[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

TW479200B - Method and apparatus for caching for streaming data - Google Patents

Method and apparatus for caching for streaming data Download PDF

Info

Publication number
TW479200B
TW479200B TW89118815A TW89118815A TW479200B TW 479200 B TW479200 B TW 479200B TW 89118815 A TW89118815 A TW 89118815A TW 89118815 A TW89118815 A TW 89118815A TW 479200 B TW479200 B TW 479200B
Authority
TW
Taiwan
Prior art keywords
name
cache
video
scope
blocks
Prior art date
Application number
TW89118815A
Other languages
Chinese (zh)
Inventor
Horng-Juing Lee
Original Assignee
Streaming21 Inc
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 Streaming21 Inc filed Critical Streaming21 Inc
Application granted granted Critical
Publication of TW479200B publication Critical patent/TW479200B/en

Links

Landscapes

  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

Streaming video data from the video title is delivered to a client from a central server and a proxy server. To reduce the backbone bandwidth required, a portion of the video title is cached at the proxy server where the portion cached is distributed over the video title so as to reduce the peak bit rate required for the central server to send the remaining portion of the video title. The amount cached at the proxy server may be adjusted in accordance with access profile of video titles to improve caching efficiency. The portion cached may be the basic layer in a scalable multilayer approach where only the basic layer may be sent to a client who has limited bandwidth while clients with higher bandwidth, capability may also receive the enhancement layer(s) for improved video reception.

Description

479200 A7 _B7_;__ 五、發明說明(1 ) 相關專利申請案 本專利申請案享有臨時專利申請案60/153,905,名稱 METHOD AND SYSTEM FOR VIDEO DELIVERY OVER DATA NETWORKS,申請曰 1999/09/14 的優先權。 上述專利申請案在此供參考,該先前專利申請案的一部 分也在説明本發明,或是説明特定實施例及範例,因此可 參考以下説明而了解本發明。 發明範疇 本發明大致有關於資料傳送的方法及/或裝置,且尤其 有關於多媒體資料方法及應用,及系統及/或裝置及/或方 法以改良快取,其特別適用於成_數據。 著作權宣告 專利申請案的部分内容包括著作權保護的範圍,該著作 權所有者不反對任何人將本發明内容(或是於出現在專利 商標局的檔案時)加以影印,但是除此以外則保有該著作 權。 前案説明 以下文件與本發明有關,或是提供背景資訊,這些參考 文件不應該視爲已完成本發明的正式前案查尋,或者這些 參考文件的任一者構成本發明的前案。 發明背景 網際網路已快速成爲世界上互聯電腦的通訊網路,數百 萬台互聯電腦形成超媒體資訊的廣大儲存區,以供用户藉 由任何互聯電腦而隨時隨地的擷取,有越來越多的用户以 -4 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 請 先 閱 讀 背 & 之 注 意 事I 再 填 1 J裝 頁 訂479200 A7 _B7_; __ 5. Description of the invention (1) Related patent applications This patent application enjoys a provisional patent application 60 / 153,905 with the name METHOD AND SYSTEM FOR VIDEO DELIVERY OVER DATA NETWORKS, and has priority of application on September 9, 1999. . The above patent application is hereby incorporated by reference, and a part of this prior patent application is also describing the present invention or specific embodiments and examples, so the following description can be used to understand the present invention. Scope of the invention The present invention relates generally to data transmission methods and / or devices, and more particularly to multimedia data methods and applications, and systems and / or devices and / or methods to improve caching, which are particularly suitable for generating data. Part of the content of the patent application for copyright declaration includes the scope of copyright protection. The copyright owner has no objection to the photocopying of the present invention (or when it appears in the file of the Patent and Trademark Office), but otherwise retains the copyright. . Explanation of the previous case The following documents are related to the present invention or provide background information. These reference documents should not be regarded as having completed the formal previous case search of the present invention, or any of these reference documents constitutes the previous case of the present invention. BACKGROUND OF THE INVENTION The Internet has rapidly become the communication network of connected computers in the world. Millions of connected computers form a vast storage area for hypermedia information for users to retrieve anytime, anywhere with any connected computer. Many users apply the Chinese National Standard (CNS) A4 size (210 X 297 mm) with a paper size of -4. Please read the back &

I 經濟部智慧財產局員工消費合作社印製 479200 A7 --------- B7_ .___ 五、發明說明(2 ) 上網方式擷取各種資訊,同時要求得到更多資訊,及要求 用各種經由網際網路而將資訊傳送給他們,目前網際網路 上最常要求得到的資訊是文字資訊,影像及圖形,視訊及 聲訊等。 連續資訊如連續視訊,聲訊,視聽或是其他多媒體作品 (以下總稱爲媒體)是時常要求的網路資源之一,因此時常 會發生同時有數千個要求取用一個受歡迎的視訊,聲訊或 視聽節目,在資料網路的一定網路頻寬下,節目的即時傳 送且網路的服務品質(Q〇S )又有保證因而變的極重要,依 賴中央媒體伺服器以同時替所有的取用請求服務是不實際 的,因此需要一種多功能的媒體傳送系統,其能同時回應 數千個取用請求,而且對於服務品質的影響減到極小。 經濟部智慧財產局員工消費合作社印製 由於視訊壓縮,網路結構,及整體電腦功能等技術的進 步,使得即時視訊(V0D )成爲可能,近年來,網際網路及 全球資訊網的歡近更加速V〇D服務的實行,如 RealNetwork公司的Real Media及微軟公司的NetSh〇w Service都成功的在網際網路上傳送低品質的聲訊/視訊, 該等服務已用於多媒體應用的許多方面包括(但不限於此) 即時新聞,遠距敎學,公司訓練及音樂播放,有更多的人 們上網以找尋想要的資訊,這觸發將更多的資訊傳送到網 站0 爲了吸引更多的人上網站,於是公司設計網頁上面有許 多媒體資訊,尤其是增加連續的媒體資訊,惟這種連續媒 體的特徵如大的儲存要求,高網路都調整要求,及即時傳 -5- 本紙張尺度適用中國國豕標準(CNS)A4規格(21〇 X 297公爱) 479200 A7 ---------B7_____ 五、發明說明(3 ) 运等已使網際網路上的媒體分送服務變成大問題,這對於 具有高視訊内容的媒體而言更爲眞實,即時視訊伺服器^ 解::此問題的良好方法但不是沒有缺點,—受歡迎的網: 可能同時有數千個取用,因此視訊伺服器系統如合處理所 有的這些請求及傳送視訊到終端用户(品質又要有保證)是 一件急迫的事,一種應付伺服器擴充性的方法是使用數= 機聚集技術以建立平行的視訊伺服器,惟這種方法不適^ 廣大的網路取用,所以中央視訊伺服器不能服務所有 對端成_請求。 另一種方法是在現存網際網路取用模型上增加視訊傳送 服務,網路服務的普及已產生許多新的商機,而網際網路 服務提供者(ISP)是其中之一,其提供必要的結構,設備 及服務以使用户可容易且快速的與網際網路連線,爲了給 用户更好的服務,I s ρ公司向一地理區域中的用户提供服 矛力所有的I S Ρ公司都以鬲容量及筇貴的主幹網路來連 接,所有的ISP公司工作在一起而成爲一單元,用户的區 域I S Ρ站台負責接收用户的取用請求,必要時將請求傳送 到原始網站,作爲代理以接收網站的回覆,及傳送請求資 Λ回到請求者’傳送請求到原始網站及接收它的回覆等步 驟會使主幹過於忙碌,當過多請求通過主幹網路時,即需 要較高的主幹頻寬,這表示][S ρ的操件成本增加。 爲了解決此問題,網際網路已開發出一種網路快取策 略’網路代理伺服器安裝在〗S P站台且作爲快取引擎以儲 存時常取用的網頁(其與用户區域有關),下次當使用快取 -6- 本紙張尺度適用中國國家標準(CNS)A4規格(21〇 x 297公釐) (請先閱讀背面之注意事項再填寫本頁) 裝 . 經濟部智慧財產局員工消費合作社印製 479200 A7 B7 五、發明說明(4 ) 頁時,代理伺服器即直接的從其區域快取(而不是從原始 網路伺服器)傳送網頁,這種方法可減少用户的瀏覽回應 時間(因爲内容是來自區域代理伺服器),也可減少主幹對 於傳送内容的要求,因而減少操件成本。 上逑網路快取模型也適用於大型端對端視訊傳送伺服 經濟部智慧財產局員工消費合作社印製 已開發出許多網路代理伺服器產品,其中快取用元件主 要疋指HTML網頁及其内嵌影像,多數的習用網路代理一 般不是指視訊代理,其他早期研究指出關於端對端視訊傳 送的Q 〇 S (服務品質)議題,但是仍未指出視訊代理,已 提議用一些方法在視訊分級設計中減少分散多媒體傳送上 王幹的頻寬要求,在這些系統中,當用户要求一特別視訊 時’邵分的視訊内容即從中央視訊伺服器經由W an連線 而成串送出,而部分的視訊内容從區域代理伺服器成串送 出,這2種資料_都進入用户的播放站台,若大部分的資 .料是從區域伺服器成申送出,則從主幹通道成_送出的資 料較少,因而減少主幹頻寬要求,因此先前的視訊分級方 法是預擷取一預設量的視訊資料且將它儲存在代理伺服 器,代理伺服器根據先前概述周期中產生的取用方式而儲 存該預設視訊内容,而不是目前取用動作。一些早期設計 僅能改良編碼成VBR (可變位元率)格式内容的主幹利 用,至於CB R (定位元率)編碼視訊内容,一些習用方法 不會減少任何W AN頻寬要求。 另一種方法是快取一滑動窗的資料供各顯示視訊用,以 479200 經濟部智慧財產局員工消費合作社印製 A7 B7_;__五、發明說明(5 ) 便視訊能部分儲存在代理伺服器,因此若用户請求的電影 到達代理伺服器,則代理伺服器在其區域快取中儲存W分 鐘電影的滑動窗,若新請求到達時間窗,則可利用區域快 取中的資料以減少資料量,其必須從中央伺服器中傳送 出,因而減少主幹頻寬要求,此一滑動窗方法可參考Chan et al.在丨’Caching Schemes for Distributed Video Services,M 1999 IEEE International Conference on Communications (Vancouver Canada),June 1999中所述的,收到相同電影的 新請求時即增加窗的大小,因而又增加W分鐘,雖然滑動 窗方法可減少主幹頻寬要求,此一減少一般是不可預測而 且是依該窗中是否有對於相同電影到達的進一步請求,以 及此一窗增加的次數有多頻繁而定,這可參考上述Chan 的文章。因此若進一步請求在滑動窗以外到達,則不會減 少主幹頻寬。 因此上述各方法沒有一個令人完全滿意,因此期望提供 一改良的快取方法以減輕上述困難。 發明總結 本發明是根據以下觀察,藉由儲存或是從請求媒體名稱 中快取資料,其中快取資料分佈在名稱上,即可克服上述 困難,當客户請求代理伺服器部分快取之名稱時,僅代理 未快取之部分名稱需要在主幹頻寬上從中央伺服器得到。 因爲代理快取之名稱資料分佈在名稱上,所以可從中央伺 服器減少峰主幹頻寬,用以傳送部分名稱,這與滑動窗方 式完全不同,其中儲存之資料集中在窗,而且窗外不快取 -8- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意再填寫本頁) 1 -—裝 ά 479200 發明說明( 6 經濟部智慧財產局員工消費合作社印製 貝,上lii及以下方法所述之指令可從資料源或是從電 可1買取媒體中下載。例如藉由從資料源下載指令之方法’ 如直接或經由網際網路或早仅hA 』格4疋任何它種電腦網路,或是從 學或磁性儲存媒體,得到之指令或節目即可使客户裝置執 仃上述及以下(方法。因此電腦讀取媒體儲存這種指令 經由電腦鏈而傳送這種指令(或是將其傳送)之過程都 發明之範圍中。 在較佳實施例中’媒體名稱分成數個區塊而各區塊又八 成數個子區塊,其中區塊循序傳送到客户,依此,即可二 少中央伺服器傳送名稱之峰傳送位元率,有數個不同的 能實施例,因此可以在名稱上随機分佈快取之子區塊, 使用亂數產生器而從區塊序列選擇子區塊,在另一實施 中,代理伺服器從名稱之各區塊快取相同數目之子區塊 或者可沿著區塊I時序而在約每η個區塊中儲存—子 塊,其中η是正整數。在所有這種實施例中,可確定中 .伺服器之岭傳送位元率之減少,肖上述滑動窗方式不同 是,快取之子區塊最好也儲存在與時間無請求時間周 中。 根據本發明之另一特徵,使用繼續媒體快取在資料網 上以傳送媒體選擇至至少一媒體客户,資料網路可以是 際網路,企業網路或是專屬網路之網路。通常將一媒體 稱視爲一組大小不等或是相等之快取單元,當—特別媒 名稱之取用頻率增加時,與中央伺服器連接之代理词服、 即開始繼續儲存更多的快取單元,因爲取用動作是受歡^ 可 如 例 區 央 的 期 路 網 名 體 器 閱 讀 背 面 之 意I1 填 寫 頁 訂 -9- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) A7 B7 、發明說明( 經濟部智慧財產局員工消費合作社印製 的。類似的,當姐;々 田媒姐名稱之取用頻率減少時,若盔空間可 取=代理伺服器最好將其快取單元去除,以快取i他時常 ^ "冉’視訊名稱之取用頻率與其對應快取大小 配置可增加代理伺服器之快取性能,以減少代理伺服 中央词服器間在資料鏈上之頻寬要求, 網路主榦卜;^ /土 。 ’因此可大幅減少媒體服務提供者之總操 件成本。 Μ-!簡單説明^ 圖疋視Λ傳送系統包括中央及區域視訊伺服器血刑 配置。 ,、土 圖2是根據本發明特定實施例的典型視訊傳送系統。 圖3 Α是3種可能快取方法的範例,在一實施例的代理伺 服斋中快取視訊串中視訊資料的四分之一。 、圖3B是從包含1〇〇個區塊(各區塊分成1〇個子區塊)的 視訊名稱Vl的第i個視訊名稱中快取子區塊示意圖。 圖3C是時間窗…中取用概述的示意圖。 圖3 D是時間窗w中取用概述的示意圖,其中窗的開始 時間在時間的開始或是時間等於〇。 圖4疋複合快取效應的範例,其中在代理伺服器中快取 視訊串中視訊資料的增加部分,以回應根據本發明特定實 施例的視訊名稱請求。 圖5是通訊系統的典型配置其中可實施本發明。 圖6是代理伺服器的典型方塊圖包括中央伺服器及終端裝置。 -10 - 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意再填寫本頁) 丨裝 · 4 479200 A7 B7 五、發明說明(8 圖7疋代理词服器的快取記憶體的典型佈局。 圖8疋中央視訊伺服器的典型方塊圖以對應圖6的視訊 伺服器 經濟部智慧財產局員工消費合作社印製 圖9是根據本發明另一實施例的中央視訊伺服器的典型 實施例。 圖1 〇是根據本發明特定實施例的代理模組的典型流程 圖11疋根據本發明特定實施例的視訊伺服器模組的血 型流程圖。 圖1 2疋邏輯裝置範例的方塊圖,其中可實施本發明的 特徵。 奋配合附圖及詳細說明即可更明了本發明及其各個特徵及 U例,纟不同附圖中相同的數字表示本發明範圍内的類 似功能。 ’ 較佳實施例説1 雖然以下户斤述的媒體傳送系、统是根據視訊成争作號,煞 :此技術者可了解本發明也適用於聲訊成事信號或其他: =或多媒體信號。本發明的詳細説明提供許多細節以更 解本發明,惟熟於此技術者該了解本發明可以在無這 疋細即(下實施,在其他情況下並未將習用方法,程序, 二件及電路等詳細説明以避免不必要的模糊本發 欲’在説明特定實施例之前,以下要大致説明—此一般奋 施例。 二紅貝 本發明部分是指—種新的快取設計稱爲繼續視訊 快取 (請先閱讀背面之注意再填寫本頁) -丨裝 訂-· % -11 - 本紙張尺度適用中國國家標準(CNS)A4規格(21G x 297公爱) 479200 五、發明說明(9 經濟部智慧財產局員工消費合作社印制衣 m根據本發明的特定實施例,將—視訊他 =加時’代㈣服器接著繼續儲存更多的快取單元,因 =取用動作受歡迎’類似的’當視訊名稱的取用频率減 /時,代理伺服器能去除其快取單元以供快取最近取用的 視訊名%用’視訊名稱的取用頻率與其快取大小的緊宓關 係可増加代理伺服器的快取性能,因而減少主幹頻寬二要 求及總操件成本。 、根據實施例的其他細節,一旦設定快取策略,即自動決 疋$曰加或減少快取單元,因而減少管理問題,如上所述, 快取單元儲存且分佈在視訊名稱中,根據其他實施例的強 化P V c設計(在此稱爲複合P v c),當代理伺服器請求名 稱時’该設計即藉由快取額外快取單元以再減少主幹 見’以回應後績對於相同視訊名稱的多重取用,代理伺 器接著能擷取儲存的額外快取單元供後續取用用,因而 少從中央伺服器所需的資料量,及將資料從中央伺服器 送出所需的主幹頻寬。後者請求的優點是,除了快取單 外(其儲存在一般P V C設計中),現在可快速且直接從 合p v c的代理中得到額外的快取單元。 此外習用上邏輯或數位系統及/或方法能包括許多不 的元件及模組形式的不同功能,熟於此技術者可從以下 解這些,本發明的不同實施例能包括元件及/或功能的 同合併,本發明的不同實施例能包括一些執行動作或 -12 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意^9^再填寫本頁) ;裝 訂· 頻 減 元 複 同 了 不 步 Φ 479200 A7I Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 479200 A7 --------- B7_ .___ V. Description of the Invention (2) Internet access to capture various information, while requesting more information, and requesting the use of various information Information is sent to them via the Internet. Currently, the most commonly requested information on the Internet is text information, images and graphics, video and audio. Continuous information such as continuous video, audio, audiovisual or other multimedia works (hereinafter collectively referred to as media) is one of the frequently requested network resources, so often there are thousands of simultaneous requests for access to a popular video, audio or For audiovisual programs, under a certain network bandwidth of the data network, the program's real-time transmission and the quality of service of the network (Q0S) are guaranteed to become extremely important. It depends on the central media server to simultaneously access all It is not practical to request services, so a multifunctional media transmission system is needed, which can respond to thousands of access requests at the same time, and the impact on service quality is minimized. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs due to video compression, network structure, and overall computer functions and other technological advances, making real-time video (V0D) possible. In recent years, the Internet and the World Wide Web have become closer together. Accelerate the implementation of VOD services, such as RealNetwork's Real Media and Microsoft's NetShow Service, which have successfully transmitted low-quality audio / video over the Internet. These services have been used in many aspects of multimedia applications including ( (But not limited to this) real-time news, distance learning, company training and music playback, more people go online to find the information they want, which triggers sending more information to the website 0 in order to attract more people Website, so the company designed the webpage with a lot of media information, especially the continuous media information, but this continuous media features such as large storage requirements, high network adjustment requirements, and real-time transmission China National Standard (CNS) A4 specification (21〇X 297 public love) 479200 A7 --------- B7_____ V. Description of invention (3) The media distribution service on the Internet has become a big problem, which is more pragmatic for media with high video content. Real-time video server ^ Solution: A good way to solve this problem, but not without its disadvantages.-Popular website: There may be thousands of accesses at the same time, so it is an urgent matter for the video server system to handle all these requests and send the video to the end user (with quality assurance). One way to deal with server scalability is to The number = machine aggregation technology is used to establish parallel video servers, but this method is not suitable for ^ wide network access, so the central video server cannot serve all the peer requests. Another method is to add video transmission services to the existing Internet access model. The popularity of network services has created many new business opportunities. Internet service providers (ISPs) are one of them, which provides the necessary structure. Equipment and services so that users can easily and quickly connect to the Internet. In order to provide better services to users, Is ρ company provides services to users in a geographical area. All IS companies are based on Capacity and expensive backbone network to connect, all ISP companies work together to become a unit, the user's regional IS site is responsible for receiving the user's access request, and if necessary, transmitting the request to the original website as a proxy to receive Website responses, and sending request data back to the requestor 'Sending requests to the original website and receiving its responses will make the trunk too busy. When too many requests pass through the backbone network, a higher backbone bandwidth is required. This means that [S ρ increases the operating cost. In order to solve this problem, the Internet has developed a network caching strategy. The 'Internet proxy server is installed on the SP site and serves as a caching engine to store frequently accessed web pages (which are related to the user area). Next time, When using the cache-6- This paper size is applicable to the Chinese National Standard (CNS) A4 specification (21 × 297 mm) (Please read the precautions on the back before filling out this page) When printing 479200 A7 B7 V. Description of the Invention (4) pages, the proxy server directly sends the web page from its regional cache (not from the original web server). This method can reduce the user's browsing response time ( Because the content comes from a regional proxy server), it can also reduce the trunk's requirements for transmitting content, thereby reducing operating costs. The online cache model is also suitable for large-scale end-to-end video transmission server. The Intellectual Property Bureau of the Ministry of Economic Affairs, employee consumer cooperatives has developed many web proxy server products. The cache components mainly refer to HTML pages and their web pages. Embedded image, most of the conventional network agents generally do not refer to video agents. Other early researches have pointed out QoS (quality of service) issues on end-to-end video transmission, but still have not pointed out video agents. Some methods have been proposed in video The hierarchical design reduces Wang Qian ’s bandwidth requirements for distributed multimedia transmission. In these systems, when the user requests a special video, the video content of Shao Fen is sent from the central video server through Wan connection, and Part of the video content is sent in series from the regional proxy server. Both types of data are entered into the user's playback station. If most of the information is sent from the regional server, the data will be sent from the main channel. Less, thus reducing backbone bandwidth requirements, so the previous video grading method was to pre-fetch a preset amount of video data and store it in Proxy server, the proxy server stores the default video content based on the retrieval method generated in the previous overview cycle, instead of the current retrieval action. Some early designs can only improve the backbone use of encoding into VBR (variable bit rate) format content. As for CB R (locating bit rate) encoding video content, some conventional methods will not reduce any W AN bandwidth requirements. Another method is to cache the data of a sliding window for each display video. It is printed as 479200 by the Consumer Property Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. A7 B7_; __V. Description of the invention (5) The video can be stored in the proxy server. , So if the movie requested by the user reaches the proxy server, the proxy server stores a sliding window of W minutes of movies in its regional cache. If a new request arrives at the time window, the data in the regional cache can be used to reduce the amount of data , Which must be transmitted from the central server, thus reducing the backbone bandwidth requirements. For this sliding window method, refer to Chan et al. In 'Caching Schemes for Distributed Video Services, M 1999 IEEE International Conference on Communications (Vancouver Canada) As described in June 1999, when a new request for the same movie is received, the window size is increased, thus increasing W minutes. Although the sliding window method can reduce the backbone bandwidth requirements, this reduction is generally unpredictable and should be Whether there are further requests in the window for the same movie to arrive, and how often this window is added, See Chan's article above. Therefore, if further requests arrive outside the sliding window, the backbone bandwidth will not be reduced. Therefore, none of the above methods is completely satisfactory, and it is desirable to provide an improved caching method to alleviate the above difficulties. Summary of the Invention The present invention is based on the observation that the above difficulties can be overcome by storing or caching data from the requested media name, where the cached data is distributed on the name, when the client requests a proxy server to partially cache the name , Only the part of the name that is not cached by the agent needs to be obtained from the central server on the backbone bandwidth. Because the name data of the agent cache is distributed on the name, the peak backbone bandwidth can be reduced from the central server to transmit part of the name. This is completely different from the sliding window method, where the stored data is concentrated in the window, and the cache is not cached outside the window. -8- This paper size is in accordance with Chinese National Standard (CNS) A4 (210 X 297 mm) (Please read the note on the back before filling out this page) 1 -—Installation 479200 Invention Description (6 Employees of Intellectual Property Bureau, Ministry of Economic Affairs Consumer cooperatives print shellfish, and the instructions described above and below can be downloaded from data sources or purchased from media. For example, by downloading instructions from data sources, such as directly or via the Internet or early Only hA "Grid 4" Any other computer network, or school or magnetic storage media, the instructions or programs can make the client device perform the above and below (methods. Therefore, the computer reads the media to store such instructions via The computer chain and the process of transmitting such instructions (or transmitting them) are all within the scope of the invention. In the preferred embodiment, the 'media name is divided into several blocks and each block is 80% Sub-blocks, in which the blocks are sequentially transmitted to the customer, so that the peak transmission bit rate of the name can be transmitted by the two central servers, there are several different embodiments, so the cached children can be randomly distributed on the name Block, using a random number generator to select sub-blocks from the block sequence. In another implementation, the proxy server caches the same number of sub-blocks from each block of the name or may Approximately every n blocks are stored—sub-blocks, where η is a positive integer. In all such embodiments, it can be determined. The reduction of the bit rate of the server ridge transmission, which is different from the above sliding window method, is cache The sub-blocks are preferably also stored in the time period with no request time. According to another feature of the present invention, the continuous media cache is used on the data network to transmit media selection to at least one media client. The data network may be the Internet Network, corporate network, or private network. A medium is usually referred to as a set of cache units of different sizes or equals. When the frequency of the use of special media names increases, it interacts with the central server. connection I will continue to store more cache units, because the access action is popular ^ You can read the meaning on the back, such as the central road network name body device I1 Fill in the page order -9- This paper size Applicable to China National Standard (CNS) A4 specifications (210 X 297 mm) A7 B7, invention description (printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Similarly, when the sister; Putian matchmaker name is used less frequently If the helmet space is available = it is better to remove the cache unit of the proxy server in order to cache it. ^ "Quoting the frequency of the video name and its corresponding cache size configuration can increase the cache of the proxy server Performance, in order to reduce the bandwidth requirements on the data chain between the central server of the proxy server, the network backbone; ’As a result, the total operating cost of a media service provider can be significantly reduced. Μ-! Brief explanation ^ Figure 疋 Λ transmission system includes central and regional video server blood penalty configuration. Fig. 2 is a typical video transmission system according to a specific embodiment of the present invention. Figure 3A is an example of three possible caching methods. In one embodiment, the proxy server caches a quarter of the video data in the video string. 3B is a schematic diagram of caching the sub-blocks from the i-th video name of the video name V1 containing 100 blocks (each block is divided into 10 sub-blocks). FIG. 3C is a schematic diagram of an overview taken in a time window. Figure 3D is a schematic diagram of the overview taken in the time window w, where the start time of the window is at the beginning of time or the time is equal to zero. Figure 4: An example of a composite cache effect, in which an increased portion of video data in a video stream is cached in a proxy server in response to a video name request in accordance with a particular embodiment of the present invention. Fig. 5 is a typical configuration of a communication system in which the present invention can be implemented. FIG. 6 is a typical block diagram of a proxy server including a central server and a terminal device. -10-This paper size is in accordance with Chinese National Standard (CNS) A4 (210 X 297 mm) (Please read the note on the back before filling in this page) 丨 4 479200 A7 B7 5. Description of the invention (8 Figure 7 疋The typical layout of the cache memory of the proxy server. Figure 8: A typical block diagram of the central video server is printed to correspond to the video server of Figure 6. The Intellectual Property Bureau of the Ministry of Economic Affairs, Employee Consumption Cooperative, Figure 9 A typical embodiment of the central video server of the embodiment. Fig. 10 is a typical flowchart of an agent module according to a specific embodiment of the present invention. 11) A blood type flowchart of a video server module according to a specific embodiment of the present invention. 1 2 疋 A block diagram of an example of a logic device, in which the features of the present invention can be implemented. The invention and its various features and U examples can be more clearly understood by working with the drawings and detailed description, and the same numbers in different drawings represent the invention Similar functions within the scope. 'Preferred embodiment said 1 Although the following media transmission system, the system is based on video contention, it is obvious to those skilled in the art that the present invention is also applicable to audio Event signal or other: = or multimedia signal. The detailed description of the present invention provides many details to better understand the present invention, but those skilled in the art should understand that the present invention can be implemented without this detail (in other cases, and in other cases and The conventional methods, procedures, two pieces, and circuits have not been described in detail to avoid unnecessary obscuration. Before describing specific embodiments, the following is a general description—this general embodiment. Two red shells The present invention refers to part —A new type of cache design is called Continue Video Cache (please read the note on the back before filling this page)-丨 Binding- ·% -11-This paper size applies to China National Standard (CNS) A4 (21G x 297) Public love) 479200 V. Description of the invention (9 Printed clothing by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. According to a specific embodiment of the present invention, the video server will be added to the server and then continue to store more caches. Unit, because the fetch action is popular 'Similar' When the fetch frequency of the video name is reduced by /, the proxy server can remove its cache unit for caching the most recently fetched video name The tight relationship between the frequency and its cache size can increase the cache performance of the proxy server, thus reducing the backbone bandwidth requirement and the total operating cost. According to other details of the embodiment, once a cache strategy is set, it is automatically determined. Add or reduce the cache unit, thereby reducing management problems. As mentioned above, the cache unit is stored and distributed in the video name. According to the enhanced PV c design of other embodiments (herein referred to as composite PVC), when When the proxy server requests the name, 'the design is to cache additional cache units to reduce the number of trunk views' in response to multiple results for the same video name, the proxy server can then retrieve the stored additional cache units For subsequent retrieval, the amount of data required from the central server is reduced, and the backbone bandwidth required to send data from the central server is reduced. The advantage of the latter request is that, in addition to the cached order (which is stored in the general P V C design), additional cached units can now be obtained quickly and directly from the agents that integrate p v c. In addition, conventional logic or digital systems and / or methods can include many different functions in the form of components and modules. Those skilled in the art can understand these from the following. Different embodiments of the present invention can include components and / or functions. In combination, different embodiments of the present invention can include some execution actions or -12 This paper size is applicable to Chinese National Standard (CNS) A4 (210 X 297 mm) (Please read the note on the back ^ 9 ^ before filling this page ); Stapling · Frequency reduction is not the same Φ 479200 A7

