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

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_; __ 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 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

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-

The size of this paper is CNS A4 (210 X 297 male iT 479200)

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, ... 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 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

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

Case B Case c (Please read the note on the back before filling this page)

-1 n I installed ·. 27 This paper size applies to China National Standard (CNS) A4 (210 X 297 Public Love 479200)

V. 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 W

W

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 A7

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 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

V. 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, 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 Α7

Few 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

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 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 D8 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 Complementary 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 The 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 Scope 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 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