JP2002525749A - Internet caching system, method and system configuration - Google Patents

Abstract

(57) [Summary] The present invention relates to an Internet caching system and a configuration and a method provided for a request for an Internet information file in the Internet caching system. The system is built as a two-tier caching system. To reduce the load on the central cache server (130), the intermediate configuration (110) interconnects the system's local server (100) with the central cache server (130). Configuration 110 communicates with a local cache server according to a protocol used for communication between cache servers. Upon requesting an Internet information file from a central cache server, configuration 110 utilizes Structured Query Language. Thus, the central cache server (130) is primarily dedicated to replying to normal SQL queries.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001] Technical Field of the Invention The present invention, the Internet caching system (Internet caching system
), And an arrangement and method for responding to requests for Internet information files in an Internet caching system.

[0002] is the background art Internet and the present day the most familiarize feature, the World Wide Web (WWW) is, in recent years, has evolved into a source of huge information. Anyone can supply any information, such as text, photos, sound and video, to the World Wide Web, where they can be easily retrieved by users anywhere in the world, as long as they have access to the Internet. obtain.

[0003] A significant problem facing the Internet is the ever-increasing demand for communication capacity when users access information from anywhere in the world. It has been estimated that the world wide web communication traffic on most international communication lines has already surpassed all conventional telephone and facsimile traffic. Although more transmission and switching capacity is continually increasing, it is a slow and expensive process and demand continues to outpace supply.

[0004] The content of the World Wide Web is becoming unmeasurable, perhaps hundreds of terabytes (as of the summer of 1998). However, a relatively small subset of all this information is a huge part of the information that is actually being accessed. Thus, in order to limit the amount of information that must be transmitted over the Internet, with the aim of minimizing the bandwidth used when accessing information on the Internet and the required latency (latency). And
Different caching techniques are currently used to limit the distance over which information is transmitted.

In the field of caching WWW objects and Internet information files, there are two approaches: client-side caching and server-side caching. The simplest method of client-side caching is today
It is substantially used by W browsers. The browser is
Maintains a cache on the user's computer along with the last accessed Internet information file. When a user wants to access a particular information file a second time, the browser retrieves it from his cache instead of requesting it over the Internet.

A proxy server caching method, which is another client-side caching method, can be used to help a nearby user. In this scheme, the cache is populated with WWW proxy nodes to which a number of nearby users are connected. Such a proxy node may be, for example, a server existing in a company. When a WWW client wishes to access a WWW server on the Internet, the client sends an http request to a proxy node or WWW proxy server, rather than directly to a global Internet server. Instead, it is the proxy server that sends the request to the global Internet, caches the response, and returns the response to the client. Thus, when an information file is first requested, it is transmitted over the Internet and stored in the cache of the WWW proxy server. That W
Subsequent requests for the same information file from any client connected to the WWW proxy server can now be resolved locally, rather than making an http request to the WWW server over the global Internet. Proxy server caching is also used outside the organization of a company or any other organization by implementing the above scheme on a regional Internet cache server to which a number of clients are directly or indirectly connected. obtain.

[0007] Depending on the size and homogeneity of the user community utilizing the server's cache, about 20-40 gigabytes of cache storage can be achieved (as of spring 1998).
30-50 Internet traffic created by the user community
% Decrease. As the information provided by the Internet and the WWW continues to grow, the cache size required over time will increase to maintain the hit rate, ie, the percentage of the requested information file transmitted from the cache server. There is a great deal that will have to increase. Furthermore, if the hit rate is 7
An increase of 5% or more would provide significant benefits to Internet performance and availability. In terms of typical end-user behavior, this requires a much larger cache, currently on the order of 200-400 gigabytes, and is now hundreds of thousands, but very large in the end-user community. Needs many members. The reason is that the larger the end-user community, the more likely it is that someone else in the community will access the necessary files first, especially if the users share some common interests. This is because the probability of being present is greater.

[0008] Installing a large cache is easy with the right computer and the right disk space. However, the cache is also required to be able to handle all requests from the end users involved. With current technology, a single processor computer cannot meet the demands of hundreds of thousands of end users. Therefore, several systems have been proposed to address this problem, and are outlined here under the name of their main proposer.

[0009] Cisco Systems Inc. has a group or dedicated cache appli- cation for all WWW requests.
es) "Farm" or "Cache Engines"
It is proposed that all end-users be connected to a backbone router that is programmed to redirect users unknowingly. Each cache engine handles a subset of all base WWW servers based on IP (Internet Protocol) address groupings. This solution can scale up to 32 cache engines in parallel. This equates to approximately 500,000 contracted end users.