經濟部智慧財產局員工消費合作社印剧衣 驟,其與任何特定範例所述的 6 k , 斤不同’本發明的不同f 犯例此包括一些群或是大元件中 ^ ^ 4s\ ΡΪΤ ΑΑ -7- 、’、疋件,其與任何特定 範例所述的不同。爲了説明 明,立包括咩夕Τ门从、 疋以系統來説明本潑 人τ、、„丄 干及新兀件與習用元件的新翻 云,不该舲本發明僅限於這收組 .^ ^ ϋ,其包括本發明的任甸 -件中峨元件,本發明的功能特徵(如以下所述)可:; 用任何適當的實施環境或程式語言(C++,c〇b〇i, Pascal,Java,Java_script等)而達成,本發明所述的所有 公告,專利,及專利申請案都在此供參考。 因此本發明的特定特徵提供一種成φ視訊/聲訊信號, 不論終端裝置中預設那-種機器都可以在任—種作業系統 下操件的各種視訊終端裝置中播出。 在特定實施例中,本發明是指快取方法及系統其適用於 通訊資料網路(有線網路,區域網路,它種專屬網路的網 路及網際網路)上提供多媒體成串。 本發明於説明時用到一些程序,步驟,邏輯區塊,處 理,及其他符號其與資料處理裝置類似。熟於此技術者於 使用這些方法及説明後可更佳了解本發明,以下所述的方 法及系統是一種一貫性的處理序列或步驟以產生期望的結 果。這些步驟或過程需要實體量的實體操件,通常(雖然 不是必要)這些量是電的信號形式其能在電腦系統或電子 十舁裝置中儲存’傳送’合併’比較,顯示及控制。它有 時k供便利(主要是因爲共同使用的原因)以使這些信號成 爲么元’値,元件,符號,操件,信息,名詞,數字等。 -13 本,,式張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意再填寫本頁) -裝 訂: m A7 B7 五、發明說明( 經濟部智慧財產局員工消費合作社印制衣 、g职、人ϋ有的&些類似名詞都與適當的實體量相關,且僅 太旅;^些里。除非特別説明,否則在以下説明中可了解 -^使用以下名詞來探討,如處理,計算,鐘定或顯示 變器:=:算裝置:動作及過程,以控制及經由常駐轉 ^ 、匕的暫存森及記憶體中的實體量所表示的資料 轉成類比輸出信號。 只打 :疋視訊傳送系統包括中央及區域視訊伺服器的典型 ^此杈型適用於網際網路(及網際網路2 )及企業網路 此模型中,區域服務站台安裝—代理視訊词服器 ^矛力視訊傳运服務至其附近用彳,區域視訊㈣器經由 f網路而與中央視訊飼服器連接,中央視訊㈣器是具 :有視訊名稱拷貝的視訊儲存區,代理視訊伺服器快取 時常取用的區域視訊内容且當它沒區域拷貝可用時僅從 央視訊伺服器得到視訊名稱,快取原理與網路取用服務 視訊傳送服務相同,惟’在考慮大儲存要求的本質,高 路頻寬要求及資料傳送的時間因素等之下,視訊内容所 的挑戰更大’南視訊品質儲存量如電影專家群⑽即巧 ^PEG-2格式)的使用使此挑戰更難。—種不正確的快取 策會浪費寶貴的資源如主幹頻寬及儲存空間。如何設計 良好的視訊代理設計以找出上述問題的解答是揭示本發 特定實施例的目的。 通訊系統概述 圖5是通訊系統的典型配置其中可實施本發明。中央 服器1〇2及視訊/聲訊資料庫1〇4形成—視訊/聲訊源, 中 及 網 需 決 出 明 伺 包 (請先閱讀背面之注意再填寫本頁) 裝 訂: Φ -14- 本紙張尺度適用中國國家標準(CNS)A4規格(21〇 x 297公爱) 479200 五、發明說明(12 經濟部智慧財產局員工消費合作社印製 括可依需求而取用的視訊/聲訊構。 訊檔或名稱是指任何視 *月中,視訊/聲 段,其可壓縮成任何二作影及/或視訊/聲訊片 <仕彳了袼式。一種典型格式是mpe ^視訊檔,需要mpeg觀^客户㈣,其可在個^ :種:理器速度及内部記憶體來執行。聲訊的另 技衔::=二式是Μ隱1聲訊層3(ΜΡ3),其爲數量 β式k壓鈿聲晋序列成爲極小的檔(約爲原始栌大 小的12分之一),同時在透過客户程式(播放機)而播^時 留聲晋品質的原音,該程式可從網站下載。惟該注音 減檔的正確格式不影響本發明的操件,將該注意及了: :疋’本發明適用於視訊或聲訊的任何檔以及其他多媒體 二料,包括(但不限於)二位元資料及檔超本文檔及片段 α,貝料網路1 0 6最好是資料網路主幹如大的傳送線,在區 域位準中,主幹是線或一組線使區域網路接到廣域網路, 或是在區域網路中以迅速擴張距離(如在建築物之間),在 網際網路或其他廣域網路中,主幹是一組路徑供區域網路 接到長程連線,與資料網路Α 1〇6連接的是2個代表性代 理词服器1 0 8,1 1 〇,經由資料網路1 1 2,1 1 4而分別與 各自服務代表性終端裝置116-119。 貧料網路1 1 2,1 1 4 一般是網際網路,區域網路或是電 話/有線網路,以便終端裝置能接收視訊檔,終端裝置包 括(但不限於此)多媒體電腦(如丨丨6,;[〗9 ),網路電視或 是其他視訊/聲訊播放機(如117,U8),通常終端裳置配 訂 螓 -15-Employees of the Intellectual Property Bureau of the Ministry of Economic Affairs 'consumer cooperatives printed dramas, which are different from the 6k described in any particular example.' Different f examples of the present invention include some groups or large components ^ ^ 4s \ ΡΪΤ ΑΑ- 7-, ', files, which are different from those described in any particular example. In order to clarify, including including the door of the door and the system to explain the new 泼, 丄, 丄, and new components and custom components of the new cloud, the invention should not be limited to this collection. ^ ^ ϋ, which includes the Rendian-Ezhong component of the present invention, the functional features of the present invention (as described below) may :; use any suitable implementation environment or programming language (C ++, coboi, Pascal, Java, Java_script, etc.), all the announcements, patents, and patent applications described in the present invention are here for reference. Therefore, the specific feature of the present invention provides a video / audio signal, regardless of the preset in the terminal device. All kinds of machines can be broadcasted in various video terminal devices operating under any kind of operating system. In a specific embodiment, the present invention refers to a caching method and system, which is applicable to a communication data network (cable network, area) Internet, which is a kind of dedicated network (Internet and Internet), provides multimedia strings. The present invention uses some procedures, steps, logical blocks, processing, and other symbols in the description, which are similar to data processing devices. Familiar with this The surgeon can better understand the present invention after using these methods and descriptions. The methods and systems described below are a consistent processing sequence or steps to produce the desired results. These steps or processes require physical quantities of physical exercises, Usually (though not necessarily) these quantities are in the form of electrical signals which can be stored, transmitted, combined, compared, displayed and controlled in a computer system or electronic device. It is sometimes used for convenience (mainly for reasons of common use) ) In order to make these signals become “Yuan Yuan”, components, symbols, operating parts, information, nouns, numbers, etc. -13 copies, the standard scale is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 mm) ( Please read the note on the back before filling in this page)-Binding: m A7 B7 V. Description of the invention (printed by the consumer co-operatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, printed clothing, job titles, and some similar terms are appropriate The physical quantity is related, and it is only too far; ^ some. Unless otherwise specified, you can understand in the following description-^ use the following nouns to explore, such as processing, calculation, time setting or display change : =: Computing device: Actions and processes to control and convert the data represented by the permanent transfer ^, the temporary storage of the dagger, and the physical quantity in the memory into analog output signals. Only hit: 疋 The video transmission system includes the central and Typical of regional video server ^ This type is applicable to the Internet (and Internet 2) and corporate networks. In this model, the regional service station is installed—proxy video server ^ Spear video delivery service to its vicinity For example, the regional video server is connected to the central video feeder via the f network. The central video server is: a video storage area with a copy of the video name, and the proxy video server caches the frequently accessed local video content. When it has no regional copy available, it only obtains the video name from the CCTV server. The caching principle is the same as the network access service video transmission service, but 'the nature of large storage requirements, high bandwidth requirements, and data transmission Due to time and other factors, the challenge of video content is greater. The use of South Korea ’s quality storage capacity (such as the film expert group (eg, PEG-2 format) makes this challenge even more difficult. — An incorrect caching strategy will waste valuable resources such as backbone bandwidth and storage space. How to Design A good video agency design to find the answers to the above questions is the purpose of revealing a specific embodiment of the present invention. Communication System Overview Figure 5 is a typical configuration of a communication system in which the present invention can be implemented. Central server 1102 and video / audio data base 104 are formed—video / audio source, China and the Internet need to determine the package (please read the note on the back before filling this page) Binding: Φ -14- Paper size applies Chinese National Standard (CNS) A4 specification (21x 297 public love) 479200 V. Invention description (12 Employees' Cooperatives of Intellectual Property Bureau of the Ministry of Economic Affairs prints video / audio structures that can be accessed on demand. The file or name refers to any video / video, video / sound segment, which can be compressed into any two filming and / or video / sound film < official style. A typical format is mpe ^ video file, which requires mpeg View customers, it can be implemented in various types: processor speed and internal memory. Another technical title of audio :: = type 2 is M hidden 1 audio layer 3 (MP3), which is the number β type k The vocalization sequence becomes a very small file (approximately one twelfth the size of the original vocalization), and at the same time, the original quality sound is preserved when playing through a client program (player), which can be downloaded from the website. The correct format of the Zhuyin downshift does not affect the operating elements of the present invention, and it should be noted that: : 疋 'The present invention is applicable to any file of video or audio and other multimedia materials, including (but not limited to) binary data and files exceeding this document and fragment α. The material network 1 0 6 is preferably a data network The trunk of a road is like a large transmission line. In the regional level, the trunk is a line or a group of lines that connects the local area network to the wide area network, or in the local area network to expand the distance quickly (such as between buildings). In the Internet or other wide area networks, the backbone is a set of paths for the local network to receive long-range connections. The data network A 106 is connected to two representative proxy servers 1 0 8, 1 1 〇, through the data network 1 12, 1 1 4 and respectively serve the representative terminal device 116-119. The lean network 1 1 2, 1 1 4 is generally the Internet, LAN or telephone / Wired network so that terminal devices can receive video files. Terminal devices include (but are not limited to) multimedia computers (such as 丨 丨 6, [〗 9), Internet TV or other video / audio players (such as 117, U8), usually the terminal is equipped with order 螓 -15-

本紙張尺㈣时關家標準(CNS)A4規格(210 X 297 公iT 479200The size of this paper is CNS A4 (210 X 297 male iT 479200)

有應用或功能以執行及顯示收到的視訊樓,如一種受歡迎 的應用是微軟公司的WIND〇ws 98中提供的鹏g播放 機,,MPEG視訊樓從一代理词服器成h人終端^ 時’ f由執行多媒體電腦中的MPEG播放機,即可在電腦 顯示幕上顯示視訊檔。 局了接收期望的視訊,一終端裝置在請求之下傳送,包 括期望視訊的名稱’此外若視訊服務僅允許授權取用則該 請求包括用户豸,收到請求時,若選擇的視訊在其中則代 理伺服器會先檢查其快取,㈣時由請求f理員記錄請求 (S ) k擇的視訊會當成成申視訊而傳送到終端裝置(若 整個視訊在快取中),否則若視訊的一些單元在其快取記 憶體中’㈣理词服器會繼續傳送請求到剩餘視訊的視訊 中央伺服器1 0 2中’或是若沒有視訊單元在其快取記憶體 中’則傳送請求到整個視訊。 視訊快取的一般策略 經濟部智慧財產局員工消費合作社印製 局了説明目的,在一實施例中,可參考圖丨的結構而了 解端對端視訊傳送,例如一終端用户使用其多媒體個人電 腦(P c或是任何其他適當的終端裝置)以傳送視訊播放請 求,涊凊求先傳送到區域代理視訊伺服器,若代理伺服器 具有請求視訊的整個内容,則它能在不傳送該請求到遠地 中央視訊伺服器之下立即服務該請求。這稱爲快取命中其 表示不必消耗主幹資源,換言之,若區域代理伺服器完全 沒有視訊拷貝,則會將該請求再引入遠地視訊伺服器,視 訊資料會成攀的從中央伺服器傳回區域代理伺服器,且接 -16 - 本紙張尺度適用中國國家標準(CNS)A4規格(21〇 X 297公釐) 479200 A7 B7 五、發明說明( 經濟部智慧財產局員工消費合作社印製 :這稱爲快取失誤,在此例中,視訊資 ^在王幹連線域動,快取策略的目的是改良快取命中比 (其包括區域代理飼服器中部分可利用視訊的快取命”, 接耆導出快取失誤比而節省主幹資源"淮,在代理伺服 器儲存整視訊拷貝需要區域儲存區,因此若需要儲存許多 視訊名稱,則必須裝設大量的記憶體而這是昂貴的,因此 根據本發明若部分的熱門視訊名稱儲存在代理伺服器,則 接著快取部分可與中央伺服器的視訊名稱失誤部分^併二 該合併視訊名稱接著成串送入請求者,因此根據本ς明 特欲,於快取視訊名稱時部分的名稱在視訊名稱中分佈 再快取。因爲這些部分在視訊名稱中分佈,所以較少的 訊貝料為要從中央伺服器擴取以及與快取部分合併,以 在不中斷之下傳送到請求者,較佳的,儲存的部分視訊 料會均勻的在名稱上分佈,以便確定可減少要求的峰位 率,用以從中央伺服器傳送該失誤部分如以下所述。 快取效率是由以下二個因素決定,即主幹使用的極小 代理词服器中區域儲存量的極小,因此若在代理词服器 取的視訊名稱不是用户時常請求的,這表示較少使用的 稱佔據區域儲存處,同時需要從中央伺服器擷取熱門〜 稱,因而減少快取效率,因此影響快取效率的重要因素是 用户取用概述,知道用户取用樣式後即能設計出較佳的 取策略,因此如何正確的預測用户取用行爲變成改良區 命中比的重要事項,先前研究顯示用户取用概述在類 Zipf的分佈之後,惟此分佈僅能作爲一般指南使用,並 的 視 資 元 及 快 名 名 快 域 似 不 (請先閱讀背面之注意#^再填寫本頁) 訂· ¾ 17- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) A7 五、發明說明(15 ) 適用於其他情況。例如在同一天不同的觀眾群在不同時段 觀看不同的廣播節目,在單一天有數種不同的取用概述, 此外取用概述一般表示過去行爲且假設近期會遵守該趨 勢’概述更新周期愈長,則取用概述愈不正確,不可能事 先對於區域词服器上快取的視訊名稱作取用概述的預測及 準備。 因此本發明的特別實施例中不想預測取用概述,反而每 吾取用行爲改變時即調整快取策略,因此本發明確定在每 時刻’快取策略會與目前取用樣式同步,且調整快取參 數以改良快取命中比,因爲取用行爲會隨時改變且改變大 小’在操任何類介入之下根據本發明特定實施例的快取策 略凋整疋完全自動的,這對於系統管理者而言是一大優 點0 見訊伺服器快敌样^ 經濟部智慧財產局員工消費合作社印製 在根據本發明特定實施例的繼續視訊快取(PVC )策略 中,將一視訊名稱當成一組大小相等(或是如下所述可變 大】)的f夬取單元來處理,當一特別視訊名稱的取用頻率 增加時,代理伺服器接著繼續儲存更多的快取單元,因爲 其取用動作爻歡迎,類似的,當視訊名稱的取用頻率減少 時,代理伺服器能去除其快取單元以供快取最近取用的視 訊名稱用,視訊名稱的取用頻率與其快取大小的緊密關係 可增加代理伺服器的快取性能,因而減少主幹頻寬的要求 及總操件成本,另-優點是_旦設定快取策略,即可自動 增減決策,如以下所述。 -18- 479200 五、發明說明(16 ) 雖然爲了易於説明而用分成一組含相同快取單元大小的 視訊名稱來説明PVC設計,但是此限制是有彈性的。 P v C設計也將視訊名稱分成一組含不同大小的快取單 元,如MPEG格式中的I,B及p框,或是在以下所述的並 他實施例中。在該等其他設計中,Pvc設計極適於視^ 編碼成爲多層格式及/或可變位元率格式如以下所述。 在根據本發明另一特徵的視訊傳送模型中,區域代理伺 服备是指將視訊資料送入用户站台,資料可以是僅來自區 ,代理(作爲整個可用視訊的快取命中),僅來自遠地伺月I 器(作爲區域代理的快取失誤),或是來自區域及遠地站台 (作爲區域代理伺服器的部分可用視訊的快取命中)。圖^ 説明部分可用視訊内容的快取命中設計的視訊傳送,在圖 中,區域代理傳送視訊内容到客户的多媒體]?(:,區域代 理伺服器擷取區域的資料段〇,8,丨6及來自遠地中央伺 服器的資料段的其他部分(如資料段i,2,3,4,$, 6 ’ 7 ’ 9 ’ 1 〇等),此例的目的只是説明,在以下説明中 提供根據本發明的額外實施例及改良的進一步敘述。 圖2的範例也揭示代理視訊伺服器的資源要求,代理伺 經濟部智慧財產局員工消費合作社印製 服器可讀取/窝入硬碟(或它種大量記憶體)的快取資料, 因此硬碟速度及其儲存容量影響快取性能,代理伺服器也 需要主記憶體空間以緩衝視訊資料,類似的,主幹連線中 的網路頻寬及用户連線決定多少資料能流過其中,以下要 説明上述資源對於快取性能的影響及根據本發明其他實施 例的其他特徵。 本紙張尺度適财酬家鮮 -19- 479200 A7 B7 五、發明說明(17 視訊擷取模刚 在足義繼續視訊快取(PVC)之前,先考慮視訊擷取模 型’視訊擴取視爲一種周期性擷取過程,在長度P的各周 期中’系統類取客户的資料量的pxri,以便在資料率ri之 下請求視訊名稱、,這可確保在播放機需要之前即準備好 貝料’根據本發明,將視訊名稱Vi視爲一序列視訊區塊: 〇 1 V:· ’ K ’ ·.· V,·,其中niXp表示視訊長度,而各資料區塊 具有pXq的大小,爲了易於了解本發明而考慮一種雙緩衝 設計’其中一緩衝藉由網路模組而將要傳送到用户的資料 儲存’同時使用其他緩衝以儲存區域硬碟或是遠地中央伺 服备的次一資料區塊,根據本發明的p V c定義的範例是 根據以上假設。 訊快取p v c的一般説明 根據本發明的繼續視訊快取(pvc)是一種快取設計以決 足要快取(或取代)的視訊内容量及視訊資料部分,每當要 求貝料快取或取代時,根據本發明的代理伺服器即遵守 p v c設計以識別視訊資料部分以達成快取動作或是操 件。 p v c定義稱爲N的參數以指定快取或取代單元,每當代 理词服器要求快取視訊名稱(或是快取資源的取代)時,視 訊内容的1/N (N 2 1 )部分即區域的儲存(或取代),換言 I ’當N等於2時,代理伺服器在各快取操件中儲存或是 取代視訊名稱的一半,在N等於1的特例中,快取單元是 整個視訊名稱,除了用以決定快取大小以外,N値也定義 20- 本紙張尺度適用中關家標準(CNS)A4規格(210 X 2Θ7公爱) 請 先 閱 讀 背 面 之 注 意 事蝙 項1 再 填 I裝 頁 訂 蟪 經濟部智慧財產局員工消費合作社印製 479200 A7 五、發明說明( 經濟部智慧財產局員工消費合作社印製 快取(取代)視訊資料的正確部分,根據本發明特定實施例 的P V C設計最好將資料的1/N部分均勻分配在視訊名稱 中〇 雖然爲了實際原因最好將資料的1/N部分均勻配置在視 訊名稱中,但這不是必要的。因此快取的子區塊可隨意分 佈在名稱上如使用亂數產生器而從區塊序列中選擇子區 塊。在以下的另一實施例中,用代理伺服器而從名稱的各 區塊快取相同數目的子區塊,或者在區塊的時間序列中於 約每η個區塊儲存一個子區塊,在所有的這些實施例中, 可確定中央伺服器的峰傳送位元率的減少。 根據本發明的另一實施例,考慮圖3 A (其説明3種可 解釋)即可更了解均勻配置,在圖3 A的上面部分顯示一 訊名稱的原始傳送時間,其以資料率r來傳送。在每一 1 段(即P),代理伺服器必須傳送rXp的資料量以維持服務 品質,對於以N = 4在P V C下操件的代理伺服器而言, 訊的快取操件將1/4的資料部分置入區域代理。 圖jA所π的方法A藉由在垂直方向記下資料區塊而 施資料的均勻要求,圖中的代理快取區塊〇,4,8,即 取每4個資料區塊之一,在次一快取命中中,中央伺服 遵守原始時間而在主幹網路上傳送剩餘的區塊。中央伺 洛在代理伺服器具有儲存的區域資料期間無動作,因此 主幹連線的峰頻寬具有與快取失誤相同的値,換言之, 法A不會減少中央伺服器的峰頻寬要求,雖然它可減少 均頻寬要求。 能 視 時 視 實 快 器 方 平 (請先閱讀背面之注咅?再填寫本頁) 裝 4 21 - 479200 五、發明說明(19 經濟部智慧財產局員工消費合作社印製 圖3 A所示的方法 間而減少峰資料率,換士使=延伸方法八中的傳送時 3個示-時段傳…中央飼服器將原始預計要在 技術者可了解七? 資料分成4個較小區塊。熟於此 T者了了解如何實施它,所以不在此説 也可改蠻偉详Hi _X ^服态段傳I可接著將4個小區塊在4個對應時二:少,資料率達25%沿至。·75Γβ惟因爲資料的傳)送 .^預疋早,所以代理伺服器需要額外的緩衝空 ^ /2)ΧΓΧΡ而不是方法A中的2xrxp)以快取先到達的 ,因而增加在代理所需的緩衝空間。 圖〇 A所7F的万法c是在水平方向記下資料區塊,在快取 二丄代理伺服器維持區域視訊名稱的每一資料區塊的1/4 P刀此方法可減少主幹連線的峰資料率從r變成〇 75i<, 同時k衝要求維持在(2xrxp)的値,這是原始(未快取)時 間要求的,根據本發明特定實施例的Pvc設計採用 二JUi理視訊伺服器模剷 如本發明的特定實施例所述,假設在視訊名稱%中的 訊區塊總數η^Ν的整數倍(即ni>Niniin〇dN=0),至 A的各j個2區塊,其可視爲一組含N個相等大小的子區塊 以4 ’ vf7 ’ · · · vfy表示,至於各視訊快取命中或取代, 鎖足取用視訊段的資料區塊的次一可用子區塊,例如傻 N= 1 〇的P V C設計,含1 〇 〇個區塊的視訊名稱Vi標記在 間序列,其中其以νί,,…,v)。。傳送。這10〇個區塊要 -22- 請 先 閱 讀 背 Sj 之 注 間 資 1·! f *Ϊ裝 頁 訂 法 視 於 則 設 時 紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐)' ------ 鼇 479200 五、發明說明( 20 經濟部智慧財產局員工消費合作社印製 以循序方式傳送到客户或是用户, 區域代理伺服哭了产甘U而可依此排列Μ鼠設 佳5 1 憶體中具有視訊名稱、,較 存#的資料(其中u^100)在觸發E域代理伺服器以错 )彺具E域硬碟中。如圖3b* :二圖3B是含100個區塊的視訊名稱v,示意圖,其中 之塊分成!0個子區塊以說明本發明的較佳實施 圖公中,當請求視訊名稱'時,區域代理伺服 取記憶體中不具有此名稱,因而開始將部分名稱儲 ==己憶體中。根據上述方法C,最好儲存1⑽個 °口 冲刀,參考圖3 B,接著儲存1 〇 0個子區塊,J· 從1到100,如圖3Β所示,在虛線132上用編號 T,112, ,llj,...,iU〇來表示,該等子區塊總稱爲單元132a。 ^再請,相同的名稱Vi時,區域代理伺服器即決定增加名 %Vi的量,其藉由儲存1 〇〇個子區塊中的另一單元134a 而快取’在快取記憶體中j從1到1 0 0如圖3 B虛線1 3 4上 的子區塊所示,相同視訊段的另一快取命中會將加 區域硬碟,假設資料g,#仍在區域伺服器中。接著 績E直到快取所有的名稱v i。 因此若區域代理伺服器具有足夠的資源且沒有發生快 取代’則區域代理伺服器決定增加視訊名稱Vi的子區塊 其在該名稱的各個額外請求之後即儲存,接著在1 0次 求視訊名稱Vi後,區域伺服器會具有Vi的整個内容。以 公式定義上述圖3所示的P V C快取效應,在C ( v i )表示 面 入 繼 取 請 下 視 (請先閱讀背面之注意再填寫本頁) 丨裝 . 4 -23 本紙張尺度適用中_家標準(cns)A4規格⑵◦ χ视公髮) 五、發明說明( A7 B7 訊名稱Vi的目前區域快取内容之下 C(v,)There is an application or function to execute and display the received video building. For example, a popular application is the Peng g player provided by Microsoft's WINDOWs 98. The MPEG video building transforms from a proxy server to a human terminal. ^ 'F by running the MPEG player in the multimedia computer, you can display the video file on the computer screen. After receiving the desired video, a terminal device transmits it under the request, including the name of the desired video. In addition, if the video service only allows authorized access, the request includes the user. When the request is received, if the selected video is among them The proxy server will check its cache first, and the video selected by the requester to record the request (S) will be sent to the terminal device as the application video (if the entire video is in the cache), otherwise if the video Some units' cache server will continue to send requests to the video central server 102 of the remaining videos' in their cache memory or if there are no video units in their cache memory 'then send the request to The entire video. General strategy for video caching. The Intellectual Property Bureau of the Ministry of Economic Affairs, the Consumer Cooperative Cooperative Printing Bureau has explained the purpose. In one embodiment, you can refer to the structure of Figure 丨 to understand end-to-end video transmission, such as an end user using his multimedia personal computer. (PC or any other appropriate terminal device) to send a video playback request, first to the regional proxy video server. If the proxy server has the entire content of the requested video, it can send the request to the The request was served immediately under the remote central video server. This is called a cache hit, which means that there is no need to consume backbone resources. In other words, if the regional proxy server has no video copy at all, the request will be reintroduced to the remote video server, and the video data will be sent back from the central server to the area Proxy server, and connect to -16-This paper size is applicable to China National Standard (CNS) A4 (21〇X 297 mm) 479200 A7 B7 V. Description of invention (printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs: This name In order to cache errors, in this example, the video information is operated in the Wang Qian connection domain. The purpose of the cache strategy is to improve the cache hit ratio (which includes some of the available video cache caches in regional proxy feeders).耆 Export cache-to-error ratio and save trunk resources " Huawei, storing the entire video copy on the proxy server requires an area storage area, so if you need to store many video names, you must install a lot of memory and this is expensive, so According to the present invention, if part of the popular video name is stored in the proxy server, then the cache part can be merged with the video server's wrong part of the central server. The merged video names are sent to the requester in a string, so according to the present invention, when the video name is cached, part of the name is distributed in the video name and then cached. Because these parts are distributed in the video name, there are fewer The information material is to be expanded from the central server and merged with the cache part to be transmitted to the requester without interruption. Preferably, part of the stored video material will be evenly distributed on the name, so that it can be reduced. The required peak bit rate used to transmit the erroneous part from the central server is as follows. The cache efficiency is determined by the following two factors, namely the extremely small area storage in the minimal proxy server used by the trunk, so if The video name retrieved by the proxy server is not often requested by the user, which means that the less commonly used scales occupy the storage area of the area, and the hot ~ scales need to be retrieved from the central server, thereby reducing the cache efficiency and therefore affecting the cache efficiency The important factor is the overview of user acquisition. Knowing the user acquisition style can design a better acquisition strategy, so how to correctly predict the user acquisition row It becomes an important matter of improving the hit ratio of the area. Previous research shows that the user access overview is behind the distribution of Zipf-like, but this distribution can only be used as a general guide, and it may not be as fast as the asset and fast name (please read first Note on the back # ^ Fill this page again) Order · ¾ 17- This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) A7 V. Description of the invention (15) Applicable to other situations. For example, in the same case Different audience groups watch different radio programs at different times of the day, and there are several different access overviews on a single day. In addition, the access overview generally represents past behavior and assumes that the trend will be followed in the near future. The less accurate the overview is, the more difficult it is to predict and prepare the cached video name on the regional server beforehand. Therefore, the special embodiment of the present invention does not want to predict the access overview, but instead adjusts the cache strategy whenever the access behavior changes. Therefore, the present invention determines that the 'cache strategy will be synchronized with the current access pattern at each moment, and the adjustment is fast. Take parameters to improve the cache hit ratio, because the access behavior will change and change the size at any time. The cache strategy according to the specific embodiment of the present invention is completely automatic under any kind of intervention. This is for system administrators. Speech is a great advantage. 0 See the server's fast enemy. ^ Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. In the Continuation Video Cache (PVC) strategy according to a specific embodiment of the present invention, a video name is treated as a set of sizes. Equal (or variable large as described below) f fetch units to handle. When the frequency of fetching a particular video name increases, the proxy server then continues to store more cache units because of its fetch action.爻 Welcome, similarly, when the frequency of video name retrieval is reduced, the proxy server can remove its cache unit for caching the most recently retrieved video name. The close relationship between the name's fetching frequency and its cache size can increase the cache performance of the proxy server, thus reducing the backbone bandwidth requirements and the total operating cost. Another advantage is that once you set the cache strategy, you can automatically increase Reduction decisions, as described below. -18- 479200 V. Description of the Invention (16) Although the PVC design is divided into a group of video names with the same cache unit size for easy explanation, this restriction is flexible. The PvC design also divides the video name into a set of cache units with different sizes, such as the I, B, and p boxes in the MPEG format, or in other embodiments described below. Among these other designs, the PVC design is well-suited for encoding into a multi-layer format and / or a variable bit rate format as described below. In the video transmission model according to another feature of the present invention, the regional proxy server refers to sending video data to the user station. The data may come from the area only, the proxy (as a cache hit for the entire available video), and only from the remote server. Monthly I (cache errors as a regional proxy), or from regional and remote sites (cache hits as part of the available video as a regional proxy server). Figure ^ illustrates the video transmission designed for the cache hit of some available video content. In the figure, the regional proxy sends the video content to the client's multimedia]? (:, The regional proxy server captures the data segment of the area 0, 8, 丨 6 And other parts of the data segment from the remote central server (such as data segment i, 2, 3, 4, $, 6 '7' 9 '1 0, etc.), the purpose of this example is only for illustration, and the basis for the following description is provided The additional embodiments and improvements of the present invention are further described. The example in FIG. 2 also reveals the resource requirements of the proxy video server. It has a large amount of memory), so the speed of the hard disk and its storage capacity affect the cache performance. The proxy server also needs the main memory space to buffer the video data. Similarly, the network bandwidth in the backbone connection Connect with the user to determine how much data can flow through it. The following will explain the impact of the above resources on the cache performance and other features according to other embodiments of the present invention. Jiaxian-19- 479200 A7 B7 V. Description of the invention (17 Video capture mode. Just before Ashikaga continues video cache (PVC), consider the video capture model. Video expansion is considered as a periodic capture process. In each period of length P, 'the system class takes the customer's data amount pxri, so that the video name is requested at the data rate ri, which can ensure that the material is prepared before the player needs it' According to the invention, the video is The name Vi is regarded as a sequence of video blocks: 〇1 V: · 'K' ··· V, ·, where niXp represents the video length, and each data block has the size of pXq. In order to facilitate understanding of the present invention, a dual video block is considered. Buffer design 'One of the buffers stores the data to be transmitted to the user through the network module'. At the same time, other buffers are used to store the secondary data block of the regional hard disk or the remote central servo backup. According to the p V c of the present invention The definition example is based on the above assumptions. General Description of Video Cache PVC The Continuation Video Cache (PVC) according to the present invention is a cache design to determine the amount of video content and video data to be cached (or replaced). In part, whenever a cache is required or replaced, the proxy server according to the present invention adheres to the PVC design to identify the video data part to achieve the cache action or operation. The PVC defines a parameter called N to specify the cache Or the replacement unit. Whenever the proxy server requests to cache the video name (or the replacement of the cache resource), the 1 / N (N 2 1) part of the video content is the storage (or replacement) of the area. In other words, I ' When N is equal to 2, the proxy server stores or replaces half of the video name in each cache operation. In the special case of N equal to 1, the cache unit is the entire video name, except for determining the cache size. N 値 also defines 20- This paper size applies the Zhongguanjia Standard (CNS) A4 specification (210 X 2Θ7 public love) Please read the note on the back first and then fill in the I page to customize the consumption of the Intellectual Property Bureau of the Ministry of Economic Affairs Cooperative printed 479200 A7 V. Description of the invention (The correct part of the cached (replaced) video information printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, the PVC design according to the specific embodiment of the present invention is best The 1 / N part of the material is evenly distributed in the video name. Although it is best to arrange the 1 / N part of the data evenly in the video name for practical reasons, this is not necessary. Therefore, the cached sub-blocks can be freely distributed in names, such as using random number generators to select sub-blocks from the block sequence. In another embodiment below, a proxy server is used to cache the same number of sub-blocks from each block of the name, or to store one sub-block at approximately every n blocks in the time series of the blocks, In all of these embodiments, a reduction in the peak transmission bit rate of the central server can be determined. According to another embodiment of the present invention, consider FIG. 3A (the explanation of which can be interpreted in 3 ways) to understand the uniform configuration more. In the upper part of FIG. 3A, the original transmission time of a message name is displayed. Send. In every 1 segment (ie P), the proxy server must send the amount of data of rXp to maintain the quality of service. For a proxy server operating under PVC with N = 4, the cache operation of the message will be 1 / The data part of 4 is placed in the regional agent. Method A in Figure jA imposes a uniform request on the data by writing down the data blocks in the vertical direction. The proxy caches the blocks 0, 4, and 8, which is one of every four data blocks. In the next cache hit, the central servo follows the original time and transmits the remaining blocks on the backbone network. The central server is inactive while the proxy server has stored regional data, so the peak bandwidth of the backbone connection is the same as the cache error. In other words, Method A will not reduce the peak bandwidth requirement of the central server, although It reduces average bandwidth requirements. Fang Ping (see note on the back? Please fill out this page first) Packing 4 21-479200 V. Invention Description (19 Printed in Figure 3 A by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Reduce the peak data rate between methods, and change the transmission time to 3 display-time periods in the extension method 8 ... The central feeder will originally estimate that the technician can understand the data in 7? The data is divided into 4 smaller blocks. Those who are familiar with this will understand how to implement it, so you can change it here. Hi _X ^ Service state pass I can then place 4 small blocks at 4 corresponding times. 2: Less, data rate is 25% Along to. · 75Γβ only because of the transmission of data. ^ Early in advance, the proxy server needs extra buffer space ^ / 2) χΓχΡ instead of 2xrxp in method A to reach the cache first, so increase in The buffer space required by the agent. The method of 7F in Figure OA is to record the data blocks in the horizontal direction. In the cache cache, the proxy server maintains 1/4 P of each data block in the video name of the area. This method can reduce the backbone connection. The peak data rate changes from r to 〇75i < and the k-pulse is required to be maintained at (2xrxp) 値, which is required for the original (uncached) time. The PVC design according to a specific embodiment of the present invention uses two JUi video server As described in the specific embodiment of the present invention, assuming that the total number of message blocks η ^ N in the video name% is an integer multiple (ie, ni > Niniin〇dN = 0), each j 2 blocks to A , Which can be regarded as a set of N equal-sized subblocks represented by 4 'vf7' · · · vfy. As for each video cache hit or replace, lock the next available subblock of the data block of the fetched video segment Blocks, such as a silly PVC design with N = 10, contain the video name Vi in the sequence of 1000 blocks, where it is marked with v, ..., v). . Send. These 100 blocks are -22- Please read back the note of Sj 1 ·! F * The binding method is based on the paper size when the Chinese national standard (CNS) A4 specification (210 X 297) is applied. %) '------ Ao 479200 V. Description of Invention (20 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs and sent to customers or users in a sequential manner. The regional agent servo cries for production and can be arranged accordingly. The MU mouse set up 5 1 memory has the video name, and the more stored # data (where u ^ 100) is triggered in the E domain proxy server by mistake) and is in the E domain hard disk. As shown in Figure 3b *: Figure 2B is a schematic diagram of the video name v with 100 blocks, where the blocks are divided into 0 sub-blocks to illustrate the preferred implementation of the present invention. In the figure, when the video name 'is requested, the area The proxy server does not have this name in the memory, so it starts to store some names in the memory. According to the above method C, it is best to store 1 ° punching knife, refer to FIG. 3B, and then store 100 sub-blocks, J. from 1 to 100, as shown in FIG. 3B, with the number T on the dashed line 132, 112, llj, ..., iU0 are used to indicate that these sub-blocks are collectively referred to as unit 132a. ^ Please again, when the same name Vi, the regional proxy server decides to increase the amount of the name% Vi, which is cached in the cache memory by storing another unit 134a in 100 sub-blocks j From 1 to 100, as shown in the sub-blocks on the dotted line 1 3 4 in FIG. 3B, another cache hit of the same video segment will add the region hard disk, assuming that the data g, # is still in the region server. Then E until all names v i are cached. Therefore, if the regional proxy server has sufficient resources and no rapid replacement occurs, the regional proxy server decides to add a sub-block of the video name Vi, which is stored after each additional request for the name, and then searches for the video name 10 times. After Vi, the area server will have the entire contents of Vi. The formula is used to define the PVC cache effect shown in Figure 3 above. C (vi) indicates that the surface is to be picked up. Please look down (please read the note on the back before filling this page) 丨 Loading. 4 -23 This paper is applicable _ 家 标准 (cns) A4 specifications ⑵◦ χ depending on the public. V. Description of the invention (A7 B7 under the current name of the cached content Vi) C (v,)

C^{^--U2,^} Vf,CMc〇rUm<k jv, |y - 1,2,... I C(vt) ^ 0 C(vJ 其中k是目前或最近的命中計數。 、第-子公式説明快取資料入區域代理,只要視訊名稱未 完全快取。第二子公式説明在取用名稱中無任何快取内 4。第二子公式説明已在區域代理中快取整個視訊名稱。 類似的根據本發明的其他實施例,每當區域代理伺服器 決定在視訊名稱作P V C取代時,它即去除相同量的視訊 資料如P V C快取所爲。選擇去除的部分或單元可以是圖 3 B中的任何組如132.a,134a等,各含1 〇 0個子區塊其已 如上所述的快取。當根據以下的多層擴充性方法而將區塊 分成數個子區塊時,即可去除最近快取的一組或單元,以 下公式説明此策略: (請先閱讀背面之注意?^再填寫本頁) i裝 訂· %· 經濟部智慧財產局員工消費合作社印製 C(v,) =C ^ {^-U2, ^} Vf, CMc〇rUm < k jv, | y-1,2, ... IC (vt) ^ 0 C (vJ where k is the current or recent hit count. -The subformula indicates that the cache data is entered into the regional agent, as long as the video name is not fully cached. The second subformula indicates that there is no cache in the access name4. The second subformula indicates that the entire video has been cached in the regional agent Name. Similar to other embodiments of the present invention, whenever the regional proxy server decides to replace the video name with PVC, it removes the same amount of video data such as PVC cache. The part or unit that is selected for removal can be Any group in Figure 3B, such as 132.a, 134a, etc., each contains 100 sub-blocks which have been cached as described above. When the block is divided into several sub-blocks according to the following multi-layer scalability method , You can remove the recently cached group or unit, and the following formula explains this strategy: (Please read the note on the back first? ^ Then fill out this page) iBinding ·% · Printed by the Consumers ’Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs v,) =

y/1 ^ C〇;,) ΐοτίύ k < N \ fi e C(v,) for U 其中k是目前或最近的命中計數。 上述2個公式説明新快取的内容C ( Vi) ’供視訊名稱、成 -24- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 479200 玉、發明說明( 爲目前V i減去最近快取的部分。 • gj史快取内容 根據本發明的其他實施例,安裝區域代理词服器 -問題是在成串服務開始前如何建立區域快取,根據 明可使用數種方法,如不快取,平均快取,或是 ; 概述的比例快取。 j 不快取是最簡單的方法,即開始時將區域快取内 白,然後於成串服務時建立伺服器内容。這會在開始時: 費整快取空間,這是主要缺點,使得開始服務時需要 的主幹頻寬。 ° 平均取用(或是平均頻寬)方法藉由储存各視訊名稱的相 等視訊資料邵分而利用率的區域代理空間,這是好方法假 設過去沒有取用或是目標應用未在視訊名稱中顯示偏好。又 根據取用概述的比例快取是根據視訊取用記錄而快取資 料,其假設過去趨勢會繼續。 / 經濟部智慧財產局員工消費合作社印製 快取系統開始操件後,如熟於此技術者可了解的,區域 快取内容會隨時改變以反映實際視訊取用動作,使用任何 演譯法(如增加快取子區塊的組數,其與時間周期中的取 用次數成正比,以及當時間周期中的取用次數減少時與取 用次數成比例減少)即可達成,除非一定需要更新快取内 容(如人工快取新發行的視訊名稱),使用自動更新的快取 内容的正確性是足夠的。 區域代理的資源限制 如上所述,根據本發明其他實施例的區域代理伺服器使 -25- 本紙張尺度適用中國國家標準(CNS)A4規格(210 x 297公爱) 479200y / 1 ^ C〇 ;,) ΐοτίύ k < N \ fi e C (v,) for U where k is the current or recent hit count. The above two formulas explain the contents of the new cache C (Vi) 'For video name, Cheng-24- This paper size applies Chinese National Standard (CNS) A4 specification (210 X 297 mm) 479200 Jade, invention description (for the present Vi minus the most recent cached part. • Gj history cache content. According to other embodiments of the present invention, install a regional proxy server-the problem is how to set up a regional cache before the string service starts. Some methods, such as not caching, average caching, or; Generalized proportional caching. J Uncaching is the easiest method, which is to cache the area at the beginning and then create server content when the service is in a string. This starts at the beginning: It costs the entire cache space, which is the main disadvantage, which makes the trunk bandwidth required when starting the service. ° The average access (or average bandwidth) method saves equal video data for each video name. And the utilization of the regional agent space is a good method. It is assumed that the target application has not been used in the past or the target application has not displayed the preference in the video name. The cache based on the proportion of the access overview is faster based on the video access record. Data, it is assumed that the past trend will continue. / After the consumer cache of the Intellectual Property Bureau of the Ministry of Economic Affairs prints the cache system and starts to operate, as those skilled in this technology can understand, the content of the regional cache will change at any time to reflect the actual video Fetch action, using any deduction method (such as increasing the number of cached subblocks, which is proportional to the number of fetches in the time period, and proportional to the number of fetches when the number of fetches in the time period decreases (Reduction), unless the cached content must be updated (such as manually caching newly released video names), the correctness of using the automatically updated cached content is sufficient. The resource limits of the regional agents are as described above. The regional proxy server of the other embodiment of the invention makes -25- This paper size applies to the Chinese National Standard (CNS) A4 specification (210 x 297 public love) 479200

經濟部智慧財產局員工消費合作社印制衣 用P V C以選擇性快取區域資料。爲了達成快取操件,區 域代理伺服器需要供給足夠的資源如硬碟(或是其他大^ 憶體)取用頻寬以便將選擇的資料寫入區域硬碟,及從= 域硬碟擷取快取資料,及硬碟(或是其他大記憶體)儲:: 間以儲存選擇的資料。 1 其他區域資源是網路頻寬(如遠地中央伺服器與區域代 理伺服器之間主幹連線頻寬的區域介面),及區域代理词 服器的前端網路連線及終端用户站台。 實際上,區域代理伺服器相對於記憶體緩衝容量,硬碟 儲存容量及取用頻寬,及網路連線頻寬具有的資源是有限 的’這些資源限制對於快取性能具有重要的影響,根據本 發明其他實施例的快取過程及其資源要求如以下所述。 根據本發明的其他實施例,區域代理伺服器於收到終端 用户的視訊播放請求後即開始快取,如以下所述·· 步驟1 :確定有足夠的前端網路頻寬以傳送新的請求,不 論視訊資料是否最後一定從遠地中央伺服器或區 域代理伺服器(或是二者)送出,區域代理都需要 將它傳送到用户站台,這需要足夠的前端網路頻 寬,否則區域代理會拒絕該請求。 步驟2 :傳送視訊資料到用户站台,依新的請求及目前快 取内容而定,有3種可能情況:快取命中,快取 失誤及部分快取命中。 情況A :快取命中,代理伺服器具有新請求的視訊名稱的 全部内容,區域代理伺服器本身能傳送視訊資 -26- 本纸張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) f請先閱讀背面之注意再填寫本頁} ,-裝 4^1_ 4 479200 A7 B7 五、 發明說明( 24 經濟部智慧財產局員工消費合作社印製 料,假設有足夠的區域硬碟頻寬供擷取區域硬碟 的資料。 快取失誤,代理伺服器不具有新請求的視訊名 稱,視訊資料只能來自遠地中央伺服器,因此必 須有足夠主幹以執行從中央伺服器至區域代理的 貝料傳送,否則,代理伺服器應該拒絕該請求。 若有足夠的主幹頻寬,則代理伺服器能滿足新的播放請 求,此時是由代理伺服器決定調整其快取内容以反映新的 請求,使用PVC的代理伺服器又考慮目前硬碟資源,包 括硬碟取用頻寬(以具有取用頻寬以便將選擇的視訊資料 寫入硬碟)及硬碟儲存容量(以具有空間以儲存選擇的視訊 資料),若2種資源都可用,則代理伺服器儲存部分新請 求的視訊内容。若區域代理具有足夠的硬碟頻寬但是該選 擇資料部分的區域硬碟空間不足,則會要求快取取代=件 (如以下所述)以提供空間給它。 對於不夠區域硬碟頻寬的情況而言(不論是否硬碟空間 夠),都會取消快取,這表示代理伺服器會將進入資料交 通(用以中央伺服器)再導入終端用户,且不快取任何區 資料。 一 部分快取命中,代理伺服器具有新請求視訊名稱 (儲存在其區域硬碟)的部分拷貝,這情況與快取 失誤設計極類似,除了部分請求資料來自代理 服器以外,這表示請求會消耗較少的主幹頻寬 源(以傳送區域代理以外的資料),區域硬碟的Printing of clothing by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Use P V C to selectively cache regional data. In order to achieve the cache operation, the regional proxy server needs to provide sufficient resources such as hard disk (or other large memory) to access the bandwidth to write the selected data to the regional hard disk, and to retrieve the data from the local hard disk. Cache data, and hard disk (or other large memory) :: to store selected data. 1 Other regional resources are network bandwidth (such as the regional interface of the backbone connection bandwidth between the remote central server and the regional proxy server), and the front-end network connection of the regional proxy server and the end-user station. In fact, the regional proxy server has limited resources with respect to memory buffer capacity, hard disk storage capacity and access bandwidth, and network connection bandwidth. These resource limitations have an important impact on cache performance. The caching process and its resource requirements according to other embodiments of the present invention are as follows. According to other embodiments of the present invention, the regional proxy server starts caching after receiving a video playback request from an end user, as described below. Step 1: Make sure there is sufficient front-end network bandwidth to transmit a new request Regardless of whether the video data must be sent from the remote central server or the regional proxy server (or both), the regional proxy needs to send it to the user site. This requires sufficient front-end network bandwidth, otherwise the regional proxy will Reject the request. Step 2: Send video data to the user's site. Depending on the new request and the current cache content, there are 3 possible situations: cache hit, cache error, and partial cache hit. Case A: The cache hits, the proxy server has all the contents of the newly requested video name, and the regional proxy server itself can transmit the video information. -26- This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 public) F) Please read the note on the back before filling in this page}, -install 4 ^ 1_ 4 479200 A7 B7 V. Description of the invention (24 The printed materials of the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, assuming there is sufficient regional hard disk frequency Wide for retrieving data from regional hard disks. Cache error, proxy server does not have a new requested video name, video data can only come from a remote central server, so there must be enough backbone to run the central server to the regional agent If the data is transmitted, the proxy server should reject the request. If there is sufficient backbone bandwidth, the proxy server can satisfy the new playback request. At this time, the proxy server decides to adjust its cache content to reflect the new content. Request, the proxy server using PVC also considers the current hard disk resources, including the hard disk access bandwidth (to have access bandwidth to write the selected video data to the hard disk Disk) and hard disk storage capacity (with space to store the selected video data), if both resources are available, the proxy server stores some of the newly requested video content. If the regional proxy has sufficient hard disk bandwidth but the Selecting a data area where there is insufficient hard disk space will require a cache replacement (as described below) to provide space for it. For situations where there is insufficient regional hard disk bandwidth (whether or not hard disk space is sufficient), Cache will be canceled, which means that the proxy server will enter the data traffic (for the central server) and then import it to the end user without caching any zone data. Part of the cache hit, the proxy server has a new request video name (stored in Partial copy of its regional hard disk), this situation is very similar to the cache miss design, except that some of the requested data comes from the proxy server, which means that the request will consume less backbone bandwidth sources (to transmit data outside the regional proxy) , Regional hard drive

情況B 情況c 伺 資 取 (請先閱讀背面之注意再填寫本頁)Case B Case c (Please read the note on the back before filling this page)

-1 n I 裝 ·. 27 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公爱 479200-1 n I installed ·. 27 This paper size applies to China National Standard (CNS) A4 (210 X 297 Public Love 479200)

五、發明說明(25 ) 經濟部智慧財產局員工消費合作社印製 用頻寬(以擷取區域快取資料及寫入新的快取資 料,其尚未從中央伺服器快取入區域硬碟),及 硬碟空間以儲存新快取的資料。在該步驟中,代 理伺服器中的請求管理員(參考圖6)追縱取用I 也恢復取用概述供各請求視訊名稱用,快取取代 演譯法使用更新的取用概述作爲重要的指南以便 用新内客取代舊的快取資料。以下説明快取取代 演譯法的詳細内容。 快取取代 快取取代用新内容取代邵分現存的視訊快取資料,爲了 快取取代,必須決定要取代那一視訊名稱及那些資料。如 上所述,可以僅取代一特別視訊的最近快取部分。 爲了解決前一問題(決定那一視訊名稱要快取),這並不 是直接的,最好的情況是選擇視訊名稱其具有最佳機會在 未來取用,因爲預測未來取用樣式是困難的,所以本發明 的特定實施例使用一種啓發式方法以預測未來取用樣式, 一種方法是使用最近取用的概述作爲預測未來取用概述的 指示,因此根據本發明的代理伺服器必須追縱各取用請 求’圖中令A i ( t)表示取用視訊v i從時間〇到時間t的序 列,且等於{义,αί2,…,^},其中义表示在時間tj取 用視訊Vi,因此tj S tj +丨(1 〈瓜),在tj = tj+ι的情況下, 它表示2個請求同時發生。 一種測量取用概述的方法稱爲相等加權取用頻率,其定 義如下: -28- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公爱) (請先閱讀背面之注意再填寫本頁) -1裝 4 479200 A7 B7 五、發明說明(26 tkii and &lt;·上 (gjjOj tk&lt; t and a!k g A.(/)jV. Description of the invention (25) Bandwidth for printing by employees' cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs (to capture regional cache data and write new cache data, which has not yet been cached from the central server into the regional hard disk) , And hard drive space to store newly cached data. In this step, the request manager (refer to FIG. 6) in the proxy server traces the retrieval request and also resumes the retrieval overview for each requesting video name. The cache replaces the rendering method and uses the updated retrieval overview as important. Guide to replace old cached data with new guests. The following explains the details of the cache replacement algorithm. Cache Replacement Cache replacement replaces Shao Fen's existing video cache data with new content. In order to replace the cache, you must decide which video name and what data to replace. As mentioned above, only the most recent cached portion of a particular video can be replaced. In order to solve the previous problem (deciding which video name to cache), this is not straightforward. The best case is to choose the video name that has the best chance to be accessed in the future, because it is difficult to predict the future access style. Therefore, the specific embodiment of the present invention uses a heuristic method to predict the future fetch pattern. One method is to use the recently fetched profile as an indication of predicting the future fetch profile. Therefore, the proxy server according to the present invention must follow each fetch. Let A i (t) in the request 'graph denote the sequence of taking video vi from time 0 to time t, and it is equal to {meaning, αί2, ..., ^}, where meaning means to take video Vi at time tj, so tj S tj + 丨 (1 <melon), in the case of tj = tj + ι, it means that 2 requests occur at the same time. A method of measuring access is called equal weighted access frequency, which is defined as follows: -28- This paper size applies to China National Standard (CNS) A4 (210 X 297 public love) (Please read the note on the back before filling This page) -1 Pack 4 479200 A7 B7 V. Description of the invention (26 tkii and &lt; · 上 (gjjOj tk &lt; t and a! Kg A. (/) J

where t s Wwhere t s W

WW

where t &lt; W 經 濟 部 智 慧 財 產 局 員 工 消 費 合 作 社 印 製 /、中Fg(vi’t)疋時間t視訊名稱V i的相等加權取用頻率,取 用頻率是時間周期間視訊名稱Vi的請求次數, 、t W爲例其測量時間周期〇及T c中的取用頻率,w表 示時間窗的値’這可參考圖3C所示,其係取用概述的示 意圖,在圖3C中,時間t可決定一特別視訊名稱的取用概 述j以決足此一名稱是否應該快取或取代,而時間窗w是 設定測量回到時間t,力冃1 ρ^ ^ 吁間t,在圖3C中,廷是以t_w (窗的開始時 間)與t (冒的結束,時間)之間的時間窗來表示&quot;匕一視訊名 %的取用次數接著相加,其中時間窗w中的取用具有相 加權取用/人數的總和接著用肖間窗w來除以得到取用 率,圖3(:説明_種情況即窗的開始時間卜w是在時 間之後(或時間=G),且對應上述取用頻率的上方部= 公式’圖3D説明的情況是時間窗t_w是在開始時間 (或時間=0),且對應上述取用頻率的第二公式。 目 另一種万法是將不同權重加在取用上,其發生在 間’在此,法中,最近發生的取用比較早發生的取用具右 更大的權重,根據本發明其他實施例的p v c設計是 此万法,而取用頻率稱爲時間加權取用頻率且定義如下:豕 g X (/〜/,))for,她恍^价 [λ x(r^j) f〇r where r &lt; ψ -29- L本紙張尺度適用中關家標準(e:NS)A4規格(21Q x视公爱 閱 讀 背 之 注 I裝 頁· 訂 囂 479200 五、發明說明(27 經濟部智慧財產局員工消費合作社印剩农 如何決定時間窗w的長度是重要的事,大的冒値表示在 較長時間收集到的取用樣式,因此較早發生的取用仍在選 擇取代視訊名稱時具有一些權重,換言之,較小w値,即 極早期發生的取用現在於取代決策中也佔有一定角色,較 大時間窗比較少取代的優點多,惟若取用樣式在短時間^ 改變,則不能快速反映該趨勢及適應它,換言之,小時間 窗値的取代策略能將快取内容儘量調整成接近最近取用^ 式,惟它會在短時間觸發過多的資料取代。 7 p v C 強化:C-PVC 先前説明的本發明實施例是指視訊名稱的單次取用,根 據本發明的其他實施例,本發明是指相同視訊名稱的不同 點的多重繼續請求取用,這對於長視訊名稱尤其爲眞,如 電影或熱門視訊名稱,在此情況下,根據本發明的其他實 施例,可以將早期的請求視爲稍後進入請求的預擷取,2 此實施例中,本發明因此能進一步減少主幹頻寬要求如以 下所述,本發明的這些其他實施例在此稱爲複合繼續視訊 快取(C-PVC)。 σ 圖4説明Ν (如上所述)設定爲4時的回應快取效應,假設 各請求會觸發快取操件,圖4(A)是PVC而圖4(Β)是c_pvc 設計,圖中有3個請求以取用相同的視訊名稱,當請求3 進入時,請求2,ί即分別快取時間卜,ti〇的^料^在 PVC設計中,請求1消耗r的主幹頻寬,請求2需要〇 的主幹頻寬,因爲代理伺服器已快取1/4的資料同時服務 請求1,類似的請求3需要0.5r的主幹頻寬,因爲代理伺^ 30- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 請 先 閱 讀 背 面 之 注 意I 事癱I 項零I 再 I 填 I裂 頁 叮where t &lt; W Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs /, in the middle of Fg (vi't) 疋 time t equal weighted access frequency of the video name Vi, the access frequency is the request of the video name Vi during the time period Times, and t W are taken as examples. The measurement frequency is taken from the time period 0 and T c, and w represents the time window 値 ′. This can be referred to FIG. 3C, which is a schematic diagram of an overview. In FIG. 3C, the time t may determine an overview of the use of a particular video name j to determine whether this name should be cached or replaced, and the time window w is to set the measurement back to time t, force 冃 1 ρ ^ ^ 间 t, in Figure 3C In the court, the time window between t_w (the start time of the window) and t (the end of the risk, time) is used to indicate the number of times the video name% is fetched, and the time window w is fetched. Take the sum of the weighted fetches / number of people and then divide the shaw window w to get the fetch rate. Figure 3 (: Explanation_cases, ie the window's start time, w is after time (or time = G), And the upper part corresponding to the above-mentioned fetched frequency = formula 'Fig. 3D illustrates the case where the time window t_w is in Start time (or time = 0), and correspond to the second formula of the above access frequency. Objective Another method is to add different weights to the access, which occurs in between 'here, in this method, the latest access With the earlier weighting, the weight is larger. The PVC design according to other embodiments of the present invention is this method. The fetching frequency is called the time-weighted fetching frequency and is defined as follows: 豕 g X (/ ~ /, )) For, she 恍 valence [λ x (r ^ j) f〇r where r &lt; ψ -29- L This paper size applies the Zhongguanjia standard (e: NS) A4 specification (21Q x as public reading Behind the Note I Binding Page · Ordering 479200 V. Invention Description (27 How to determine the length of the time window w is important for India ’s leftover farmers, the Intellectual Property Bureau employee consumer cooperative of the Ministry of Economic Affairs. Access style, so earlier access still has some weight when choosing to replace the video name, in other words, smaller w 値, that is, very early access now also occupies a certain role in the replacement decision, a larger time The advantages of less windows replacement are many, but if the style is used in a short time If the time changes, it cannot quickly reflect the trend and adapt to it. In other words, the replacement strategy of the small time window can adjust the cached content as close as possible to the most recently accessed mode, but it will trigger too much data replacement in a short time. 7 pv C enhancement: The embodiment of the present invention previously described by C-PVC refers to a single retrieval of the video name. According to other embodiments of the present invention, the present invention refers to multiple continued request retrievals at different points of the same video name. This is particularly embarrassing for long video names, such as movie or popular video names. In this case, according to other embodiments of the present invention, the earlier request can be considered as a pre-fetch of a request that enters later. 2 In this embodiment, Therefore, the present invention can further reduce the backbone bandwidth requirements as described below. These other embodiments of the present invention are referred to herein as composite continuous video cache (C-PVC). σ Figure 4 illustrates the response cache effect when N (as described above) is set to 4. Assume that each request will trigger the cache operation. Figure 4 (A) is PVC and Figure 4 (B) is c_pvc design. 3 requests to use the same video name, when request 3 enters, request 2, which means to cache the time respectively, ti〇 ^^ In the PVC design, request 1 consumes r's backbone bandwidth, request 2 A trunk bandwidth of 0 is required, because the proxy server has cached 1/4 of the data while serving request 1, a similar request 3 requires a trunk bandwidth of 0.5r, because the proxy server ^ 30- This paper standard applies Chinese national standards ( CNS) A4 size (210 X 297 mm) Please read the note on the back I I will be paralyzed I item zero I then I fill in a cracked page

I 479200 A7I 479200 A7

五、發明說明(28 ) 器已快取2/4的資料同時服務請求2,3。 圖4B顯示C-PVC設計,代理伺服器接收中央伺服器的 整個資料及快取它的1/4資料以服務請求i,當請求2進入 時’請求1在時間t3,代理伺服器仍從中央伺服器接收供 請求1用’藉由了解進入的請求要求相同的視訊名稱,代 理伺服器也能快取額外資料,除了開始時選擇的資料以 外,在此例中,請求2僅於時間η之後消耗〇·5γ的頻寬, 與PVC設計中的〇.75r相比。當請求3於時間tl〇進入時, 請求2僅於時間t1()之後需要0.25r的頻寬,新進入的請求3 於時間t7之後僅消耗〇.25r的頻寬。 在此例中,若整視訊長度是1100,則P V C設計中的總主 幹頻寬交通等於225r ( = 10〇r + 0.75r + 〇.5r分別對應於請 求1,2,3 ),在C-PVC設計中,總主幹頻寬使用量是 155r 〇 在這些請求中,這是100r (請求1 ),( 3 χ〇·75χγ + 7x0.5xr .+ 90x0.25xr)(請求 2 ),及(7x0.5xr + 93x0.25xr )(請求 3 ), 這是與PVC設計相比,此例中的C-PVC可省下31.11%的 頻寬。 擴充性多層方法 經濟部智慧財產局員工消費合作社印製 根據國際儲存量ISO/IEC 13818-2,第一版1996/05/15 (其 在此供參考)而提議數種擴充性視訊編碼,此儲存量提議 三種擴充性:S N R擴充性,空間擴充性及暫時擴充性, 在各種擴充性中,假設基層會提供基本的視訊資料其足以 複製一視訊影像,因此對於有限頻寬的觀看者而言,僅傳 -31 - 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) &quot; 479200 五、發明說明(29 运到基層,對於高頻寬的觀看者而言,每_種擴充性都具 有至少-個強化層’當與基層資訊合併時即可強化觀看者 看到的視訊品質。θ此在SNR擴充性中,所有的層都且 咖的空間解析度,基層提供基本的視訊品質,‘化層 藉由提供基層的DCT係數的精密資料而增加視訊品質。曰 在空間擴充性中,基層提供基本空間解析度及暫時率,'強 化層使用空間插入基層以增加空間解析度,在暫時擴充性 中,基層提供基本暫時率,強化層使用相對於基層的暫時 預測,基層及強化層可合併以產生全暫時率輸出,MPEG 2支援上述擴充性方法,關於1^1&gt;£(}2方法的擴充性可表 1 1 章,名稱&quot;MPEG 2” Of Video Demystified,A nk for the Digital Engineer, 2nd Edition, HighText Interactive 參V. Description of the invention (28) The server has cached the data of 2/4 while serving requests 2,3. Figure 4B shows the C-PVC design. The proxy server receives the entire data of the central server and caches 1/4 of it to serve request i. When request 2 enters, 'request 1 is at time t3, the proxy server is still from the central The server receives the request for request 1 '. By knowing that the incoming request requests the same video name, the proxy server can also cache additional data. In addition to the data selected at the beginning, in this example, request 2 is only after time η Consuming a bandwidth of 0.5γ, compared with 0.75r in the PVC design. When request 3 enters at time t10, request 2 requires a bandwidth of 0.25r only after time t1 (), and newly entered request 3 consumes only 0.25r of bandwidth after time t7. In this example, if the entire video length is 1100, the total trunk bandwidth traffic in the PVC design is equal to 225r (= 10〇r + 0.75r + 0.5r corresponding to requests 1, 2, and 3, respectively). In the PVC design, the total backbone bandwidth usage is 155r. Among these requests, this is 100r (request 1), (3x0 · 75xγ + 7x0.5xr. + 90x0.25xr) (request 2), and (7x0 .5xr + 93x0.25xr) (request 3). This is compared to the PVC design. The C-PVC in this example can save 31.11% of the bandwidth. Expandable multi-layer approach Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, based on the international storage volume ISO / IEC 13818-2, the first edition 1996/05/15 (which is here for reference) proposes several expandable video coding, this Three expansions of storage capacity are proposed: SNR expansion, space expansion, and temporary expansion. Among the various expansions, it is assumed that the basic layer will provide basic video data that is sufficient to copy a video image, so for viewers with limited bandwidth , Only pass -31-This paper size is applicable to China National Standard (CNS) A4 (210 X 297 mm) &quot; 479200 V. Description of the invention (29 shipped to the grassroots level, for high-frequency viewers, each _ kind of expansion All have at least one enhancement layer. When combined with the information of the base layer, the video quality seen by the viewer can be enhanced. Θ In the SNR scalability, all the layers have spatial resolution, and the base layer provides basic video Quality, 'The layer increases the video quality by providing precise data on the DCT coefficients of the base layer. In space expansion, the base layer provides basic spatial resolution and temporary rate,' Use space to insert the base layer to increase the spatial resolution. In temporary scalability, the base layer provides the basic temporary rate. The enhancement layer uses the temporary prediction relative to the base layer. The base layer and the enhancement layer can be combined to produce a full temporary rate output. MPEG 2 supports the above expansion For the method of 1 ^ 1 &gt; £ (} 2, the extensibility of the method can be shown in Chapter 1 of Chapter 1, the name &quot; MPEG 2 "Of Video Demystified, Ank for the Digital Engineer, 2nd Edition, HighText Interactive

Inc·,San Dieg0, CA 1996, pp 5〇3_512,該章節在此供, 考。 施 除了上述擴充性方法以外,也可以有它種擴充性實 例 數 數 在 的 經濟部智慧財產局員工消費合作社印制衣 2 土於:ϊ:間擴充性編碼,例如在消費者電子產品中,多 的視訊源使用交錯的顯示格式:掃描各框成爲2個(奇 及偶數)場其在垂直方向暫時分開且作空間補償,因此 一種空間擴充性編碼形式中,低層包括交錯擴充性格式 2個場之一而強化層包括另一場,因此從代理伺服器將 個場傳送到客户的頻寬不夠,只有基層包括的場會傳送, 其中唯一傳送的場會提供稍微劣化但是足夠的視訊影像。 惟若代理伺服器至客户的頻寬足夠,則包括另一場的強化 -32- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公爱) 479200 A7Inc., San Dieg0, CA 1996, pp 503_512, this section is provided here for examination. In addition to the above-mentioned expansive methods, there may also be other types of expansive examples. The printing of clothing by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 2 is based on: expansive coding, such as in consumer electronics, Multiple video sources use an interlaced display format: each frame is scanned into 2 (odd and even) fields which are temporarily separated in the vertical direction and are spatially compensated. Therefore, in a form of spatial expansion coding, the lower layer includes 2 interlaced expansion formats The enhancement layer includes another field, so the bandwidth from the proxy server to the individual field to the client is not sufficient. Only the field included in the base layer will transmit. The only field transmitted will provide a slightly degraded but sufficient video image. However, if the bandwidth from the proxy server to the client is sufficient, another session of enhancement is included -32- This paper size applies to China National Standard (CNS) A4 (210 X 297 public love) 479200 A7

五、發明說明(3〇 ) 層會傳送,所以2個場一起提供較佳品質的視訊影像給客 户。換言之,空間擴充性編碼方法中的基層包括沿著一組 掃描線(如奇數掃描線)的視訊資料,這是根據任一種通用 儲存量如NT S C,而強化層沿著另一組掃描線(如偶數掃 描線)的視訊資料以包括根據相同儲存量的另一場。當代 理伺服器與客户之間的連線的頻寬有限因而僅能傳送基層 時’基層中唯一場的資訊可轉成全視訊框,這是藉由一種 習用演譯法如掃描線插入及場合併,例如在9章,名稱 Video Processing&quot; 〇f Video Mystified, A Handbook for theV. Description of the invention The (30) layer will transmit, so the two fields together provide a better quality video image to the customer. In other words, the base layer in the spatially scalable coding method includes video data along one set of scan lines (such as odd scan lines). This is based on any common storage capacity such as NT SC, and the enhancement layer follows another set of scan lines ( (E.g., even scanning lines) to include another field based on the same amount of storage. When the bandwidth of the connection between the proxy server and the client is limited and therefore only the grassroots can be transmitted, the information of the only field in the grassroots can be converted into a full video frame. This is done by a conventional interpretation method such as scan line insertion and field Merged, for example in Chapter 9, the name Video Processing &quot; 〇f Video Mystified, A Handbook for the

Digital Engineer,2nd Edition,HighText Interactive,Inc.,San Diego, CA 1996, pp. 3 86-394,該章節在此供參考。雖然上 述假設將視訊資訊分成2個相等部分以再分成2層,但是 它也能將它分成不相同的部分,如基層佔四分之一而強化 層佔四刀之二’而且也可能將該四分之三再分成一個以上 的強化層’其中以可用頻寬的函數來決定傳送的強化層數 目。 經濟部智慧財產局員工消費合作社印製 暫時擴充性方法的另一範例如以下所述,當視訊資源以 特別取樣率提供視訊資料時,來自該資源的部分視訊資料 4八本即結成一群以包括以一取樣率傳送的基層,該率低於 資源的取樣率,而且剩餘樣本也結成一群且以較低取樣率 傳送以包括強化層。以上述的空間擴充性方法的情況爲 例’當頻寬不准傳送基層及強化層時,僅有基層會傳送且 可使用習用方法以便在低取樣率之下將視訊資料轉成全視 訊框,當頻寬允許傳送基層及強化層時,這二層的樣本會 -33- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 479200 發明說明(31 經 濟 部 智 慧 財 產 局 員 工 消 費 合 作 社 印 製 以習用方式合併以顯示較佳的視訊品質,再度,可執行將 資源的樣本分成二層,以便這二層不具有相同的樣本數, 且更可能將樣本分成一基層及一個以上的強化層。 上述設計可用以實施圖3八中的方法c,因此以上述的空 間擴充性方法爲例,可以快取沿著1/4掃描線的視訊資 料,其中這些掃描線最好均勻地在框上間隔,以便僅需要 從中央伺服器擷取沿著視訊框的剩餘3 /4掃描線的視訊資 料:類似的,在暫時擴充性方法中,也可快取1/4掃描線 ”最好枭時均勻地在視訊框上間隔,同時從中央伺服器擷 取剩餘的樣本,以實施圖3 A的方法c。 ilg伺服器及快取記憶體之結摄_例 局了更了解本發明,圖6顯示代理伺服器的方塊圖,立 對應圖5的代理词服器108或11()以説明一通訊系統的/ 型配置,其中可實施本發明。 根據本發明的實施例,代理伺服器載有本發明實施例… 編譯及連接版纟,當該方法執行它時,冑編譯及連接版本 即執行以下方法: 決定請求一視訊名稱之次數; 儲存在-代理伺服器之快取中以回應該視訊名稱之請求 次數,即對應複數個視訊名稱之快取單元; 每當收到一新請求時即更新快取,·其中更新包括: 若新請求對應視訊名稱則在快取中儲存更多視訊名稱 快取單元;及 若新請求不對應視訊名稱則從快取中刪除視訊名稱之 典 的 之 至 (請先閱讀背面之注意再填寫本頁) ;裝 . · %· 34-Digital Engineer, 2nd Edition, HighText Interactive, Inc., San Diego, CA 1996, pp. 3 86-394, this section is here for reference. Although the above assumption divides the video information into two equal parts and then into two layers, it can also divide it into different parts, such as the base layer accounts for a quarter and the enhancement layer accounts for two of the four '. Three quarters are further divided into more than one enhancement layer, where the number of enhancement layers transmitted is determined as a function of available bandwidth. Another example of a temporary expansion method for printing by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs is, for example, described below. When video resources provide video data at a special sampling rate, some of the video data from that resource are grouped into a group of The base layer is transmitted at a sampling rate that is lower than the resource sampling rate, and the remaining samples are also grouped and transmitted at a lower sampling rate to include the enhancement layer. Take the case of the above-mentioned spatial expansion method as an example. When the bandwidth is not allowed to transmit the base layer and the enhancement layer, only the base layer will transmit and the conventional method can be used to convert the video data into a full video frame at a low sampling rate. When the bandwidth allows transmission of the base layer and the enhancement layer, the samples of these two layers will be -33- This paper size applies to the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 479200 Invention description (31 Employees of the Intellectual Property Bureau, Ministry of Economic Affairs Consumption cooperatives print custom combinations to show better video quality. Once again, you can divide the sample of resources into two layers, so that the two layers do not have the same number of samples, and it is more likely to divide the samples into one base layer and more than one. The above design can be used to implement method c in Figure 3-8, so using the above-mentioned space expansion method as an example, video data along the 1/4 scan line can be cached, and these scan lines are preferably uniform Space the frames so that you only need to retrieve video data from the central server along the remaining 3/4 scan lines of the video frame: similarly, in a temporarily extensible method It is also possible to cache 1/4 scan line "It is best to evenly space on the video frame at the same time, and at the same time retrieve the remaining samples from the central server to implement method c of Figure 3 A. Ilg server and cache memory The conclusion of the body_ example to understand the present invention better, Figure 6 shows a block diagram of the proxy server, corresponding to the proxy server 108 or 11 () of Figure 5 to illustrate the / configuration of a communication system, which can be implemented According to the embodiment of the present invention, the proxy server carries the embodiment of the present invention ... a compiled and linked version. When the method executes it, the compiled and linked version executes the following method: determining the number of times a video name is requested ; Stored in the cache of the proxy server to respond to the number of requests for the video name, that is, the cache unit corresponding to a plurality of video names; The cache is updated whenever a new request is received. The update includes: If If the new request corresponds to the video name, store more video name cache units in the cache; and if the new request does not correspond to the video name, delete the code name from the cache (please read the note on the back before filling This page); · install · 34-%.

47920U Α747920U Α7

少單元’或者若無足夠儲存空間供快取新的快取單 兀或是 Λ 他理由時,則不删除任何快取單元。 夂頭不代理伺服器的快取記憶體的典型佈局,其中假 5又刼各視訊均等分成Ν個快 如上所述,各快取單 兀,括的子區塊於不同區塊中都不相同,如單元132a, 、依取用頻率而足,各視訊的快取單元數目會改變, ,K,對於—視訊名稱的請求次數越多,則在代理伺服 -的丨夬取e己憶體中快取的視訊的快取單元數目越多。當再 有視訊名稱的請求有時,即更新快取記憶體,如減少:是 完全刪除視訊的快取單元的數目,以留下空間供視訊名稱 用。繼續視訊快取方法會導致擴充性快取機構且大幅改良 媒體傳送系統的總體性能。 圖8頑7F對應圖1伺服器i 0 2的中央視訊伺服器的方塊 圖,載入官理員接收代理伺服器的請求且確定以最迅速方 式傳送請求的視訊,内容管理員提供一般背景給終端裝置 以頭不琢視訊,背景包括額外資訊,以便服務提供者令終 端裝置用户最好去參加如服務/產品促銷資訊。 圖9顯示根據本發明的中央視訊伺服器的實施例。 經濟部智慧財產局員工消費合作社印製 爲了更了解本發明,圖1 〇,圖1 1分別顯示(且根據本發 明的實施例)代理模組及視訊伺服器模組的流程圖,分別 在代理伺服器及中央視訊伺服器中執行。 根據本發明特徵的媒體傳送系統是強力,可迅速操件且 合乎經濟,繼續快取機構提供代理伺服器中快取記憶體的 最佳使用,且允許成_視訊無缝的傳送,以保證服務品 -35- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 479200 A7Few units' or if there is not enough storage space to cache a new cache unit or other reasons, no cache unit is deleted. The gimmick does not represent the typical layout of the cache memory of the server. Among them, the false 5 is divided into N caches. As described above, each cache is single, and the enclosed sub-blocks are different in different blocks. For example, if the unit 132a, depends on the frequency of use, the number of cache units for each video will change, K, the more requests for the video name, the more it will be in the proxy server The greater the number of cached units of the cached video. When there is another request for the video name, the cache memory is updated. If it is reduced: Yes, the number of cache units of the video is completely deleted to leave space for the video name. Continuing the video caching approach will result in a scalable caching mechanism and significantly improve the overall performance of the media delivery system. FIG. 8 is a block diagram of the central video server corresponding to the server i 0 2 in FIG. 1. It loads the administrator to receive the request from the proxy server and determines to transmit the requested video in the fastest way. The content manager provides a general background to The terminal device does not pay attention to the video, and the background includes additional information, so that the service provider can make the terminal device user better to participate in such service / product promotion information. FIG. 9 shows an embodiment of a central video server according to the present invention. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. To better understand the present invention, FIG. 10 and FIG. 11 respectively show (and according to the embodiment of the present invention) a flowchart of the agent module and the video server module, respectively, in the agent Server and central video server. The media transmission system according to the features of the present invention is powerful, can be operated quickly and is economical, and the cache mechanism continues to provide the best use of cache memory in the proxy server, and allows seamless transmission of video to ensure service品 -35- This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) 479200 A7

質。此外本發明可以與任—種顯示方式配合使用,包括銷 售展示及產品/服務促銷,以提供視訊服務提供者的額外 收入來源。 本發明的方法,序列或步驟及特徵都是互相相關,且都 合乎新穎性’該揭示的方法及序列可單獨執行或合併執行 以提供適用媒體傳送系統的新且非顯著的檔案結構系统, 該了解的是該等方法及序列的合併也可產生相同的新箱 性,即使是以最廣義的方式合併。 其他實施例 已參考特足實施例説明本發明,熟於此技術者可了解其 他實施例’尤其是已將用户數位資訊裝置以個人電腦爲例 作大致說明,惟數位計算裝置是指任何裝置其可以與遠地 資料應用互動,也可包括這些如數位電視,行動^ 人數位助理器等。 請 先 閱 讀 背 S 之 注 意 事編 項1 再 填 寫裝 頁 訂 讎 經 濟 部 智 慧 財 產 局 員 工 消 費 合 作 社 印 製 此外已用主從應用環境在-些範例中説明本發明,^ 本發明並不僅限於這些邏輯環境如主從等,本發明的客疋 ,廣義的包括任何邏輯用以從遠地系統取用資料,而词月尸良 器廣義的包括任何邏輯用以向遠地系統提供資料。 該了解本文的範例及實施例只是用以説明, :::本發明的敎示了解各種改良或變化,這些都包:: 申知專利的精神及範圍之中,以及其同等者。 在 置之實施例 以入,%狀骚术貫施,在本發明 …財,本發明的不同特徵可以實施在客户邏輯 或 -36-quality. In addition, the present invention can be used in conjunction with any display method, including sales display and product / service promotion to provide an additional source of income for video service providers. The methods, sequences or steps and features of the present invention are all interrelated and all of novelty. The disclosed methods and sequences can be performed separately or combined to provide a new and non-significant file structure system suitable for a media delivery system. It is understood that the merging of these methods and sequences can also produce the same new box properties, even in the broadest way. Other embodiments have described the present invention with reference to specific embodiments. Those skilled in the art can understand other embodiments. In particular, the user's digital information device has been roughly described with a personal computer as an example. However, a digital computing device refers to any device Can interact with remote data applications, including these such as digital TV, mobile ^ personal assistant and so on. Please read the Caution Item 1 on the back of S first, then fill in the binding page, printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. In addition, the master-slave application environment has been used to illustrate the invention in some examples. The invention is not limited to these. Logical environments, such as master-slave, etc., the subject matter of the present invention, broadly includes any logic to retrieve data from remote systems, and Ciyue Liangyi broadly includes any logic to provide data to remote systems. The examples and embodiments in this article are just for illustration. ::: The present invention shows various improvements or changes. These include: the spirit and scope of the claimed patent, and their equivalents. In the embodiment of the present invention, the implementation of the tactics, the different features of the present invention can be implemented in the customer logic or -36-

% 479200 A7 經濟部智慧財產局員工消費合作社印製 五、發明說明( 是伺服器邏輯,可了解本發明或其元件可以實施在固定媒 體節目元件,其包括邏輯指令及/或資料,當载入適當的 計算裝置時,可以使該裝置根據本發明而執行。可了解固 定媒體節目可以在固定媒體向用户傳送以載入用户電腦, 或者固定媒體代理可置於遠地伺服器,而用户可經Z通訊 媒體而取用以下載節目。 圖12顯示—資訊裝置(或數位裝置)7〇〇,可了解它是邏 輯裝置能從媒體717及/或網路埠7 :ι 9讀取指令,裝置7〇〇 然後能使用這些指令以導引伺服器或客户邏輯,這是習用 的,以實施本發明的特徵。一種邏輯裝置其可實施本^明 是電腦系統7〇〇,包括CPu 707,額外輸入裝置7〇2, 711,硬碟715及額外的螢幕70 5。固定媒體717可用以 設足此一系統,且表示碟型光學或磁性媒體,磁帶,固態 記憶體等。可以將本發明完全或部分的作爲記錄在此固= 媒體的軟體,通訊埠7 i 9也可用以於開始時接收指令,用 以设足此一系統及表示任一種通訊連線。 也可以將本發明完全或部分的作爲應用特定積體電路 (ASIC)的電路或可程式邏輯裝置(pLD)。在此_情況下 本發明可實施在電腦可了解的説明性語言,用以產生 ASIC或P L D以便如本文所述的操件。 -37 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) (請先閱讀背面之注意再填寫本頁) 裝 訂* - 4% 479200 A7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. 5. Description of the invention (is server logic. It can be understood that the invention or its components can be implemented in fixed media program components, which include logic instructions and / or information. When a suitable computing device is used, the device can be implemented in accordance with the present invention. It can be understood that fixed media programs can be transmitted to the user on the fixed media to load the user's computer, or that the fixed media agent can be placed on a remote server, and the user can pass Z Communication media to download programs. Figure 12 shows-information device (or digital device) 700, it can be understood that it is a logical device that can read instructions from media 717 and / or network port 7: ι 9, device 7 〇〇 These instructions can then be used to guide server or client logic, which is customary to implement the features of the present invention. A logic device that can implement the present invention is a computer system 700, including CPu 707, additional input Devices 702, 711, hard disk 715 and additional screen 70 5. Fixed media 717 can be used to set up such a system, and indicates disk-type optical or magnetic media, magnetic tape Solid-state memory, etc. The present invention can be used in whole or in part as software recorded on this solid = media, and the communication port 7 i 9 can also be used to receive instructions at the beginning to set up this system and indicate any kind of communication connection. The present invention can also be used in whole or in part as a circuit using an application specific integrated circuit (ASIC) or a programmable logic device (pLD). In this case, the present invention can be implemented in a computer-readable descriptive language using To produce ASIC or PLD to operate as described in this article. -37 This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) (Please read the note on the back before filling this page) Binding *- 4

Claims (1)

479200 A8 B8 C8 D8479200 A8 B8 C8 D8 申請專利範圍 1. 一種在系統中快取之方法,用以從_ 丁天服斋及一代 理伺服器傳送複數個媒體資料名稱 ' 王主乂 一客户,其中 各名稱分成諸區塊在一定時序下傳送至至少—客户,、而 各區塊分成諸子區塊,該方法包括·· 識別那些子區塊來自要快取之各名稱之不同區塊,並 中識別出之子區塊包括在至少—名稱之區塊上分佈者/及 在代理词服器控制下快取識別出之子區塊以減 伺服器之位元率用以傳送名稱。 2·如申請專利範園第!項之方法,其中在與時間無請求時 段快取該快取子區塊。 3. 如申請專利範圍第β之方法,其中該快取從該至少— 名稱之各區塊實質快取相同數目之子區塊。 4. 如申請專利範圍第卜頁之方法,其中媒體名稱包括視訊 名稱,而子區塊包括視訊框,而各區塊分成要循序傳送 之視訊框,且更包括從中央伺服器將快取視訊框插入— 視訊框争以形成一合併串及傳送合併争至客户。 經濟部智慧財產局員工消費合作社印製 5·如申請專利範圍第丨項之方法,其中媒體名稱包括視訊 名稱,而子區塊包括視訊框之部分資訊,其中視訊框要 循序傳送,更包括將代理伺服器之視訊框之部分資訊與 中央伺服器之此視訊框之互補部分資訊合併在 框,及傳送完全視訊框至客户。 。 6·如申請專利範圍第5項之方法,其中部分資訊包括沿著 視訊框之一些掃描線之視訊資訊,更包括將沿著此掃描 線之視訊資訊與沿著中央伺服器之此視訊框之其他掃描 -38 ‘纸張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐Patent application scope 1. A method of caching in the system, used to transmit multiple media data names from _ Ding Tianfuzhai and an agent server 'Wang Zhuo a client, where each name is divided into blocks at a certain timing Down to at least-customers, and each block is divided into sub-blocks, the method includes ... identifying those sub-blocks from different blocks of each name to be cached, and the identified sub-blocks are included in at least- The distributor on the block of the name and / or cache the identified sub-blocks under the control of the proxy server to reduce the bit rate of the server to transmit the name. 2 · If you apply for a patent, please! Method, in which the cache sub-block is cached in a time period without a request. 3. If the method of applying for the scope of the patent β, wherein the cache substantially caches the same number of sub-blocks from each block of the at least-name. 4. For the method of applying for the page of the patent scope, where the media name includes the video name, and the sub-block includes the video frame, and each block is divided into video frames to be transmitted in sequence, and it also includes caching the video from the central server. Frame Insertion-Video frame contention to form a merged string and send the merged contention to the customer. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs5. For the method of applying for item No. 丨, where the media name includes the video name, and the sub-block includes part of the information in the video frame. Part of the information of the video frame of the proxy server and the information of the complementary part of this video frame of the central server are combined in the frame, and the full video frame is sent to the client. . 6. The method according to item 5 of the patent application, where some of the information includes video information along some scanning lines of the video frame, and also includes the video information along this scanning line and the video frame along the central server. Other Scan-38 'Paper Size Applicable to Chinese National Standard (CNS) A4 Specification (210 X 297 mm 申請專利範圍 線之互補視訊資訊合併成完全視訊框,及傳送完全視訊 框至客户。 7. 如申請專利範圍第5項之方法,其中部分資訊包括在_ 組取樣時間得到之視訊資訊及一第一取樣率,低於該資 訊起源之視訊源之取樣率,更包括將較低第一取樣率之 代理飼服器之視訊資訊與取樣時間擷取之互補視訊資訊 合併,孩取樣時間與中央伺服器之此視訊框之取樣時間 =不同,其在高於第一取樣率之取樣率之下合併成視訊 貝料,及在較高取樣率傳送視訊資料至客户。 8. 如申請專利範圍第5項之方法,其中部分資訊包括一基 本層中之視訊貪訊,而互補部分資訊包括一強化層中之 視訊資訊,該基本層及強化層係根據空間,訊噪或暫時 擴充性而定義。 9·如申請專利範圍第Μ之方法,其中識別係、代理處名稱 之取用概述函數。 讥如申請專利範圍第!項之方法,其中在客户之任何名稱 取用(前’―平均快取方式藉由儲存各名稱之實質相等 數目之子區塊而利用代理伺服器之儲存區。 經濟部智慧財產局員工消費合作社印製 11·如申請專利範圍第i項之方法, 具中在客尸之任何名稱 取用I前,一比例快取方式利 J W取用歷史資料以決足要 快取多少各名稱。 I2.如申請專利範圍第1項之方法 ^ 々次其中系統開始操件之 後,代理伺服器之快取内容备喃 ^ ^ 奋曰随時改變以反映實際取用 動作。 -39- 479200 A8 B8 C8 D8Complementary video information from patent application lines is combined into a full video frame, and the full video frame is sent to the customer. 7. For the method in the fifth item of the patent application, part of the information includes the video information and a first sampling rate obtained at the sampling time of the _ group, which is lower than the sampling rate of the video source from which the information originated. The video information of an agent feeder with a sampling rate is combined with the complementary video information acquired by the sampling time. The sampling time of this sampling frame is different from the sampling time of this video frame of the central server, which is higher than the sampling rate of the first sampling rate. Combined into video materials, and send video data to customers at a higher sampling rate. 8. For the method in the fifth item of the patent application, part of the information includes video corruption in a basic layer, and the complementary part of information includes video information in an enhanced layer. The basic layer and the enhanced layer are based on space, noise and noise. Or temporarily expandable. 9. The method according to the scope of patent application, wherein the identification system and the name of the agency are summarized.讥 If the scope of patent application is the first! Item method, in which any name of the customer is accessed (pre -'- average caching method uses the storage area of the proxy server by storing a substantially equal number of sub-blocks of each name. Employee Intellectual Property Bureau of the Ministry of Economic Affairs, Consumer Consumption Cooperative Print System 11: If the method of item i in the scope of the patent application is applied, before any name of the corpse is taken by I, a proportional cache method will be used by JW to access historical data to determine how many names are to be cached. I2. The method of applying for the first item of the patent scope ^ 々 times where the system starts to operate, the cache content of the proxy server is prepared ^ ^ Fen Yue changes at any time to reflect the actual access action. -39- 479200 A8 B8 C8 D8 申睛專利範圍 U如申請專利範圍第i項I方法, 理黾朴令、七… ^ 尺匕括猎由確定從該π 理土此一各尸有足夠頻寬而於接收客户之名 在代理伺服器開始一快取過程以傳土 w 頻寬則拒絕請求。 ,及若無足夠 14.如申請專利範圍第13項之方法,更包括當此_内容在 β代理伺服ϋ之區域儲存區之中時,傳送請求名稱之完 全内容。 15·如申請專利範圍第13項之方法,更包括: 口當該代理词服器不具有請求名稱之完全内容時,判定 是否有足夠可用主幹頻寬俾從該中央伺服器攜帶該名稱 至3代理伺服為,及若無足夠可用主幹頻寬則拒絕請 求。 16·如申請專利範圍第15项之方法,更包括啓動繼續快取 過程以調整該代理伺服器之快取内容以反映請求名稱。 17.如申請專利範圍第!項之方法,更包括藉由删除此一名 稱之最近快取邵分而取代一特別名稱之快取部分。 18·如申請專利範圍第1項之方法,更包括使用一最新取用 概述作爲一未來概述指示,以決定要快取取代那些名 稱。 經濟部智慧財產局員工消費合作社印製 19·如申請專利範圍第i項之方法,更包括在代理伺服器追 縱各取用請求以決定要快取取代那些名稱。 20·如申請專利範圍第i項之方法,更包括使用一最新取用 概述作爲一未來概述指示,以決定要快取取代那些名 稱,其中該決定包括: -40- 本紙張尺度適用中國國家標準(CNS)A4規格(2iq x 297八爱) 479200 申請專利範圍 在快取取代時間結束時定義一時間窗; 在一代理伺服器儲存區計算各名稱之取用頻率,該取 用頻率係時間窗期間或其—部分取用此一名稱之函數;及 執行快取取代以回應儲存區中名稱之取用頻率。 2L如申請專利範圍第20項之方法,其中該取用頻率與時 間窗期間或其-部分取用此—名稱之總和成正比。 22. 如申請專利範圍第2〇項之方法,其中該取用頻率 頁 間窗期間或其-部分取用此—名稱之時間加權總和成正 比,而時間加權偏好窗中最近發生之取用。 23. 如申請專利範圍第1項之方法,更包括在快取期間對於 不同時間收到名稱之多重繼續客户請求以回應該名稱之 初始請求,及增加從至少—名稱之區塊快取之子區塊數 目以回應該名稱之後續請求。 24. —種決定快取動作中資料名稱量之方法,包括·· 定義一參數N以指定快取單元;及 每次啓動-快取動作時,在該資料名稱之至少— 部分執行孩動作(N2丨),其中 佈在該資料名稱中。 “科名稱e部分分 25. 如申請專利範圍第24項之方 部分藉由以下而均勻分佈: 科名稱之⑽ 智 慧 財 產 局 員 工 消 費 將該資料名稱分成諸區塊; 將各區塊分成諸子區塊;及 於各快取動作時,從至少—名稱之各 目之子區塊。 心评稍问數 印 _ -41 - 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公釐) 26·如申請專利範圍第24項之方法,其中該快取勤作包與 知貝料與该名稱之快取部分相加。 27. 如申請專利範圍第24項之方法,其中該快取動作包招 將資料從該名稱之快取部分去除。 28. —種傳送媒體資訊之系統,該系統包括: 復數個代理伺服器,各服務數個終端裝置及當該終端 裝置 &lt; 用尸欲取得一媒體名稱時,即接收該終端裝罾之 請求,各該代理伺服器包括一快取記憶體用以儲存=八 名稱之單元,其中由各該名稱之請求頻率決定各該名^ 之單元數目,·及 與該代理伺服器連接之中央媒體伺服器,該中央媒體 伺服器具有一儲存空間用以儲存該複數個名稱,及去: 收一該代理词服器之代理請求時,從一該名稱I 料。 ’、只 29. —種傳送媒體資訊之系統,該系統包括: 經濟部智慧財產局員工消費合作社印製 複數個代理伺服器,各服務數個終端裝置及當該終端 裝置之用户欲從複數個媒體名稱取得一媒體名稱時,即 接收該終端裝置之請求,其中至少—該代理伺服器包括 -快取記憶體用以儲存至少—該名稱之數個單元,其中 在此一名稱上分佈至少一儲存名稱之單元;及 ^ 與該代理伺服器連接之中央媒體伺服器,該中央媒體 伺服器具有-儲存空間用以儲存該複數個名稱,及當』 收-該代理伺服器之代理請求時,從一該名稱提: 料。 / -42- 479200 A8 B8 C8 D8 六、申請專利範圍 30.如申請專利範目第29项之系,统,丨中該μ—代 服器在與時間無請求時段儲存至少—名稱之單元。d .如申請專利範圍第29項之系、统,其中至少—名 諸區塊要在—時序下傳送至至少一用户,而各區塊分成 諸子區塊,其中至少一代理伺服器從該至少一名稱之各 區塊快取相同數目之子區塊。 32·如申請專利範圍第29項之系統,其中至少一名稱包括 一視訊分成諸區塊要在一時序下傳送,而各區塊分成諸 子區塊,而各子區塊包括視訊框之部分資訊,其中要循 序傳送視訊框。 33. 如申請專利範圍第32項之系統,其中至少一代理伺服 器將視訊框之部分資訊與中央伺服器之此視訊框之互補 部分資訊合併成完全視訊框,及傳送完全視訊框至用 户。 34. 如申請專利範圍第3 3項之系統,其中部分資訊包括沿 著視訊框之一些掃描線之視訊資訊,而其中至少一代理 伺服器將沿著此掃描線之視訊資訊與沿著中央伺服器之 此視訊框之其他掃描線之互補視訊資訊合併成完全視訊 框,及傳送完全視訊框至用户。 35·如申請專利範圍第3 3項之系統,其中部分資訊包括在 一組取樣時間得到之視訊資訊及一第一取樣率,低於該 賀訊起源之視訊源之取樣率,而其中至少一代理词服器 將較低第一取樣率之視訊資訊與取樣時間擷取之互補視 訊資訊合併,該取樣時間與中央伺服器之此視訊框之取 -43- 本紙張尺度適用中國國家標準(CNS)A4規格(210 X 297公® ) 請 先 閱 讀 背 面 之 注 意 事 項 填 , 寫驻 t 頁I 訂 # 經濟部智慧財產局員工消費合作社印製 479200The scope of the patent application for the eyes is the same as the method of item i of the scope of the patent application, which is based on Pu Ling, Qi ... ^ The rule is determined that each corpse has sufficient bandwidth from the π physical land and is in the name of the receiving client. The server starts a cache process to pass the bandwidth and rejects the request. And if not enough 14. The method of item 13 of the scope of patent application, further includes transmitting the full content of the request name when this content is in the area storage area of the β proxy server. 15. The method according to item 13 of the scope of patent application, further comprising: when the proxy server does not have the complete content of the requested name, determine whether there is sufficient backbone bandwidth available; carry the name from the central server to 3 The proxy server is, and the request is rejected if there is not enough backbone bandwidth available. 16. The method according to item 15 of the scope of patent application, further comprising initiating a continuous cache process to adjust the cache content of the proxy server to reflect the request name. 17. As for the scope of patent application! The method also includes replacing the cached part of a special name by deleting the recent cached name. 18. The method of applying for item 1 of the scope of the patent application further includes using an up-to-date access profile as a future overview instruction to decide which names to cache. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 19. If the method in the scope of patent application for item i, it also includes pursuing each access request on the proxy server to determine which names to cache. 20 · If the method of applying for item i of the patent scope further includes the use of an up-to-date access summary as a future overview instruction to decide which names to be cached, the decision includes: -40- This paper size applies Chinese national standards (CNS) A4 specification (2iq x 297 Eight Love) 479200 The scope of the patent application defines a time window at the end of the cache replacement time; calculates the access frequency of each name in a proxy server storage area, and the access frequency is the time window Period or a function that takes part of this name; and performs a cache replacement in response to the frequency of taking names in the store. 2L The method according to item 20 of the scope of patent application, wherein the frequency of access is proportional to the sum of the time window period or its partial access to this name. 22. The method of claim 20 in the scope of patent application, in which the fetched frequency page window period or its-partial access to this-name is proportional to the time-weighted sum, and the time-weighted preference window has the most recent access. 23. If the method of applying for item 1 of the patent scope further includes receiving multiple continual client requests for names at different times during the cache to respond to the initial request for the name, and adding a sub-area for caching from at least-name blocks The number of blocks responds to subsequent requests for the name. 24. — A method for determining the amount of data names in a cache action, including defining a parameter N to specify a cache unit; and each time a cache action is initiated, at least-part of the data name is executed ( N2 丨), which is listed in the profile name. "The section name e is divided into 25. For example, the 24th part of the scope of the patent application is evenly distributed by the following: The name of the section The intellectual property bureau staff consumes the data name into blocks; Each block is divided into sub-areas Blocks; and in each cache action, from at least-the sub-blocks of each item of the name. Mindful comment number printed _ -41-This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) 26. If the method of applying for the scope of the patent item 24, wherein the cached work package and the knowledge part and the cache part of the name are added. 27. For the method of applying for the scope of the patent item 24, where the cache action The package removes the information from the cached part of the name. 28. A system for transmitting media information, the system includes: a plurality of proxy servers, each serving a number of terminal devices and when the terminal device &lt; When a media name is received, the request of the terminal device is received, and each proxy server includes a cache memory for storing a unit of eight names, where each of the names is determined by the request frequency of each name ^ The number of units, and a central media server connected to the proxy server, the central media server has a storage space for storing the plurality of names, and: when receiving a proxy request for the proxy server, From the name I. ', only 29. — a system for transmitting media information, the system includes: the consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs printed a plurality of proxy servers, each serving a number of terminal devices and when the terminal When a user of a device wants to obtain a media name from a plurality of media names, he receives a request from the terminal device, wherein at least—the proxy server includes—a cache memory for storing at least—the units of the name, where At least one unit storing the name is distributed on this name; and ^ a central media server connected to the proxy server, the central media server has-a storage space for storing the plurality of names, and when "received"-the proxy When requested by the agent of the server, please refer to the following: / -42- 479200 A8 B8 C8 D8 6. Application for patent scope 30. The system of item 29 of the Lifan Project, the system, in which the μ-server has stored at least the unit of the name in the time period without a request. D. If the system and system of item 29 of the scope of the patent application, at least-the Blocks are transmitted to at least one user in a time sequence, and each block is divided into sub-blocks, where at least one proxy server caches the same number of sub-blocks from each block of the at least one name. 32. If applying for a patent The system of scope item 29, wherein at least one name includes a video divided into blocks to be transmitted at a time sequence, and each block is divided into sub-blocks, and each sub-block includes part of the information of the video frame, which is to be transmitted sequentially Video frame. 33. If the system of the scope of patent application is No. 32, at least one of the proxy servers merges part of the information of the video frame with the complementary part of the information of the central server into a complete video frame, and sends the complete video frame to the user. 34. If the system of item 33 of the scope of patent application, part of the information includes video information along some scanning lines of the video frame, and at least one of the proxy servers will share the video information along this scanning line and the central servo The complementary video information of other scanning lines of the video frame of the device is merged into a full video frame, and the full video frame is transmitted to the user. 35. If the system of item 33 of the scope of patent application, part of the information includes video information obtained at a set of sampling times and a first sampling rate, which is lower than the sampling rate of the video source from which the message originated, and at least one of them The proxy server merges the video information of the lower first sampling rate with the complementary video information acquired by the sampling time. The sampling time and the video frame of the central server are taken. -43- This paper size applies to the Chinese national standard (CNS ) A4 size (210 X 297 male ®) Please read the precautions on the back and fill in the following pages: I # Printed by the Ministry of Economic Affairs Intellectual Property Bureau Employee Consumption Cooperative 479200 申請專利範圍 經濟部智慧財產局員工消費合作社印製 樣時間組不同,其在高於第一取樣率之取樣率之下合併 成視訊資料,及在較高取樣率傳送視訊資料至用户。 36.如申請專利範圍第3 2項之系統,其中子區塊包括一擴 充性多層系統之基本層中之資訊。 37·如申請專利範圍第29項之系統,其中單元數目係在至 少一代理伺服器之至少一名稱之取用概述函數。 土 38· —種在系統中快取用以從一中央伺服器及一代理伺服 傳送複數個資料名稱至至少一客户之方法,其中各名 分成諸子區塊要在一時序傳送至至少一客户,而各區 分成諸子區塊,該方法包括: 識別那些子區塊來自要快取之各名稱之不同區塊, 中識別出之子區塊包括在至少—名稱之區塊上分佈者. 在代理伺服器控制下快取識別出之子區塊以減少中 伺服器之位元率用以傳送名稱。 39_如申請專利範圍第38項之方法,其中在與時間鮮 時段儲存該快取子區塊。 月 40·如申請專利範圍第3 8項之方法, _ *肀涿快取從該至 一名稱之各區塊快取相同數目之子區塊。 、種電腦可讀取之儲存裝置’實施由電腦執行 式以執行系統中之快取方法,用以從_中央 代理伺服器傳送複數個資料名稱至至少—客户^ 名稱分成諸子區塊要在一時序傳送至至少其 區塊分成諸子區塊,該方法包括·· 各尸’ 識別那些子區塊來自要快取之 谷名私又不同區塊, 器 稱 其 及央 求 少 中各 而各 44- 479200 AS B8 C8 D8 k、申清專利乾圍 中識別出之子區塊包括在至少一名稱之區塊上分佈者;及 在代理伺服器控制下快取識別出之子區塊以減少中央 伺服器之位元率用以傳送名稱。 42. 如申請專利範圍第41項之裝置,其中在與時間無請求 時段儲存該快取子區塊。 43. 如申請專利範圍第41項之裝置,其中該快取從該至少 一名稱之各區塊實質快取相同數目之子區塊。 44·如申請專利範圍第41項之裝置,其中媒體名稱包括視 訊名稱,而子區塊包括視訊框,而各區塊分成要循序傳 送之視訊框,該方法更包括從中央伺服器將快取視訊框 插入一視訊框_以形成一合併_及傳送合併♦至客户。 45·如申請專利範圍第41項之裝置,其中媒體名稱包括視 訊名稱,而子區塊包括視訊框之部分資訊,其中要循序 傳送視訊框,該方法更包括將代理伺服器之視訊框之部 分資訊與中央伺服器之此視訊框之互補部分資訊合併在 完全視訊框,及傳送完全視訊框至客户。 46·如申請專利範圍第4 5項之裝置,其中部分資訊包括沿 著視訊框之一些掃描線之視訊資訊,該方法更包括將沿 著此掃描線之視訊資訊與沿著中央伺服器之此視訊框之 其他掃描線之互補視訊資訊合併成完全視訊框,及傳送 完全視訊框至客户。 47·如申請專利範圍第4 5項之裝置,其中部分資訊包括在 一組取樣時間得到之視訊資訊及一第一取樣率,低於該 資訊起源之視訊源之取樣率,該方法更包括將較低第一 -45- 本纸張尺度適用中國國家標準(CNS)A4規格(210 X 297公爱) 請 先 閱 讀 背 面 之 注 意 事 項i 再· 填 | ί裝 頁I 一 I I 訂 # 經濟部智慧財產局員工消費合作社印製 479200 A8 B8 C8 D8 六、申請專利範圍 取樣率之代理㈣器之視訊資訊與取樣時間掏取之互補 視訊資訊合併,該取樣時間與中央伺服器之此視訊框之 取樣時間組不同,其在高於第—取樣率之取樣率之下合 併成視訊資料,及在較高取樣率傳送視訊資料至客户。 48. 如申請專利範圍第45項之裝置,其中部分資訊包括一 基本層中之視訊資訊,而互補部分資訊包括一強化層中 之視訊資訊,該基本層及強化層係根據空間,訊噪A暫 時擴充性而定義。 49. 如申請專利範圍第41項之裝置,其中識別係代理處名 稱之取用概述函數。 50·如申請專利範圍第4丨項之裝置,其中在客户之任何名 稱取用之前,一平均快取方式藉由儲存各名稱之實質相 等數目之子區塊而利用代理伺服器之儲存區。 51·如申請專利範圍第4丨項之裝置,其中在客户之任何名 稱取用之前,一比例快取方式利用取用歷史資料以決定 要快取多少各名稱。 52·如申請專利範圍第4丨項之裝置,其中系統開始操作之 後,代理伺服器之快取内容會隨時改變以反映實際取用 動作。 53·如申請專利範圍第41項之裝置,該方法更包括藉由確 疋從該代理至此一客户有足夠頻寬而於接收客户之名稱 請求後,在代理伺服器開始一快取過程以傳送請求,及 若無足夠頻寬則拒絕請求。 54.如申請專利範圍第5 3項之裝置,該方法更包括當此一 -46 - 私纸張尺度適用中國國家標準(CNS)A4規格⑵〇 χ视公爱) ——r——.----_裝 (請先閱讀背面之注意事項再填寫本頁) ----訂---------. 經濟部智慧財產局員工消費合作社印製 479200 A8 B8 C8 D8Scope of patent application The sampling time groups of the consumer cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs are different. They combine video data at a sampling rate higher than the first sampling rate, and transmit video data to users at a higher sampling rate. 36. The system according to item 32 of the scope of patent application, wherein the sub-block includes information in the basic layer of an extended multi-layer system. 37. The system of claim 29, wherein the number of units is an access summary function of at least one name of at least one proxy server. Tu 38 · —A method of caching in a system for transmitting a plurality of data names from a central server and an agent server to at least one client, wherein each name is divided into sub-blocks to be transmitted to at least one client at a time sequence, While each is divided into sub-blocks, the method includes: identifying those sub-blocks from different blocks of each name to be cached, and the identified sub-blocks include those distributed on at least-named blocks. In proxy server The sub-blocks identified by the cache are cached under the control of the processor to reduce the bit rate of the server to transmit the name. 39_ The method of claim 38, wherein the cache sub-block is stored in a fresh time period. Month 40. If the method of item 38 of the scope of patent application is applied, _ * 肀 涿 cache will cache the same number of sub-blocks from each block of that name. "A computer-readable storage device" implements a computer-executable method to execute the cache method in the system for transmitting multiple data names from the _central proxy server to at least-the client ^ name is divided into sub-blocks at a time The sequence is transmitted to at least its block and divided into sub-blocks. The method includes ... Each corpse 'identifies those sub-blocks from different famous blocks in the valley to be cached, and each of them is called 44- 479200 AS B8 C8 D8 k. The sub-blocks identified in Shenqing Patent Confinement include those distributed on at least one name of the block; and the identified sub-blocks are cached under the control of the proxy server to reduce the central server's The bit rate is used to transmit the name. 42. The device as claimed in item 41 of the patent application, wherein the cache sub-block is stored in a time period without a request. 43. The device of claim 41, wherein the cache substantially caches the same number of sub-blocks from each block of the at least one name. 44. If the device under the scope of patent application 41, wherein the media name includes the video name, and the sub-block includes the video frame, and each block is divided into video frames to be transmitted sequentially, the method further includes caching from the central server Insert a video frame into a video frame to form a merge_ and send the merge to the client. 45. If the device of the scope of patent application 41, the media name includes the video name, and the sub-block includes part of the information of the video frame, in which the video frame is to be transmitted sequentially. The method further includes the part of the video frame of the proxy server. The information and the complementary part of this video frame of the central server are combined in the full video frame and sent to the client. 46. If the device in the 45th scope of the patent application, part of the information includes video information along some scanning lines of the video frame, the method further includes combining the video information along this scanning line and this along the central server. The complementary video information of the other scan lines of the video frame is merged into a full video frame, and the full video frame is sent to the customer. 47. If the device in the 45th scope of the patent application, part of the information includes video information obtained at a set of sampling times and a first sampling rate, which is lower than the sampling rate of the video source from which the information originated, the method further includes Lower first -45- This paper size applies to Chinese National Standard (CNS) A4 specifications (210 X 297 public love) Please read the precautions on the back i, then fill in | tiling page I 一 II ## Economic Ministry wisdom Printed by the Consumer Cooperative of the Property Bureau 479200 A8 B8 C8 D8 VI. The video information of the proxy server that applied for the patent scope sampling rate and the complementary video information of the sampling time are merged, and the sampling time is sampled with this video frame of the central server The time group is different, it is combined into video data at a sampling rate higher than the first sampling rate, and the video data is transmitted to the client at a higher sampling rate. 48. For the device in the 45th scope of the patent application, some of the information includes video information in a basic layer, and the complementary part of information includes video information in an enhanced layer. The basic layer and the enhanced layer are based on space, signal noise A Defined for the time of expansibility. 49. For the device in the scope of patent application No. 41, the identification is an access summary function of the name of the agency. 50. For the device under the scope of patent application No. 4 丨, in which an average cache method utilizes the storage area of the proxy server by storing a substantially equal number of sub-blocks of each name before any name of the customer is used. 51. For the device under the scope of patent application No. 4 丨, in which a proportion of caches uses the retrieval history data to determine how much of each name to cache before any name of the customer is retrieved. 52. For the device under the scope of patent application No. 4 丨, after the system starts to operate, the cache content of the proxy server will change at any time to reflect the actual access action. 53. If the device of the scope of patent application 41 is applied, the method further includes ensuring that the client has sufficient bandwidth from the agent to receive a client's name request, and then starts a cache process on the agent server to transmit Request, and reject the request if there is not enough bandwidth. 54. If the device in the scope of the application for the 53rd item of the patent application, the method further includes when this one -46-private paper size applies the Chinese National Standard (CNS) A4 specification ⑵〇χ as public love) ——r ——.- ---_ Loading (Please read the notes on the back before filling this page) ---- Order ---------. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 479200 A8 B8 C8 D8 申請專利範圍 内容在該代理伺服器之區域儲存區之中時,傳送請求名 稱之完全内容。 55. 如申請專利範圍第5 3項之裝置,該方法更包括: 當該代理伺服器不具有請求名稱之完全内容時,判定 是否有足夠可用主幹頻寬俾從該中央伺服器攜帶該名= 至該代理伺服器,及若無足夠可用主幹頻寬則拒絕= 求0 56. 如申請專利範圍第5 5項之装置,該方法更包括啓動繼 續快取過程以調整該代理伺服器之快取内容以反映請求 名稱。 57. 如申請專利範圍第41項之装置,該方法更包括藉由刪 除此一名稱之最近快取部分而取代一特別名稱之快取部 分。 58. 如申請專利範圍第41項之裝置,該方法更包括使用一 最新取用概述作爲-未來概述指示,以決定要快取取代 那些名稱。 59·如申請專利範圍第4丨項之裝置,該方法更包括在代理 词服器追蹤各取用請求以決定要快取取代那些名稱。 60·如申請專利範圍第4 1項之裝置,該方法更包括使用一 最新取用概述作爲-未來概述指示,以決定要快取取代 那些名稱,其中該決定包括: 在快取取代時間結束時定義—時間窗; 在一代理伺服器儲存區計算各名稱之取用頻率,該取 用頻率係時間窗期間或其—部分取用此一名稱之函數;及 執行快取取代以回應儲存區中名稱之取用頻率。 c請先閱讀背面之注意事項再填寫本頁) -裝 tr---------Φ, 經濟部智慧財產局員工消費合作社印制取 -47- 479200 A8 B8 C8 D8 申請專利範圍 61·如申請專利範圍第6 0項之裝置,其中該取用頻率與時 間窗期間或其一部分取用此一名稱之總和成正比。 62.如申请專利範圍第6 0項之裝置,其中該取用頻率與時 間窗期間或其一部分取用此一名稱之時間加權總和成正 比’而時間加權偏好窗中最近發生之取用。 63·如申請專利範圍第4 1項之裝置,該方法更包括在快取 期間對於不同時間收到名稱之多重繼續客户請求以回應 該名稱之初始請求,及增加從至少一名稱之區塊快取之 子區塊數目以回應該名稱之後續請求。 64·—種電腦可讀取之儲存裝置,實施由電腦執行之指令程 式以執行決定快取動作中資料名稱量之方法,該方法包 定義一參數Ν以指定快取單元;及 每次啓動一快取動作時,在該資料名稱之至少一 ι/ν 邵分執行該動作⑽D,丨中該資料名稱之1/Ν部分分 佈在該資料名稱中。 65•如申請專利範圍第64項之裝置,其中該資料名稱 邵分藉由以下而均勻分佈: 將該資料名稱分成諸區塊; 將各區塊分成諸子區塊;及 於各快取動作時,從至少_ 目之子區塊。 …各區塊選擇相同數 66·如申請專利範圍第64項之裝置,其中該 將資料與該名稱之快取部分相加。 &quot; 作包括 L___________ - 48 - 本纸張尺度剌帽时鮮 填 頁 訂 麗 等 1· / 經濟部智慧財產局員工消費合作社印製 「、申請專利範圍 67.如申清專利範圍第6 4 ;之奘 甘rb、n、、 負 &lt; 裝置,其中茲快取動作包括 和貝枓仗孩名稱之快取部分去除。 68· —種傳送電腦可執行v私式 &lt; 万法,以在系統中執行 一快取過程用以從一中 甲夬伺服詻及一代理伺服器傳送複 數個媒體資料名稱至至少_ 甘士々々 ^ ^ ^ 各尸,其中各名稱分成諸子&amp;塊要在一時序傳m 5 &gt; ^ A 斤傅迗至至少—客尸,而各區塊分成諸子 區塊,該方法包括: 傳送一指令程式至一客户裝置; 藉由此程式令客户裝置執行以下過程要識別那些子 塊來自要快取之各名稱之不同區塊,其中識別出之子 塊包括在至少一名稱之區塊上分佈者;及 在代理伺服器控制下快取識別出之子區塊以減少中 伺服器之位元率用以傳送名稱。 69· 申叫專利範圍第6 8項之方法,其中程式令在與時 典請求時段儲存該快取子區塊。 7〇·如申請專利範圍第68項之方法,其中程式令該快取從 孩至少一名稱之各區塊實質快取相同數目之子區塊。 71·如申清專利範圍第6 8項之方法,其中媒體名稱包括視 訊名稱,而子區塊包括視訊框之部分資訊,其中要循 傳送視訊框,其中程式令客户裝置藉由此程式更執行 下和代理词服器之視訊框之邵分資訊與中央伺服器 此視訊框之互補部分資訊合併成一完全視訊框及傳 全視訊框至客户。 72·如申请專利範圍第7 1項之方法,其中部分資訊包 請 先 閱 讀 背 面 之 注 意 事 區 區 填 寫裝 本衣 頁 I 央 間 序 以 之 送完 括 訂 -49- 本纸張尺度適用中國國家標準(CNS)A4規格(21〇 x 297公釐 479200 申請專利範圍 基本層中之視訊資訊,而 之禎訊资#,兮I + a 補#刀貝矾包括一強化層中 視訊貝㉛该基本層及強化層 時擴充性而定義。 「1 w &gt; a s 73·如申請專利範圍第6 8 ^ . ^ m 4 ^ ^ 万决,其中識別係代理處名 稱之取用概述函數。 74· —種傳送電腦可執行指 V柱式&lt;万法,以執行一快取動 作中決疋-資料名稱量之過程,該方法包括: 傳送一指令程式至一客户裝置; 令客户裝置藉由此程式執行以下過程·· 定義一參數N以指定快取單元;及 每次啓動-快取動作時,在該資料名稱之至少一_ 部分執行該動作(N2 1),其中嗲资拉々γ、 佈在該資料名稱中。 “枓名%〈而部分分 乃.如申請專利範圍第74項之方法,其中該資科名稱之 部分藉由以下而均勻分佈: 將該資料名稱分成諸區塊; 將各區塊分成諸子區塊;及 數 經 濟 部 智 慧 財 產 局 員 工 消 費 合 作 社 印 製 於各快取動作時,從至少一名稱之各區塊選 目之子區塊。 其中該快取動作包括 其中該快取動作包括 76·如申請專利範圍第74項之方法 將資料加入該名稱之快取部分。 77·如申請專利範圍第74項之方法 從該名稱之快取部分去除資料。 _ -50- 本纸尺度剌中義家標準(CNS)A4規格⑽χ挪公爱)When the content of the patent application is in the regional storage area of the proxy server, the full content of the request name is transmitted. 55. If the device of claim 53 is applied, the method further includes: when the proxy server does not have the complete content of the requested name, determining whether there is sufficient backbone bandwidth available; carry the name from the central server = To the proxy server, and reject if there is not enough available backbone bandwidth = seeking 0 56. If the device in the scope of patent application for item 55 is applied, the method further includes starting the continuous cache process to adjust the cache of the proxy server Content to reflect the request name. 57. If the device of the scope of patent application 41 is applied, the method further includes replacing the cache part of a special name by deleting the most recent cache part of this name. 58. In the case of a device applying for item 41 of the patent scope, the method further includes using a latest access overview as a -future overview instruction to decide which names to cache to replace. 59. As for the device under the scope of patent application, the method further includes tracking the fetch requests at the proxy server to decide which names to cache. 60. If the device of the scope of patent application 41 is applied, the method further includes using a latest fetch overview as a -future summary instruction to decide which names to cache to replace, wherein the decision includes: at the end of the cache replacement time Definition—time window; calculating the access frequency of each name in a proxy server storage area, the access frequency being a function of the time window or its partial access to this name; and performing a cache replacement in response to the storage area The frequency of the name. c Please read the notes on the back before filling this page)-Install tr --------- Φ, printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs -47- 479200 A8 B8 C8 D8 Patent Application Scope 61 · For a device in the 60th scope of the patent application, the frequency of access is directly proportional to the sum of the names used during the time window or part of the time. 62. The device of claim 60 in the scope of patent application, wherein the frequency of access is directly proportional to the time-weighted sum of the time window period or a part of the time that this name is used, and the most recent access in the time-weighted preference window. 63. If the device of the scope of patent application No. 41 is applied, the method further includes receiving multiple continuation customer requests for the name at different times during the cache to respond to the initial request for the name, and increasing the number of blocks from at least one name. Take the number of child blocks in response to subsequent requests for the name. 64 · —A computer-readable storage device that implements a program executed by a computer to execute a method for determining the amount of data names in a cache action. The method package defines a parameter N to specify a cache unit; When the action is cached, the action ⑽D is performed at least one ι / ν of the data name, and the 1 / N part of the data name is distributed in the data name. 65 • The device according to item 64 of the patent application scope, wherein the data name is divided evenly by: dividing the data name into blocks; dividing each block into sub-blocks; and during each cache action From at least _ head of the child block. … Choose the same number for each block 66. For the device under the scope of patent application 64, which adds the data to the cached part of the name. &quot; Work includes L___________-48-This paper is scaled when the cap is freshly filled and ordered, etc. 1 / Printed by the Consumers' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, "Application for patent scope 67. Such as application for patent scope 6 4; Rb, rb, n, and negative &lt; devices, where the cache action includes the removal of the cache part of the name of the child and the child. 68 · — a transmission computer executable v private type &lt; A cache process is used to send multiple media data names from at least one Chinese server and one proxy server to at least _ Gan Shi 々々 ^ ^ ^ Each corpse, where each name is divided into sons & blocks at a time Sequential transmission of m 5 &gt; ^ A is equal to at least-guest corpse, and each block is divided into sub-blocks. The method includes: sending a command program to a client device; using this program to cause the client device to perform the following process: Identify those sub-blocks from different blocks of each name to be cached, where the identified sub-blocks include those distributed over at least one name of the block; and cache the identified sub-blocks under the control of a proxy server to reduce the number of server The bit rate is used to transmit the name. 69 · The method of claiming item 68 of the patent scope, wherein the program order stores the cache sub-block during the request period with the code. 70. The method of item 68 of the patent scope , Where the program causes the cache to substantially cache the same number of sub-blocks from each block of at least one name. 71. If the method of claim 68 is claimed, the media name includes the video name, and the sub-blocks Include part of the information of the video frame, in which the video frame must be transmitted, in which the program enables the client device to execute the program and the sub-information of the video frame of the proxy server merges with the complementary part of the video frame of the central server. Make a complete video frame and pass the full video frame to the customer. 72. If you apply for the method of item 71 of the patent scope, please read the note on the back of the page to fill in the page of the book.订 定 -49- This paper size applies to the Chinese National Standard (CNS) A4 specification (21x 297 mm 479200).资 # , 曦 I + a 补 # Sword and shell alum are defined by the expansion of the basic layer and the enhanced layer in a reinforced layer. "1 w &gt; as 73 · As the scope of patent application 6 8 ^. ^ m 4 ^ ^ Wan Ju, where the identification is an overview function of the name of the agency. 74 · — A transmission computer can execute the V-column &lt; Wan Fa to perform a cache action-the amount of data name The method includes: sending a command program to a client device; causing the client device to perform the following process by this program ... defining a parameter N to specify a cache unit; and each time a cache operation is started, the data is At least one part of the name performs the action (N2 1), of which 嗲 拉, γ, is placed in the name of the material. "Anonymous %% and part of the method is. For example, the method of the scope of patent application No. 74, wherein the part of the asset name is evenly distributed by: Dividing the data name into blocks; Dividing each block into sub-areas Block; and the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, printed sub-blocks selected from each block of at least one name during each cache action. The cache action includes the cache action including 76 · 如The method of applying for the scope of patent scope item 74 adds data to the cached part of the name. 77. If the method of applying for the scope of patent scope 74 removes the data from the cache part of the name. _ -50- Standard of this paper: Zhongyijia Standard (CNS) A4 specifications
TW89118815A 1999-09-14 2000-12-11 Method and apparatus for caching for streaming data TW479200B (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US15390599P 1999-09-14 1999-09-14

Publications (1)

Publication Number Publication Date
TW479200B true TW479200B (en) 2002-03-11

Family

ID=22549215

Family Applications (1)

Application Number Title Priority Date Filing Date
TW89118815A TW479200B (en) 1999-09-14 2000-12-11 Method and apparatus for caching for streaming data

Country Status (1)

Country Link
TW (1) TW479200B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101689191A (en) * 2005-09-12 2010-03-31 惠普开发有限公司 The system and method for process information or data on computers

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101689191A (en) * 2005-09-12 2010-03-31 惠普开发有限公司 The system and method for process information or data on computers

Similar Documents

Publication Publication Date Title
US7028096B1 (en) Method and apparatus for caching for streaming data
US11089358B2 (en) Method of unscrambling television content on a bandwidth
CN102740159B (en) Media file storage format and self-adaptation transfer system
CN102439578B (en) Dynamic variable rate media delivery system
CN102598688B (en) Streaming encoded video data
CN102282804B (en) Adaptive network content distribution system
CN106209892B (en) Streaming is requested using the enhanced block of scalable coding
JP4619353B2 (en) System for distributing multimedia content
Acharya et al. Middleman: A video caching proxy server
US20070078954A1 (en) Apparatus and method for providing high speed download service of multimedia contents
CN1484920A (en) A method and apparatus for acquiring media services available from contnt aggregators
Shen et al. An information-aware QoE-centric mobile video cache
Zink Scalable video on demand: adaptive internet-based distribution
Acharya Techniques for improving multimedia communication over wide area networks
TW479200B (en) Method and apparatus for caching for streaming data
US20240121287A1 (en) Extended w3c media extensions for processing dash and cmaf inband events
CN101465861A (en) System, method and computer programming product for transmitting and/or receiving medium current
Min et al. Design of an adaptive streaming system in ubiquitous environment
Friedman et al. A framework for caching multimedia objects in the Internet

Legal Events

Date Code Title Description
GD4A Issue of patent certificate for granted invention patent
MM4A Annulment or lapse of patent due to non-payment of fees