傳來 ;;即指定其本身當作資料元件、或:::特定考節 含具有共同url資訊的資料元件的儲存節點。件其包 ^如中請專利範圍糾項之系統,其中複數個節點包 複數個i視節點’其每一個監視節點被用 訊息作回應、以指出對應的回應監。 盆疋第參考節點所規範的貪料元件的監視節點, 其中第一參考節點也在選取回應最佳的監視節點,者 料元件的監視節點。 田作資 51、如申請專利範圍第5〇項之系統,其中該第_參 也在其發出訊息後之特定時間内,收集從回應監視節=發Page 65 Λ Day--New York 89m 6. The scope of the patent application is in response to the transmission of the local network's ^ ~ ^: domain, method to correctly identify the storage node;; 5: the consistency of the node is signaled by the wide area network from a A reference node that is connected to the local area network tries to pass the data element, and then fetches the data node and the storage node of the desired data. 'Ti Li 4th' itself is the specific data element 4 9 As stated in the scope of patent application No. 48 top, there is no monitoring insert #, 二 ,,, where, if the first reference node is from 1, # 皿That is to point out that the storage node M ^., · Information, which leads to the identification of the consistency of the storage node ^ 2 is required to confirm the consistency of the monitoring node: 2: from the J network; Designate itself as a data element, or a storage node containing ::: a specific test section containing data elements with common url information. The system includes a patent correction system as described in the above, wherein a plurality of nodes include a plurality of i-view nodes' and each monitoring node is responded with a message to indicate a corresponding response supervisor. The monitoring node of the raw material element specified by the reference node of the basin, and the first reference node is also selecting the monitoring node that responds best, and the monitoring node of the component. Tian Zuozi 51. If the system of the scope of patent application No. 50, where the _ reference also collects from the response monitoring section = issued within a specific time after the message is sent. 第66頁 1223167 _案號89112368_年月曰 修正_ 六、申請專利範圍 出的每一訊息,並根據預先決定的一標準從收集訊息中選 取最佳的監視節點。 52、如申請專利範圍第5 1項之系統,其中由第一參考節點 所選出的最佳監視節點是能識別最大數目的資料元件彼等 具有共同URL資訊且是第一參考節點所想要擷取者。Page 66 1223167 _Case No. 89112368_ Year Month Amendment_ VI. For each message in the scope of patent application, and select the best monitoring node from the collected messages according to a predetermined standard. 52. The system according to item 51 of the scope of patent application, wherein the best monitoring node selected by the first reference node is the one that can identify the largest number of data elements. They have common URL information and the first reference node wants to retrieve it. To take. 5 3、如申請專利範圍第5 1項之系統,其中第一參考節點所 選取最佳的監視節點是,在其他回應的監視節點發出的其 他訊息之前,其訊息被第一參考節點所接收的監視節點。 54、 如申請專利範圍第30項之系統,其中複數個節點包 括:5 3. According to the system of claim 51 in the scope of patent application, the best monitoring node selected by the first reference node is that the other reference node's message is received by the first reference node before other messages sent by the responding monitoring node. Monitoring nodes. 54. If the system for applying for item 30 of the patent scope, the plurality of nodes include: 被指定儲存複數個資料元件的複數個儲存節點;以及 被指定作為複數個資料元件的複數個監視節點,每一 監視節點藉由提供特定的一個儲存節點其作為想要的資料 元件、或一組資料元件其包含想要的資料元件的一個儲存 節點之一致性,而去回應參考節點發出、去識別該參考節 點所想要存取的複數個資料元件中的一個資料元件之儲存 節點的要求。 55、 如申請專利範圍第54項之系統,其中複數個節點包 括: 一第一參考節點,其想要去存取、但卻缺少該複數個A plurality of storage nodes designated to store a plurality of data elements; and a plurality of monitoring nodes designated as a plurality of data elements, each monitoring node providing a specific storage node as a desired data element, or a group The data element contains the consistency of a storage node of the desired data element, and responds to the request from the reference node to identify the storage node of one of the plurality of data elements that the reference node wants to access. 