[0010] Inktomi Corporation is a so-called Layer 4
A switch, called a (layer 4) switch, is suggested to be used to redirect all requests for WWW pages to an "Inktomi Traffic server". Powerful computers are used, all sharing the same disk storage system. This solution can scale up to 16 workstations in parallel. This is also equivalent to serving approximately 500,000 contracted end users. However, having several computers accessing the same disk storage system adds complexity and management, i.e., some of the power of each computer is unavailable for processing demands.

[0011] Network Appliance Incorporated
c.) proposes a multi-row caching solution arranged in layers. This system has several local caches near the end user. These local caches communicate with the central cache using the Internet Cache Protocol (ICP) in the event of a local-level miss. If the requested file is in the central cache, it is transmitted to the local cache and then transferred to the end user. If the requested file is not in the central cache, the central cache makes a request to the origin server, causing the file to be transferred to the local cache, which in turn causes the file to be transferred to the end user. The central cache thus handles ICP requests from the local cache, and communicates with the origin server in the event of a cache miss in the central cache. For scale-up, several central caches, each handling a subset of the origin servers, may exist in parallel. This means that the local cache can direct each request to the correct central cache server. Since this protocol is not standardized, that means that all local caches must be delivered from Network Appliance Incorporated.

All of these solutions have the disadvantage that the central cache server needs to handle a wide range of communications in different ways. This results in reduced availability of server capabilities and difficulty in handling the hundreds of thousands of users required to achieve high hit rates. By adding more servers, the system becomes more expensive and complex. The complexity of the system results in added overhead and, therefore, reduced availability of the relatively expensive resources represented by the server.

[0013] An object of the present invention is to overcome the shortcomings of the presently known techniques for caching information files on the Internet, and, in a cost-effective approach, the information file Providing a solution for caching.

Another object of the present invention is to provide a solution in which a user's request for a cached information file is provided in a quick and cost-effective manner by a caching system.

[0015] Yet another object is to provide a solution to the problem of a cache server that can better handle the growing number of information files provided by the Internet and the World Wide Web.

Yet another object is to provide a solution for obtaining a high hit rate percentage for requests for information files directed to the caching system at a minimum cost.

[0017] Yet another object is to provide a scalable caching system that is scalable in a standardized manner.

The above objects are realized by an internet caching system and method for a request for an internet information file in an internet caching system according to the claims.

According to a first aspect of the present invention, there is provided a method for serving an Internet information file request in an Internet caching system, the method comprising, at a local Internet cache server, a user requesting an Internet information file from a user. Receiving a request; querying the information file in response to the received request if the information file is not cached by the local server; seeking an information file in response to a reply to the query Generating a file request, wherein if the reply indicates that the central file server storing the cached Internet information file is caching the information file, the file request is Sending the information file to the central file server from the feeder unit in response to the file request, thereby reducing the load on the central file server. .

According to a second aspect, there is provided an arrangement in an Internet caching system, the system including at least one local cache server and at least one central file server, both servers being cached. An internet information file is stored and arranged to include a feeder in communication with the local cache server and the central file server to reduce the load on the central file server, wherein the feeder is configured to receive Internet information from the local cache server. First means for receiving a request for a file;
A second means for deriving an inquiry from the alphanumeric string received from the local cache server; and a third means for inquiring the central file server of the Internet information file using the inquiry derived by the second means. Means.

According to a third aspect, there is provided an Internet caching system, the system comprising: a set of local Internet cache servers, each configured to receive a request for an Internet information file from a user; At least one central file server included in a cache site for storing cached Internet information files; and feeder means for interconnecting the set of local cache servers and the central file server, comprising: Means for communicating with at least one local cache server in accordance with the protocol used for the communication of, and retrieving an Internet information file from said central file server using a database query And feeder means including at least one feeder that reduces the load on the central file server.

The present invention is based on the idea of connecting a number of dedicated computers to a central file server or central cache server that stores Internet information files. In connection with the central cache server, these additional computers are low performance (low end) computers. The dedicated computer is configured to reduce the load on the central cache server by performing some of the tasks normally handled by the central cache server itself. In this way, the central cache server provides the local cache server connected to the central server, more precisely via a dedicated computer, to the central server in a fast and cost-effective manner. be able to. Expensive hardware that forms the actual central file server and file repository where files are cached, while specialized inexpensive machines around the file server perform time-consuming and time-critical tasks in parallel (repository) is used to the maximum.

As a result, the feeder means or the feeder of the present invention is realized by a machine different from any machine realizing the central file server. This reduces the burden on the central file server, which allows the central file server to occupy additional processing time for the actual retrieval of cached files. Thus, the central file server can serve a large community of users in an effective manner. The number of served users may increase as the user's requests are served more effectively through requests to the local cache server. This allows the central file server to gradually gain a higher hit rate percentage for its cache.

According to the embodiment of the present invention, the feeder means communicates with the local cache server instead of the central file server according to a protocol used for communication between the Internet cache servers. The commonly used protocol is either the Internet Cache Protocol (ICP) or Cache Digest, but may be any other conventional or future protocol used for the same purpose. In this way, by offloading the task of receiving and replying to queries and / or requests for information files to a machine different from the central file server machine, the load on the central file server is significantly reduced.

When the local cache server receives a request from a user for an information file that is not cached on the local server, the local server initiates an inquiry about the file. In one embodiment, the query is sent to a table or database inside or directly connected to the local server. If the table indicates that the queried file is cached by the central file server, the local server will request the file from the feeder means or feeder. The query and request are then preferably performed according to the cache digest protocol. However, a request from the local server to the feeder, such as a request from a user to the local server, is, for example, an HTTP request.
It may follow any layer three protocol.

In another embodiment, a query from a local server is sent to a feeder. In a query, for example, an ICP query, the URL of the queried information file is included. The feeder derives the inquiry number from the alphanumeric URL of the inquiry about the received information file. This query number is used by the feeder to query the central file server for the information file. The feeder queries the file server for the information file using a standard SQL query (structured query language). If the queried file is on the central file server, ie, there is a cache hit, the queried file is transmitted from the central server to the local server via the feeder. Having the central file server perform file transmission in response to an SQL query, rather than in response to a query from a local cache server, such as an ICP query, means that the central file server contains a significant amount of files. ing.

Alternatively, a query number is derived from the alphanumeric URL and from a portion of the header information included in the query. This part of the header information contains the specific user information of the originating requester. This specific user information is
For example, the language used by the user, and the central file server can respond according to this specific information. The query number corresponding to the information file is derived by using any hash algorithm, preferably the MD5 hash algorithm.

In an embodiment where the local server makes an internal query for the information file, the feeder derives the query number from the next request sent by the local server to the feeder. The alphanumeric string used to derive the inquiry number is the string included in the above request, for example, in the URL of an HTTP request. When querying the central file server about the information file, preferably using an SQL query, the query number is used by the feeder. Further, it is convenient to include at least a part of the header information field of the request as a criterion for deriving the inquiry number.

[0029] To further reduce the load on the central file server, the feeder preferably includes a table that stores information about each information file cached by the central file server. This table is, for example, a memory resident MD
5 is a hash table with an index of 5; By searching the table, the feeder can determine whether or not the queried information is cached by the central file server without having to query the server, and thus respond to queries from the local server. Feeders can respond faster.

According to another embodiment of the present invention, the Internet caching system further includes an updating unit or an updating unit for updating the plurality of information files cached by the central file server. The update procedure involves transmitting a copy of the file cached on the local server to the central server. As a result of a cache miss at the central server when querying for a file, the transmitted file is a file retrieved from the base server by the local server and then a file cached by the local server.

In this way, the central file server, or central cache server, does not retrieve files that are not cached by itself, and therefore, when the central file server, or central cache server, is serving local cache server requests. There is no obligation to request a file from the base server for a cache miss. Instead, when evaluating a query from the local server for an information file and determining that the queried file is not cached on the central file server, the feeder is querying for a reply indicating that the file is unavailable. Send to the local server and then tell the updater to update the central file server. Upon receiving a reply indicating a cache miss, the local cache server retrieves the corresponding file from the base server. Upon receiving the command to update the central file server, the update unit requests a copy of the file from the local server and thereby transmits the received copy of the file to the central cache server where the file is stored. . Preferably, the transmission and storage procedures are performed when the overall load on the central file server is low or when the local server is given enough time to retrieve the file from the base server.

However, if the local server is located behind a firewall, the update unit requests a copy of the file from the origin server. The copy is then stored at the central cache server. In this case, it is preferable that the feeder does not instruct the updating unit to start the updating procedure until a certain number of queries for a specific information file are received. In this case, these queries originate from a local server located behind the firewall. Preferably, the updating unit is operated by a machine different from the machine that moves the feeder, and further by a machine different from any of the file server machines. This is convenient because file requests to the origin server, for example HTTP requests, cannot be predicted in duration, and therefore load on the machine executing the request cannot be predicted. However, even in a simplified system where the updater can be run on the same machine that runs the feeder, the machine is different from any central file server machine.
The machine running the updater and feeder connects the local cache server and the central file server, and in an embodiment where the machine itself is not included in the central cache site with the central file server, these machines are the central file server machines Obviously, it is different.

Certain Internet information files are not suitable for caching. Such files are often referred to as dynamic information files, and the term dynamic comes from the fact that these files are continuously updated at the origin server. Examples of such files are files for stock price information, weather forecasts, and the like. One preferred method of handling when dynamic files are present is to maintain a list of known non-cacheable files either at the updater or at the local server. In this way, communication in the system is minimized as a result of the user requesting such a file.

In yet another embodiment of the present invention, several central file servers are included in a central cache site, and each file server caching information file includes, within a defined range, an originating host name. , IP address, or derived inquiry number. The feeder addresses the query to the appropriate range of the file server caching file, as it is based either on the source hostname, IP address, or derived query number of the requested information file. With this scalable solution, each file server has its own disk system, thus minimizing overhead. Further, because of the standardized protocol used by the site, the central cache site is scalable by a third party file server.

To speed up the communication between the central file server and the low-end computers, ie between the feeder and the updater, each low-end computer is preferably dedicated, if there are several file servers instead, if there are several file servers. It is connected to a central file server by a network. This network is either a private network or a public network. In the latter case, at least part of the network capacity is preferably reserved for the communication.
Of course, the network used may be part of the Internet or a non-dedicated line. The type of communication used between the central file server and the low-end computer depends on whether the low-end computer, or the feeder and the update unit, are located at the same site as the central file server or at a different location than the central file server. It depends very much on what is being done.

Further, preferably, the central cache site is served to a defined set of local cache servers, and the set of local cache servers is served to linguistically and culturally homogeneous communities. This further increases the hit rate percentage at the central cache level, as the same information file is likely to be requested more than once.

By using the present invention, an operator of the Internet cash system handling information file requests according to the present invention can quickly and inexpensively provide a large number of subscribing customers.
In addition, a method of efficiently serving can be provided. The customer is preferably a different internet service provider, company or other organization, and the customer is connected to the central cash site of the invention, or the feeder / updater of the invention, with his local cash server, or , As a customer, to the entire cache system of the present invention, including a central cache site including a feeder and an update unit, and a local cache server connected thereto. Of course, the customer may be a single user making up a single WWW client connected directly to the system of the present invention. Also, large companies or Internet service providers may choose to operate the system of the present invention on their own system rather than being connected to a system operated by another. Further, since the cache system of the present invention is built around standardized protocols such as ICP or SQL, as long as these protocols are supported, a local cache server from any manufacturer and a centralized File servers are included in this system.

Within the scope of the present invention, a local Internet cache server is to be understood as a proxy node, preferably a WWW proxy node, which maintains a cache for users or WWW clients connected to the proxy node.

What is cached at the local Internet cache server or file server at the central cache site is any non-dynamic file accessible on the Internet and containing any type of information. Thus, the term Internet information file as used in the present invention includes many different types of files and their different names (eg, binary files, text files, image files, audio files, video files, HTTP (hypertext transfer). Protocol) files, WWW files, FTP (file transfer protocol) files, WWW pages, WWW objects, etc.). In addition to files accessed using the HTTP or FTP protocols, any file accessed on the Internet by any layer 3 protocol is included in the term Internet information file. Another example of a protocol that can be used is the WTP (Wireless Transport Protocol) protocol used in the WAP (Wireless Application Protocol) standard.

[0040] According to a fourth aspect of the invention, the invention includes a computer-readable medium. This medium stores one or more programs of instructions for one or more general-purpose computers. The medium includes means for causing one or more general-purpose computers to perform the steps recited in claims 1-17.

According to a fifth aspect of the invention, the invention relates to one or more sequences of instructions for one or more general-purpose computers for performing the steps recited in claims 1-17. Including a program storage device.

The above and other aspects, features and advantages of the present invention will be more fully understood from the following description of embodiments thereof, taken in conjunction with the accompanying drawings.

Detailed Description of the Preferred Embodiment One embodiment of the invention is described with reference to the block diagram shown in FIG. Figure
At 1, a number of local cache servers 100 are shown. These cache servers 100 are connected to feeder means 110 via the Internet. Here, one feeder 110 is shown as an example. The number of feeders 110 and the number of local cache servers 100 shown in FIG. 1 are merely examples, and the present embodiment is not limited to these numbers.

However, regardless of the number of feeders, each feeder is connected to a single central file server in this embodiment. In FIG. 1, the feeder 110
Are connected to the central file server 130. This central file server includes a storage medium (not shown) in which Internet information files are stored, ie, cached, and is embodied by a high-performance computer such as Sun Ultraspark or DEC Alpha Computer. On the other hand, each feeder 11
0 is embodied by a low-performance computer, such as a normal personal computer, and constitutes a front-end machine that handles communication between the local cache server 100 and the central file server 130.

The feeder 110 communicates with the local cache server 100 using the Internet cache protocol. The Internet cache protocol is a message-based protocol used for communication between cache servers via the Internet. Thus, feeder 110 responds to ICP queries for cached Internet information files. This query is received by one of the local cache servers 100 along with the ICP reply. This ICP reply indicates either a cache hit (ICP_OP_HIT) or a cache miss (ICP_OP_MISS).