55. If the system of claim 54 is applied for, the plurality of nodes includes: a first reference node that wants to access but lacks the plurality of nodes 第67頁 --塞號 89112368 六、申請專利範圍 — 貧=元件中的一個特定資料 j :決定Ϊ被指定用來儲存 即··、、、目Μ並沒有儲存資料 門争^比該特定儲存節點能夠 間更長的時間區間内、即從一 =的複製備f分,並將擷取的 存在該第一參考節點, ί:! 复數個監視節點中的-個 多考郎點當作特定資料元件的 6如申睛專利範圍第3 0項之 分屬複數個非重疊、非介 荨子集合是由I數個 揆版:數個虛擬快取記憶體, =記憶體,其中每個虛擬 的;:I集合的節點的快取記 自特ΐ Ϊ個儲存節點,每個儲 牲〜 貝#凡件、並對相同 取i的:料元件的複製備份, 其;::!該各自資料元件的 ^儲存卽點的指定,不會芦 個份給想要存取該“ 固參考卽點而來作單獨改變。 曰 修正 件的複製備份,若該第一參 該特定資料元件的一特定儲 70件,則該第一參考節點、 提供特定資料元件的經過時 词服器節點擷取該特定資料 特定資料元件複製備份、儲 特定監視節點,指定該第一 儲存節點。 系統,再包括: 無節點子集合的各節點,該 割而來的; 包含每個子集合其有一個虛 快取記憶體其係由:只有對 憶體,才被鏈結起來而形成 存節點係被指定用來儲存各 子,合的其他參考節點提供 ,每個參考節點是想要去存 複製備份的節點, t提供一個特定的資料元件 貝料元件的相同子集合的一Page 67--Serial No. 89112368 VI. Scope of patent application-poor = a specific piece of information in the component j: decided to be designated to store i.e., ..., and there is no data door to store the content ^ than that specific storage Nodes can prepare f-points from a complex of one = for a longer time interval, and use the extracted and stored in the first reference node, ί :! One of the multiple monitoring nodes in the plurality of monitoring nodes as a specific The 6th element of the data element belongs to the 30th of the patent scope. The non-overlapping and non-intermediate net set is composed of several virtual versions: several virtual cache memories, = memory, where each virtual The cache of the nodes of the I set is recorded from a special storage node, each storage node ~ ###, and a duplicate backup of the same: i ::! The designation of the storage points of the respective data elements will not be changed individually for those who want to access the "solid reference point." A copy of the correction copy will be backed up if the first reference to the specific data element A specific storage of 70 pieces, then the first reference node, the elapsed time server node that provides the specific data component retrieves the specific data, the specific data component, copies the backup, stores the specific monitoring node, and designates the first storage node. It also includes: the nodes of the nodeless sub-collection, which are cut off; including each sub-collection, which has a virtual cache memory, the reason is: only the memory body is linked to form a storage node system, which is designated It is used to store each child, and other reference nodes are provided. Each reference node is a node that wants to save a duplicate backup. T provides a specific data element. 1223167 _案號89112368_年月曰 修正_ 六、申請專利範圍 5 7、根據申請專利範圍第5 6項之系統,其中複數個節點中 的一個節點、在提供特定的資料元件的一複製備份給一特 定的參考節點的過程中,在區域網路發出一訊息;又其中 和該參考節點所屬子集合不同子集合的每一節點、以及和 該參考節點不同虛擬快取記憶體的每一節點,都應忽略該 訊息。 58、根據申請專利範圍第30項之系統,其中複數個節點包 括: 一瀏覽節點,其執行一瀏覽器應用程式,以從別的節 點要求擷取包含特定資料元件的一或多個資料元件;以及 一參考節點,其為一瀏覽節點而從儲存節點獲取該特 定資料元件。1223167 _ Case No. 89112368 _ revised month of the year _ 6, the scope of the patent application 5 7, according to the scope of the patent application of the 56th system, where one of a plurality of nodes, a copy of a specific data element to provide a backup copy to In the process of a specific reference node, a message is sent on the local area network; and each node in the subset different from the subset to which the reference node belongs, and each node in the virtual cache memory different from the reference node, Should ignore this message. 58. The system according to item 30 of the scope of patent application, wherein the plurality of nodes include: a browsing node that executes a browser application program to retrieve one or more data elements containing specific data elements from other nodes; And a reference node, which obtains the specific data element from the storage node for a browsing node. 第69頁Page 69
TW089112368A 1999-06-23 2000-09-22 Distributed virtual web cache implemented entirely in software TWI223167B (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US14064599P 1999-06-23 1999-06-23