According to the Internet cache protocol, the I received by the feeder
The CP inquiry includes the URL of the queried information file. This URL
, The feeder 110 derives the inquiry number corresponding to the information file using the MD5 hash algorithm. Then, using this query number, an MD5 indexed hash table 115 resident in memory.
Is searched. RAM (random access memory) in feeder 110
116, where the indexed table is stored. This indexed table includes an entry for each query number corresponding to an Internet information file cached at central file server 130. Searching the indexed table includes searching for an entry with a query number that matches the derived query number. If a matching query number is found in the table, this indicates that the queried information file was cached by the central file server 130, so that an ICP reply to the local server 100 indicates a cache hit. . Also, if a matching query number is not found in table 115, this indicates that the queried information file is not cached by central file server 130. As a result, the ICP reply indicates a cache miss.

As means for deriving the inquiry number using the MD5 hash algorithm and searching the indexed table, there is a microprocessor 120 included in the feeder 110. The microprocessor 120 has appropriate software modules mounted thereon. The microprocessor executes the software module, which results in the derivation of a query number and a search of the indexed table 115. The use of this software module is readily conceivable to a person skilled in the art of programming techniques.

If the reply from the feeder 110 to the local server 100 indicates a cache hit, the local server requests an information file from the feeder using Hypertext Transfer Protocol (HTTP). HTTP is a protocol for accessing a WWW object via the Internet. That is, the HTTP request including the URL of the requested file is transmitted to the feeder.

When communicating with the central file server 130, the feeder 110
Use QL queries. When the HTTP request is received, the feeder extracts the inquiry number derived in advance from the URL of the corresponding ICP inquiry.
Alternatively, the URL of the HTTP request is reused for query number derivation. The feeder uses the inquiry number in the standard SQL inquiry sent to the central file server. In response, the central file server 130 transmits the information file to the feeder 110, and the feeder subsequently transmits the information file to the local server 100 that issued the request for the file.

If the reply from the feeder 110 to the local server 100 indicates a cache miss, the local server generates an HTTP request to a base server (not shown) where the requested file exists, and caches the received file. Then, a copy of the file is transmitted to the requesting user (not shown).

The means for implementing the implementation of the Internet Cache Protocol (ICP) is a microprocessor 120 included in feeder 110. The microprocessor serves not only as a means for inquiring the central file server 130 using SQL, but also as a means for receiving an HTTP request from the local server 100. The operation of the microprocessor is controlled by a software module that is part of the above means. The adoption of these software modules is well known to those skilled in the art of programming techniques and familiar with the protocol in question.

FIG. 2 shows another embodiment of the Internet cache system according to the present invention. 2 includes a central file server 230 and a feeder 210.
And connected to the local cache server 200 via the Internet.
It differs from the one shown in FIG. 1 in that it has an updating unit 240 which is a so-called updating means. Therefore, in FIG. 2, the updating unit 240 is represented by an arrangement included in each component similarly to the feeder 210.

In addition to what is described below with respect to the components of FIG. 2, the components of FIG. 2 that correspond to the components of FIG. 1 operate and interact according to the content already described with reference to FIG. Therefore, only the features of the components related to the embodiment shown in FIG. 2 will be described below.

The role of the updating unit 240 is to update a recording medium (not shown) associated with the central file server 230 having a new cache information file. As described with reference to FIG. 1, when the local server 200 receives a cache miss in the ICP reply from the feeder 210 as a response to the previous ICP inquiry, the local server 200 sends an HTTP request for a file to its base point. Create for server (not shown). The requested file is stored on the local server 20
0 and cached. After a predetermined amount of time, as a result of reporting a cache miss in the ICP reply, feeder 210 instructs updater 240 to update the central file server.

The updating unit 240 sends the URL of the file inquired from the feeder 210
And the identity of the local server that queried the file
And receive. Then, an HTTP request for a file is generated, and the updating unit 240
To a specific local server. Upon receiving the requested file, the update unit stores or caches the file at central file server 230. When the file is stored, the updating unit instructs the feeder to add the inquiry number corresponding to the file to the indexed table 215 stored in the RAM area 216.

As a means for requesting an information file from the local cache server 200 and a means for caching the information file received in the central file server 230, there is a microprocessor 260 included in the update unit 240 and equipped with an appropriate software module. . The adoption of software modules is well known to those skilled in programming techniques.

An example of the execution procedure of the local cache server 200 in the embodiment of FIG. 2 will be described with reference to the flowchart of FIG.

In step 300, the local cache server 200 receives a request for an Internet information file from a client under a specific local cache server. However, the file request is also received from the updating unit 240. This is processed according to the description with reference to FIG. Next, in step 301, the local cache server searches the locally cached file for the requested file. If the file is found, the file is transmitted to the requesting client or updater 240. This procedure is shown in step 302.

If the local cache server 200 cannot find the requested file, ie, has not cached the requested file, then in step 303, the local cache server 200 determines whether the request came from the update To inspect. If this condition is true, a message to the effect that the requested file cannot be used is returned to the updating unit in step 304. If the condition in step 303 is false, that is, if the request is from a client, an ICP inquiry is sent to the feeder 210 in step 305. Subsequently, at step 306, the local cache server receives from the feeder 210 an ICP reply indicating whether the central file server 230 has cached the requested file. At step 307, the ICP reply is evaluated.
If the reply indicates a cache miss, that is, if the requested file is not cached on the central file server, the local cache server 20
0 generates an HTTP request for the file and sends it to the origin server for the file. On the other hand, if the reply indicates a cache hit, the local cache server generates an HTTP request for a file from the feeder 210. This procedure is shown in step 309. Next, at step 310, the local cache server receives the requested file from feeder 210. Finally, in step 311, the file is transmitted to the client that requested the file.

Hereinafter, the operation executed by the feeder 200 in the embodiment shown in FIG. 2 will be described with reference to the flowchart in FIG.

At step 400, feeder 210 receives an ICP inquiry for an Internet information file from one of local cache servers 200 commanded by the feeder. This query includes the URL of the queried information file. In step 401, the feeder 210 derives an inquiry number from the URL using the MD5 hash algorithm. This inquiry number is stored in the memory 216 of the feeder 210 in step 402.
Used to look up the indexed MD5 hash table present in it.

If this number is not found when searching the hash table, the feeder sends an ICP reply indicating a cache miss at step 403 to the ICP
The inquiry is returned to the received local cache server 200. Thereafter, in step 404, the feeder 210 forwards the URL of the queried file to the updating unit 240, thereby instructing the updating unit to retrieve the queried file that has not been cached. In step 405, the feeder 210 adds an inquiry number corresponding to the queried file to the indexed hash table 215. This is because the queried file is transmitted from the local server 200 to the central file server 230.
Is performed in response to the update unit 240 notifying the feeder that the data is stored in the feeder. Further, the operation of the updating unit 240 will be described with reference to FIG.

If the feeder 210 finds the inquiry number when searching the hash table 215 in the conditional step 402, it will retrieve an ICP reply indicating a cache hit in step 406, and the local cache server from which the ICP inquiry was received. Reply to 200. Thereafter, in step 407, the feeder receives an HTTP request from the local cache server 200 that has previously issued the ICP inquiry. HTT as in the case of ICP inquiry
The P request contains the URL of the requested information file. In step 408, the feeder 210 extracts a previously derived inquiry number corresponding to the file. In step 409, the feeder uses this inquiry number to
The requested information file is queried to the central file server 230 using a standard SQL query. In step 410, the feeder receives the cached information file from central file server 230 as a response. And
In the next step 411, the requested and cached Internet information file is retrieved from the feeder 210 by the requesting local cache server 20.
0 is transmitted.

Hereinafter, the operation performed by the updating unit 240 in the embodiment shown in FIG. 2 will be described with reference to the flowchart in FIG.

In step 500, the updating unit 240 receives an instruction from the feeder 210 indicating that a specific file is to be requested. The file in question has been previously requested by the local cache server 200, but the feeder finds that the central cache server 230 did not cache the file. This instruction includes the URL of the file in addition to the address of the local cache server 200 that has requested the file from the central cache server 230. Thereafter, in step 501, the updater will check the requested file for instructions for a list of known non-cacheable files. If the list contains the requested file, the instructions will be discarded. If the list does not include the requested file, the indication will be retained by updater 240, so that it will be time for local cache server 200 to retrieve the file from its original file.

At a convenient time for the central file server 230, ie, when the load on the central server is relatively low, the central server may send a message indicating that any pending instructions should be performed. The message is transmitted to the updating unit 240, and the reception of this message by the updating unit 240 is displayed in step 502. In the next step 503, execution of the instruction starts, and the update unit requests a copy of the file. This copy is
It should be retrieved locally from the local cache server 200 from which the file request originates and cached. Thereafter, in step 504,
A copy of the file is received from a local cache server. Step 50
At 5, a copy of the received file is transmitted to the central file server 230,
This is cached. In the last step 506, the updating unit 240
The feeder 210 is contacted with the query number corresponding to the file cached in the central file server 230 by indexing the hash table 2
Instruct to add to 15.

The operation of the central file server 230 is simple. Basically, it does two things. One is that it answers SQL queries from feeder 210 by transmitting them a cache file, and the other is that it stores a new information file in its cache, Is transmitted to it from the updating unit 240.

Hereinafter, another exemplary embodiment of the Internet cash system according to the present invention will be described with reference to FIG. In FIG. 6, the system differs from that shown in FIG. 2 in that the system has more than one central file server, where three central cache servers 630 are illustrated. FIG. 6 also includes two feeders 610, each connected to its own set in local cache server 600. Feeder 610 and update unit 6
40 is located at the central cache site 690 along with the central file server 630. An updating unit 640 and each feeder 610 are connected to all central file servers 630 by an Ethernet network 680 located in the central cache site.

The more central file servers in this embodiment have more files cached and more SQL compared to the embodiment shown in FIG.
Enables queries to be answered by a central file server. The system is completely scalable, so theoretically any number of feeders, updates or central file servers can be added to the system.

The basic difference between the operation of the system shown in FIG. 6 and that of the system shown in FIG. 2 is that the feeder 610 selects one server from a plurality of central file servers 630 to which an SQL query is to be made. That is what you need to do. Each central file server 630 caches the information file in the original host name within a predetermined range. Therefore, one selection of the central file server is performed as part of an ICP query or HTTP request based on the host name contained in the URL received from the local server. When one of the central file servers is selected by the feeder, an SQL query with the derived query number is made to the selected file server.

It is understood that the configuration and function of each element described with reference to the drawings will be apparent to those skilled in the art.

Although the present invention has been described in terms of particular exemplary embodiments, many different alternatives, modifications, and the like will be apparent to those skilled in the art. Therefore, the above embodiments are not intended to limit the scope of the present invention, and the present invention should be determined by the appended claims.

[Brief description of the drawings]

FIG. 1 schematically shows one embodiment of the Internet cash system of the present invention.

FIG. 2 schematically shows another embodiment of the Internet cash system of the present invention.

FIG. 3 schematically shows a flowchart of an operation performed by the local cache server of FIG. 2;

FIG. 4 schematically shows a flowchart of an operation performed by the feeder of FIG. 2;

FIG. 5 schematically shows a flowchart of an operation performed by the updating unit in FIG. 2;

FIG. 6 schematically shows still another embodiment of the Internet cache system of the present invention.

[Explanation of symbols]

 Reference Signs List 100 Local Cache Server 110 Feeder 115 Indexed Table 116 RAM (Random Access Memory) 120 Microprocessor 130 Central File Server

──────────────────────────────────────────────────の Continuation of front page (81) Designated country EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE ), AU, BR, CA, CN, CZ, HU, ID, IL, IN, IS, JP, KR, LT, LV, MX, NZ, PL, RO, RU, SG, TR, US

Claims (46)

[Claims] 1. Receiving a user request for an Internet information file from a user at a local Internet cache server; and responding to the received request if the information file is not cached by the local server. Querying the information file, and generating a file request for the information file in response to a reply to the query, wherein the central file server storing the cached Internet information file contains the information. If indicating that the file is cached, the file request is sent to feeder means; and in response to the file request, the information is sent from the feeder means to the central file server. Querying an information file, thereby reducing the load on said central file server, wherein the Internet caching system serves the request for an Internet information file. 2. The method of claim 1, wherein the query is made by the local cache server according to a protocol used for communication between Internet cache servers. 3. The Internet cache protocol (I)
3. The method according to claim 2, which is CP). 4. The method according to claim 1, wherein the protocol is a cache digest.
3. The method of claim 2, wherein 5. The method according to claim 1, wherein the inquiry is sent by the local cache server to the feeder means, and the feeder means returns the reply as a response. 6. The method according to claim 5, comprising the step of deriving, at said feeder means, an inquiry number corresponding to said information file concerning said inquiry. 7. The method of claim 6, wherein the step of querying uses the derived query number when querying the central file server for the information file. 8. The method of claim 6, wherein the query provides an alphanumeric string associated with the information file, wherein the alphanumeric string is used in the step of deriving the query number. 9. The method according to claim 8, wherein the alphanumeric string is a Uniform Resource Locator (URL), and the inquiry number is derived from the URL and at least a part of a header information field of the inquiry. Method. 10. The file request providing an alphanumeric string associated with the information file, wherein the alphanumeric string is used by the feeder means to derive an inquiry number corresponding to the information file. 5. The method according to any of 2 or 4. 11. The method according to claim 10, wherein the alphanumeric string is a Uniform Resource Locator (URL), and the inquiry number is derived from the URL and at least a part of a header information field of the file request. the method of. 12. The method of claim 1, further comprising the step of generating an indexed table having an entry for each of the Internet information files cached at the central file server. 13. A search for the information file in the indexed table, the reply to the query indicating whether the information file was found during the search. 13. The method according to 12. 14. The method according to claim 1, wherein the querying step uses a Structured Query Language when querying the central file server for the information file. 15. The step of inquiring comprises: selecting one central file server from a set of central file servers based on an original host name or an IP address of the information file; Each configured to cache an Internet information file with an original host name or IP address within a predetermined range, and querying the selected central file server for the information file. 15. The method according to any one of 1 to 14. 16. The step of inquiring is a step of selecting one central file server from a set of central file servers based on the inquiry number derived for the information file,
Each of said set of servers is configured to cache an Internet information file with a corresponding inquiry number within a predetermined range; and querying said selected central file server for said information file. A method according to any of claims 6 to 14. 17. If the reply to the inquiry indicates that the information file is not cached at the central file server, retrieving the information file from the base server at the local cache server. Caching the information file in the local cache server; requesting a copy of the information file from the local cache server, and caching the copy in the central file server, 17. The method according to any of the preceding claims, further comprising: updating. 18. At least one local cache server and at least one central file server, wherein each server stores cached Internet information files and reduces the load on said central file server. A feeder for communicating with a cache server and the central file server, the feeder receiving a request for an Internet information file from the local cache server; and receiving from the local cache server a request for the Internet information file. Second means for deriving a query from an alphanumeric string; and third means for querying the central file server for the Internet information file using the query derived by the second means. The configuration of the Internet caching system. 19. The arrangement according to claim 18, wherein said first means is configured to operate according to a layer 3 internet protocol. 20. The configuration according to claim 18, wherein the third means is configured to use Structured Query Language (SQL) when inquiring the Internet information file. 21. The configuration according to claim 18, wherein the alphanumeric string is included in the request received from the local cache server. 22. The arrangement of claim 21, wherein the query is derived from the alphanumeric string received from the local cache server and at least a portion of a header information field of the request. 23. The arrangement of claim 22, wherein the query includes a query number, the query number being derived by applying a hash algorithm to the alphanumeric string and the portion of the header information field. . 24. A feeder comprising: a fourth means for receiving a query for an Internet information file from the local cache server; and a fifth means for providing a reply to the received query to the local cache server. The configuration according to any of claims 18 to 20, comprising: 25. The configuration according to claim 24, wherein said fourth means and said fifth means are configured to operate according to a protocol used for communication between Internet cache servers. 26. The Internet cache protocol (26).
26. The configuration of claim 25, wherein the configuration is ICP). 27. The configuration according to claim 24, wherein the alphanumeric string is included in the inquiry received from the local cache server. 28. The method of claim 27, wherein the query derived by the second means is derived from at least a portion of the alphanumeric string received from the local cache server and a header information field of the query. The configuration described. 29. The query of claim 28, wherein the query includes a query number, wherein the query number is derived by applying a hash algorithm to the alphanumeric string and the portion of the header information field. Constitution. 30. The arrangement according to claim 24, wherein the feeder comprises a table having a copy of all indexes of all Internet information files cached at the central file server. 31. The arrangement according to claim 30, wherein the reply to the inquiry received by the fifth means is based on the contents of the table. 32. To further reduce the load on the central file server,
An updating unit for communicating with the local cache server and the central file server, wherein the updating unit requests a copy of the Internet information file stored in the local cache server; The configuration according to any one of claims 18 to 31, comprising: storage means for storing the data in a server. 33. The method according to claim 32, wherein the request unit is configured to request a copy of the information file from the base server when a local cache server storing the information file exists behind a firewall. The configuration described. 34. The arrangement of claim 32 or 33, wherein the updating unit is configured to communicate with the feeder to receive an instruction requesting a copy of the information file. 35. The updating unit includes a list of well-known non-cacheable information files, and no copying is requested for those files.
35. The configuration according to any one of -34. 36. The configuration according to claim 18, wherein the feeder is realized by a low-end computer, and the central file server is realized by a high-end computer. 37. The updating unit is realized by a low-end computer,
4. The central file server is implemented by a high-end computer.
The configuration according to any one of 2 to 35. 38. The configuration according to claim 37, wherein the updating unit and at least one feeder are realized by a single low-end computer. 39. A set of local Internet cache servers, each configured to receive a request for an Internet information file from a user, and at least one of a central cache site for storing a cached Internet information file. Feeder means for interconnecting the central file server with the set of local cache servers and the central file server, according to a protocol used for communication between the Internet cache servers;
At least one means for reducing the load on the central file server by providing means for communicating with at least one local cache server; and means for retrieving Internet information files from the central file server using queries to a database. An internet cash system, comprising: a feeder unit including a feeder. 40. The system of claim 39, wherein said feeder means is included at said central cash site. 41. Each of said feeder means includes a plurality of feeders,
41. The system of claim 39 or 40, wherein each of the feeders interconnects a portion of the set of local cache servers and the central file server. 42. The method of claim 39, wherein the central cache site is configured to serve a predetermined set of local cache servers, each set serving a linguistically and culturally homogeneous user community. The system described in Crab. 43. The method according to claim 39, wherein the protocol used is either an Internet cache protocol or a cache digest.
The system according to any one of the above. 44. The system of claim 39, wherein each of the feeders includes a table having a copy of all indexes of all information files cached at the central cache site. 45. The system of claim 39, wherein said central file server includes a cached Internet information file having an original host name within a predetermined range. 46. The central file server interconnects with at least one local cache server of the predetermined set of local cache servers, and copies an Internet information file from its base server or the at least one local cache server. 46. A system according to any of claims 39 to 45, further comprising updating means for retrieving a copy and storing the copy on the central file server.