Publications (1)

Publication Number Publication Date
TWI223167B true TWI223167B (en) 2004-11-01

Family

ID=22492183

Family Applications (1)

Application Number Title Priority Date Filing Date
TW089112368A TWI223167B (en) 1999-06-23 2000-09-22 Distributed virtual web cache implemented entirely in software

Country Status (6)

Country Link
EP (1) EP1259864A2 (en)
JP (1) JP2003515197A (en)
AU (1) AU5756000A (en)
CA (1) CA2341595A1 (en)
TW (1) TWI223167B (en)
WO (1) WO2000079362A2 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008065341A2 (en) 2006-12-01 2008-06-05 David Irvine Distributed network system
USRE47411E1 (en) 2005-08-16 2019-05-28 Rateze Remote Mgmt. L.L.C. Disaggregated resources and access methods

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2366406A (en) * 2000-09-01 2002-03-06 Ncr Int Inc Downloading and uploading data in information networks
US7720959B2 (en) 2000-10-17 2010-05-18 Avaya Inc. Method and apparatus for characterizing the quality of a network path
US7756032B2 (en) 2000-10-17 2010-07-13 Avaya Inc. Method and apparatus for communicating data within measurement traffic
US8023421B2 (en) 2002-07-25 2011-09-20 Avaya Inc. Method and apparatus for the assessment and optimization of network traffic
US7689601B2 (en) * 2004-05-06 2010-03-30 Oracle International Corporation Achieving web documents using unique document locators

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5778368A (en) * 1996-05-03 1998-07-07 Telogy Networks, Inc. Real-time embedded software respository with attribute searching apparatus and method
US5878218A (en) * 1997-03-17 1999-03-02 International Business Machines Corporation Method and system for creating and utilizing common caches for internetworks

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
USRE47411E1 (en) 2005-08-16 2019-05-28 Rateze Remote Mgmt. L.L.C. Disaggregated resources and access methods
WO2008065341A2 (en) 2006-12-01 2008-06-05 David Irvine Distributed network system

Also Published As

Publication number Publication date
WO2000079362A2 (en) 2000-12-28
EP1259864A2 (en) 2002-11-27
CA2341595A1 (en) 2000-12-28
JP2003515197A (en) 2003-04-22
AU5756000A (en) 2001-01-09
WO2000079362A9 (en) 2002-05-02
WO2000079362A3 (en) 2002-09-12

Similar Documents

Publication Publication Date Title
Stribling et al. Flexible, Wide-Area Storage for Distributed Systems with WheelFS.
JP6261758B2 (en) Implementation of services to coordinate container placement and execution
US11403262B2 (en) Local networked storage linked to remote networked storage system
JP6448555B2 (en) Content class for object storage indexing system
US9032393B1 (en) Architecture for incremental deployment
TW501032B (en) Method and system in a computer network for the reliable and consistent ordering of client requests
US8996610B1 (en) Proxy system, method and computer program product for utilizing an identifier of a request to route the request to a networked device
US9218405B2 (en) Batch processing and data synchronization in cloud-based systems
JP2018507480A (en) Method and apparatus for storing instant messaging chat records
US20100030818A1 (en) System and method for applying once a transaction delivered in a message published asynchronously in a distributed database
US10572466B2 (en) Multiple collections of user-defined metadata for self-describing objects
US20110055156A1 (en) Method for efficient storage node replacement
JP7331073B2 (en) Enhanced online privacy
WO2021007752A1 (en) Return-to-source method and related device in content delivery network
TW200935245A (en) System and method for providing visibility for dynamic webpages
US8984162B1 (en) Optimizing performance for routing operations
WO2022127762A1 (en) Cloud platform and bucket management method for object storage service provided thereby
US20020133597A1 (en) Global storage system
TWI223167B (en) Distributed virtual web cache implemented entirely in software
WO2005109252A1 (en) System and method for facilitating a consistent point-in-time copy
CN108874472B (en) Method and system for optimally displaying user head portraits
US20130054691A1 (en) Flexible rule based multi-protocol peer-to-peer caching
US10567499B1 (en) Unsupervised round robin catch up algorithm
US7143244B2 (en) System and method for invalidating data in a hierarchy of caches
Nelson et al. Syndicate: Virtual cloud storage through provider composition

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees