CN100514309C - System and method of storing and searching objects through multi-tiered cache memory - Google Patents

Abstract

Systems and methods for improved performance in the storing and retrieving of objects. In one embodiment, the invention comprises a multi-tiered caching system implemented in a network transformation proxy. The proxy performs transformations on Web content received from a Web server and stores the transformed content in the caching system. The lowest tier of caches stores the content as cache objects, while higher tiers store references to lower tiers (such as the tier which stores the objects) as their cache objects. Cache entries are looked up using a plurality of keys. Each of the keys is used to look up an entry in a different tier of the cache.

Description

System and method for storing and retrieving objects using multi-level cache memory

Cross-references to related applications

This application claims priority from the following application: U.S. Provisional Patent Application No. 1, entitled "Multilayer Caching Mechanism for Storing and Retrieving Multiple Versions of Content," filed January 18, 2002 by Jeremy S. de Bonet. .60/349,553, U.S. Provisional Patent Application No. 60/349,424, entitled "Modular Pluggable Transaction Processing Architecture," filed Jan. 18, 2002 by de Bonet et al. U.S. Provisional Patent Application No. 60/349,424, entitled "Network Proxy Platform Supporting Data Transformation, Storage, and Processing of Multiple Protocols Simultaneously," filed January 18, the entire contents of which are hereby incorporated by reference. Also, U.S. Patent Application Serial No. _ (Attorney Docket No. IDET 1130- 1), incorporated herein by reference.

Refer to the appendix

The Appendix, which is incorporated by way of attachment into this application, is hereby incorporated by reference in its entirety for the sole purpose of forming an integral part of this application. Appendix 1 is titled "Network Agent Platform and Its Applications" and includes 17 pages of text and 8 drawings.

technical field

The present invention generally relates to storage and retrieval of electronic entities. In particular, the invention relates to the use of multi-level caches for storing and retrieving objects, wherein groups of objects may be related to each other, such as storing multiple versions of web content on a web transformation agent.

Background technique

There are various methods for storing data. One such approach is through the use of conjugation arrays. In an associative array, the objects to be stored are associated with a key. Stores the object in a special location identified by the relative key. When you want to retrieve an object, you only need to look up the key, because the key identifies the location of the object.

There are various implementations of combining arrays. For example, databases, file systems, and cache memories are all bonded arrays. Caches are of particular interest here.

A cache memory is an associative array that provides local storage of data. "Local" as used here is somewhat relative. "Local" may mean that the cache includes memory fabricated on the same chip as the microprocessor, where caches are connected to microprocessors to allow them to work faster and more efficiently. However, in the case of using the cache for a network proxy, "local" may mean that the cache is implemented in a disk drive within the proxy enclosure.

Caching proxies store and retrieve web content using URLs associated with web pages as respective keys. However, one of the problems that arises in this case is that there may be a large number of different web pages with the same URL. For example, the content of Web pages may be approximately the same, but they may each be adapted for viewing on different types of devices (eg, desktop computers or Internet-enabled cellular phones). This keyword therefore needs to include additional information in order to uniquely identify the Web page that must be retrieved. Keywords can thus be combined with other characteristics of the web page, such as cookies or the type of browser the web page is designed for.

The caches implemented in prior art proxies are typically planar. In other words, there is a single cache with multiple entries. Each cache entry contains a web page associated with the corresponding keyword. As noted above, this key may combine both the URL and other characteristics necessary to uniquely identify the cached content. Therefore, if the proxy needs to store 1000 web pages with different URLs, 1000 cache entries are required. If the proxy server is required to store 10 different versions of each of these Web pages, 10,000 cache memory entries are required.

Because a cache is flat, the time and/or memory it takes to store and retrieve an entry increases with the number of entries. [[Depending on the data structure used, the lookup time can vary from O(n) to O(log(n)), or even to O(1) (constant time), so the following is not exact]] The similarity of the entry does not bring Any benefits (i.e. the fact that dozens of entry points might just be different versions of the same web page)

Furthermore, when using a flat cache structure to store multiple versions of content, there is no way to handle collections of related content. For example, there is no way to store data common to all related content (eg, storing HTTP headers or other information common to multiple versions of the same Web page). Shared information must be stored separately for each individual version. Again, there is no way to handle this collection of related content as a group. For example, if one wants to update each version of an outdated Web page, there is no way to take a single action that affects all versions -- they must be located and updated separately in the cache structure.

It should be noted that while there are multiple layers of storage mechanisms for the database, these are not the same as cache structures. The database is not designed to be used as a library of internal functions for other programs. In database systems, trees and multi-level storage and retrieval structures must be explicitly constructed by the database programmer and, due to the effort, expense, and overhead involved in implementing the database system, this technique is not suitable for high performance cache memory retrieval .

Contents of the invention

One or more of the problems discussed generally above may be addressed by various embodiments of the present invention. Generally, the present invention includes systems and methods for improving storage and retrieval performance of objects. In one embodiment, the present invention includes a multi-level cache system in which multiple keys can be used to look up cache entries. The lowest tiers of a cache system store objects, while higher tiers store references to lower tiers, such as tiers that store objects. Each key is used to find entries located at different levels of the cache system.

An exemplary embodiment is implemented in a web proxy. A network transformation proxy is configured to handle communications between a web server and one or more clients (eg, web browsers). Therefore, the Network Transformation Proxy will work more efficiently if it is configured to cache Web pages that serve clients. In this embodiment, the network transformation agent is configured to store multiple versions of each Web page, where each version corresponds to, for example, a different client device, each client device having its own display features and capabilities. Instead of storing all the different versions of a Web page in a flat cache, a Web page is stored in a multi-level cache. More specifically, web pages are stored in two levels of cache memory, where the URL of the web page is used as the key for the first level entry of the cache memory, and the version of the web page is used as the cache memory (which actually includes multiple cache memory) level 2 entry key. When a client requests a Web page, the URL of the desired page and the client's device type are identified from the request. The Network Transformation Agent indexes the URL as a key into the first level cache. The entry corresponding to this key (URL) contains the object identifying the second level cache. The Network Transformation Agent indexes into the identified Layer 2 cache with the second key (device type). The identified second-level cache entry corresponding to the second key contains the object as the desired Web page. The Web page can then be retrieved and served to the client.

Alternative embodiments include a method of storing and retrieving objects in a multi-level cache. Each object to be stored has multiple keys associated with it. Each key is indexed into a different level of cache memory. Each cache in all tiers, except the cache in the last tier, contains objects that reference caches in the next tier. The caches at the last level store the objects themselves, rather than references to other caches. Alternatively, the last level of cache may contain references to stored objects rather than the objects themselves. Thus, storing an object in a multi-level cache system involves storing an entry in the first level cache containing the first key and a reference to the second level cache, possibly repeating this for additional levels. operation (eg, storing an entry including the second key and a reference to a third-level cache in the second-level cache), and storing the object with the last key in the lowest-level cache. Retrieving an object involves indexing into the first level cache with the first key to get a reference to the second level cache, and repeating this step for other lower levels until the last level is reached, where Indexes into the last level of cache with the last key to retrieve the object.

Another embodiment of the invention includes a software application. The software application is embodied on a computer readable medium such as a floppy disk, CD-ROM, DVD-ROM, RAM, ROM, database schema, or the like. A computer readable medium contains instructions configured to cause a computer to perform the methods outlined above. It should be noted that computer readable media may include RAM or other memory forming part of a computer system. A computer system will therefore be permitted to perform methods according to what is now disclosed and is considered to be within the scope of the appended claims.

Numerous other embodiments are also possible.

Description of drawings

Other objects and advantages of the present invention will become apparent by reading the following detailed description and by referring to the accompanying drawings.

FIG. 1 illustrates an exemplary block diagram of a network-based system according to one embodiment of the present invention.

Figure 2 illustrates a basic configuration diagram of a computer suitable for use as a network transformation agent according to one embodiment of the present invention.

FIG. 3 illustrates a block diagram of a multi-level cache memory according to one embodiment of the present invention.

FIG. 4 is a flowchart illustrating a general method applicable to an N-level cache structure according to one embodiment of the present invention.

While the invention is capable of various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawings and the accompanying detailed description. It should be understood, however, that the drawings and detailed description are not intended to limit the invention to the particular embodiments described. On the contrary, the intention of this disclosure is to cover all modifications, equivalents and alternatives falling within the scope of the invention as defined by the appended claims.

Detailed ways

Preferred embodiments of the present invention will be described below. It should be noted that this and any other embodiments described below are exemplary only, and are intended to illustrate the invention rather than limit it.

Generally, the present invention includes systems and methods for improving the performance of storing and retrieving objects. In one embodiment, the present invention includes a multi-level cache system in which multiple keys are used to query cache entries. The lowest level of the cache memory stores objects, while higher levels store references to lower levels such as those storing objects. Each key is used to query entries at different levels of the cache system.

An exemplary embodiment is implemented in a network transformation agent. A network transformation proxy is configured to intercept and manipulate communications between a web server and one or more clients (eg, web browsers). Therefore, the Network Transformation Proxy cache will work more efficiently if it is configured to serve the client's Web pages. In this embodiment, the network transformation agent is configured to generate multiple versions of each Web page, where each version corresponds to, for example, a different optimization for each type of client device, each client device having its own display features and performance. Instead of storing all the different versions of a web page in one flat cache, storing the versions of a web page in a multi-level cache. More specifically, Web pages are stored in two levels of cache memory, where the URL of the Web page is used as the key into the first level of cache memory, and the device for which the page has been transformed is used as the key into the second level of cache memory. Second level keywords.

When a client requests a Web page, the URL of the desired page and the client's device type can be identified from the request. The Network Transformation Agent uses the URL as a key to index into the first level cache memory. The entry corresponding to this key (URL) contains an object that is an identifier of the cache memory of the second layer. The network transformation server indexes into the cache memory identified in the second tier with the second key (device type). The identified second-level cache entry corresponding to the second key contains the object that is the desired Web page. The Web page can then be retrieved and provided to the client.

Although the preferred embodiment is implemented at the network translation proxy, it should be noted that the present invention can be applied generally to multi-level caches of a variety of systems. Therefore, even though the present disclosure mainly focuses on the implementation of the present invention in the network conversion proxy, such purpose is only for illustration, not limitation.

The preferred embodiment of the present invention works in a network environment. A network and its components are used to distribute web content (eg, web pages) from one or more servers to one or more clients. An exemplary structure is shown with reference to FIG. 1 .

As shown in FIG. 1 , the architecture includes a client 12 connected to a network transformation proxy 14 which in turn is connected to a Web server 16 . The network translation agent 14 includes a cache subsystem 18 . Clients 12 are connected to brokers 14 via a first network 13 . The proxy 14 is connected to a web server 16 via a second network 15 . It is contemplated that at least one of network 13 and network 15 includes the Internet. Another of these networks may include a particular enterprise's internal or external network. It should be noted, however, that the connections of client 12, proxy server 14, and Web server 16 need not be configured in any particular way for the purposes of the present invention.

The proxy handles communication between a client device or program such as a Web browser and a server device or program such as a Web server. In Web-based systems, proxies process client requests for Web content, and Web servers provide Web content in response to those requests. In handling these communications, the proxy is responsible for simulating the web server and thus reducing the burden on the system (on the web server and on the network itself). The proxy does this by storing some content provided by the web server, and when possible serving the stored content to the client in response to a request for the content. In this way, the proxy relieves the web server of the burden of serving a portion of the client's requests.

In a preferred embodiment, a network transformation proxy is configured to perform transformations of web content provided by the server. The transformation may depend on the client making the request for the content and the way the request is made. Transformations may include modifications to optimize content for use on particular types of client devices. For example, modifying Web pages to accommodate the capabilities of different display devices (eg, performing color attenuation or black-and-white transformations on images on Web pages). Thus, an agent can generate multiple versions of a particular Web page (or other Web content). Agents can also perform transformations to make more content modifications for clients, such as inserting different advertisements for different clients. The proxy then needs to store these different versions of the web content.

To create and identify different versions of a Web page, the network transformation agent uses information about the type of transformation made on the page and the version provided by the origin server. Version cache keys may also indicate parameter values used in the transformation. For example, if the network transformation agent is color attenuating an image, then the key will include the label of the color to which the image is to be attenuated. Proxies according to the present invention provide a fast and efficient mechanism for storing and retrieving Web content, even though multiple versions of the content increase the total amount of information that must be stored.

Referring to Figure 2, there is shown a basic configuration of a computer suitable for use as a network transformation agent according to one embodiment of the present invention. Server 14 is implemented in computer system 100 . Computer system 100 includes central processing unit (CPU) 112 , read only memory (ROM) 114 , random access memory (RAM) 116 , hard disk drive (HD) 118 , and input-output devices (I/O) 120 . Computer system 100 may have more than one CPU, ROM, random access memory, hard drives, input output devices, or other hardware components thereof. However, computer system 100 is described as having only a single type of component. It should be noted that the system shown in Figure 2 is a simplified illustration of an exemplary hardware configuration and that many other alternative configurations are possible.

Appropriate software applications residing in memory such as ROM 114, RAM 116, or hard drive 118 may perform portions of the methods described herein. A software application may include program instructions configured to cause a data processor executing the instructions to perform the methods described herein. These instructions may be contained (stored) in internal storage devices such as ROM 114, RAM 116 or hard drive 118, other devices, as well as external storage devices, or storage media readable by a data processor such as computer system 100 or even CPU 112. Such media may include, for example, floppy disks, CD-ROMs, DVD ROMs, magnetic tapes, optical storage media, and the like.

In an exemplary embodiment of the present invention, the computer executable instructions may be several lines of compiled C ⁺⁺ , Java or other language codes. Other configurations may also be used. For example, the functions of any one computer can be performed by different computers shown in FIG. 2 . In addition, a computer program carrying such code, or software components thereof, may also be stored on readable media of more than one computer and more than one data processing system.

In the above hardware configurations, the various software components may reside on a single computer or any combination of separate computers. In alternative embodiments, some or all of the software components may reside on the same computer. For example, one or more software components of agent computer 100 may reside on a client computer or server computer or both. In yet another embodiment, the proxy computer itself may not be required if the functionality implemented by the proxy computer is incorporated into the client computer or server computer. In such an embodiment, the client computer and the server computer may be directed to connect to the same network.

Communications between any client, server and proxy computers may be accomplished electronically, optically, radio frequency or other signals. For example, when a user is at a client computer, the client computer may convert signals into human understandable form when sending communications to the user, and may convert human input into appropriate electrical, optical, radio frequency or other signals. Likewise, when the operator is at the server computer, the server computer may convert the signal into a human understandable form when communicating to the operator, and may convert human input into the appropriate electrical, optical, radio frequency, or other Signal.

As explained above, the proxy is responsible for storing information previously provided by the web server so that this information can be provided to clients in response to their requests. This information is stored in the proxy's cache subsystem. The cache subsystem actually includes a large number of cache memories organized in two or more layers. The upper and middle tiers refer to the cache memory of the lower tiers. The lowest layer actually stores the desired information. Information stored in the cache subsystem is accessed by accessing each of these layers in turn.

Referring to FIG. 3 , it shows the structure of a multi-level cache storage system according to an embodiment of the present invention. As shown, the cache memory subsystem 18 includes a first tier 22 and a second tier 24 . The first level 22 actually includes a single cache memory 30 with multiple entries (eg 31-33). The second tier 24 includes a plurality of cache memories (eg, 40, 50, 60).

Each entry of each cache consists of a key and an object. The key is used to identify the desired entry into the cache memory. Objects are desired information stored in cache memory. Since the cache storage subsystem 18 is designed to store Web content, the keys of the first level cache memory 26 entries (eg, 35 ) include URLs (Uniform Resource Locators). The entry object (eg 36 ) includes a reference to the cache memory of the second tier 24 . For example, the object 36 of the entry 31 of the first level cache is a reference to the second level cache 40 .

Cache 40 (like other second-level caches) is very similar in that each entry (eg, 41, 42, 43) includes a key and an object. However, since cache 40 is at the lowest level of the cache system structure, its cache entries contain objects including objects of the type that cache subsystem 18 is designed to store (eg, Web pages). If the cache subsystem 18 has more than two levels, the objects contained in the cache (eg, 40) at the second level 24 will include references to the third level cache. This third tier may be the lowest tier, or still another intermediate tier, where cached objects include references to caches of subsequent tiers. Therefore, the structure can be extended to any number N layers.

The flowchart of Figure 4 summarizes the method employed in retrieving stored objects using this cache structure. The flowchart of Figure 4 represents a general method applicable to an N-level cache structure. As shown, the first key is used to index into the first structure to retrieve a reference to the second level cache. This operation is repeated according to the number N of levels in the cache memory structure. For the last level (level N), the Nth key is used to index into the cache to retrieve the stored object.

In the case of the preferred embodiment, where the network translation agent implements the cache subsystem, the keys of the first level cache 26 include the names (eg, URLs) of the respective content stored there. Objects in the first tier cache 26 include references to the second tier 24 cache. The keys in the second level cache are based on parameters specifying different versions of the content identified by the URL. Objects in the second tier cache include Web content stored by the cache subsystem 18 . Keys in the first and second level caches in combination can be used to store or retrieve any version of any piece of content stored by cache subsystem 18 .

In a simple embodiment of the invention, two different caches, or as many caches as are required, each use a key to store a value. Each cache memory functions similarly. A cache may include any cache or related memory structure.

One embodiment of the present invention is a multi-level cache system. In a simple embodiment, the present invention uses a two-level cache system. At the first level, a key based on the content name is used to identify one of many secondary caches. In the second cache memory, the particular version of the content specified by the first-level key is identified by a key encapsulating the version-specific information. The program can be expressed as follows:

Level_1_Cache:=CacheOf<

f(Content_Name),

Level_2_Cache

>

      Level_2_Cache:=CacheOf<

g(Description_Of_Content_Version),

Content_Version

>

The abstract functions f() and g() transform their arguments into concise and easy-to-match keywords. In the preferred embodiment, MD5sum is used, but any (approximately) unique encoding will be used. CacheOf<Key, Content> is a cache data structure that stores and retrieves content by associating it with a key.

The first level cache (Level_1_Cache) is a cache of cache memory in which the second level cache (Level_2_Cache) is stored and retrieved with a key based on a content name. The second level of cache is a standard cache that associates content version descriptions with the appropriate content version. It is worth noting that the keys of the second-level cache do not necessarily encapsulate the content name, because all entries in each second-level cache are different versions of the same content-more specifically, as determined by the Content_Name identified content.

A second level cache storing keys including versions and values including Web pages is similar to prior art in the art. However, the first level cache storing the key including the URL and the value including the second cache is unique in storing the cache.

In a preferred embodiment, the first level cache (Level_1_Cache) is based on the name of the content (eg URL) and it points to the second level cache. The second level of caching (corresponding to Level_2_Cache) is based on the type and parameter settings of the transformations that have been applied to the content.

In the preferred embodiment, the MD5Sum of the content URL is used to provide the key for the first level cache (corresponding to the Level_1_Cache described above for the simple embodiment). The MD5 algorithm developed by Professor Ronald L. Rivest of MIT is a commonly used one-way function, which has the property that two unequal inputs are extremely unlikely to produce the same result. This algorithm makes it more efficient to identify the keywords of the data.

The second level cache contains multiple versions of the content identified by the URL. Proxies create these diverse versions by performing parameter transformations on the web content. MD5Sum and its parameters provide keys for the second-level cache based on the transform name. Each keyword identifies the version of the transformed content with those settings.

Using the MD5sum function, this structure can be described as:

     Level_1_Cache:=OneCacheOf<

MD5Sum(URL),

Level_2_Cache

>

       Level_2_Cache:=OneCacheOf<

MD5Sum(Transformation_Parameters),

       Transform(Content, Transformation_Parameters)

>

The preferred embodiment uses C ⁺⁺ templates to create caches within a multi-level cache structure. C ⁺⁺ templates make it unnecessary to write separate bodies of code to accomplish the same task. It abstracts tasks, allowing one C ⁺⁺ object to perform multiple tasks. To accomplish a specific task, any type of keyword and value can be assigned to the template. In the case of the present system, the C ⁺⁺ templates make it unnecessary to write two separate bodies of code for the first and second level caches. Key and value types for two different caches can be loaded simultaneously into the same C ⁺⁺ template structure. Exemplary systems and methods for using C ⁺⁺ templates in this manner are described in January 16, 2003 by Jeremy S. de Bonet, Todd A. Stiers, Jeffrey R. Annison, Phillip Alvelda VLL, and Paul M. Scanlan Described in detail in US Patent Application No._ (Attorney Docket No. IDET1150-1) entitled "Design for Storage and Retrieval of Arbitrary Content and Application Data" filed on .

Use multiple levels of caching to make cache lookups more efficient. For example, a caching proxy may store 10 separate versions of each of 1000 URLs. With the current invention, there would be no need to store URLs as 10,000 separate entities with 10,000 separate keywords. Instead, URLs can be stored as 1,000 separate entities using 1,000 separate keywords. When looking for a specific version of a specific page, the proxy only needs to search 1,000 URLs, and then search 10 versions of that URL. This lookup only requires searching 1,010 unique entities instead of 10,000.

Furthermore, the present invention creates a method of storing data common to all content. For example, the date the content was created or various other HTTP headers may be common to all versions (as in the case of changing proxies), and the present invention provides a common place to store such information. It is not necessary to store this information separately for each version of the content, and if there are any changes, these changes can be made to multiple versions of the content at the same time.

Also, since all versions of a Web page are stored together in a single cache system, developers can manipulate, dump, or delete them all together without having to identify each one individually.

Developers can extend the cache storage system beyond two tiers. Additional keys that may be used to index into the caches of these tiers may include other identifiers, such as device type, browser, or payment level.

As the differences between web customers increase, it becomes necessary to create multiple versions of content, each optimized for a different type of customer. Prior to this invention, there was no way to organize such multi-versioned content into a single unified cache structure.

The benefits and advantages which may be provided by the present invention have been described above with reference to specific examples. None of these benefits and advantages, nor any element or limitation which causes or makes them more specific, should be construed as an essential, necessary or essential feature of any or all of the claims. As used herein, the phrases "comprises," "comprises," or any other variation thereof are intended to be construed as non-exclusive inclusions of the elements or limitations that follow those phrases. Accordingly, a system, method, or other embodiment comprising a set of elements is not limited to only those elements, but may include other elements not expressly listed or inherent to the described embodiment.

While the invention has been described with reference to specific embodiments, it should be understood that the embodiments are illustrative and do not limit the scope of the invention to these embodiments. Many variations, modifications, additions or improvements are also possible to the embodiments described above. It is intended that such changes, modifications, additions or improvements will fall within the scope of the present invention as detailed in the following claims.

Claims (14)

1. system that is used to store and retrieve the object of the first kind comprises:

A plurality of cache memories, wherein each inlet at each cache memory comprises a key word and an object;

Have a plurality of (n+1) floor height speed cache memory in wherein said a plurality of cache memory altogether, wherein comprise the object of a first kind at each object of each inlet of (n+1) floor height speed cache memory; And

Further have a n floor height speed cache memory in wherein said a plurality of cache memory, wherein comprise the identifier of correspondence (n+1) floor height speed cache memory at each object of each inlet of n floor height speed cache memory,

Wherein, the key word in described n floor height speed cache memory comprises URL(uniform resource locator);

Wherein, the key word in described (n+1) floor height speed cache memory comprises the version identifier of corresponding uniform resource locator, and the object in (n+1) floor height speed cache memory comprises the Web page.

2. system according to claim 1, wherein said system comprise a Web page cache memory system of realizing on network transformation proxy; Wherein said n floor height speed cache memory comprises the ground floor cache memory, wherein the key word in the ground floor cache memory comprises URL(uniform resource locator), and wherein the object in the ground floor cache memory comprises the identifier of second layer cache memory; Wherein (n+1) floor height speed cache memory comprises second layer cache memory, wherein the key word in second layer cache memory comprises the version of corresponding uniform resource locator, and wherein the object in second layer cache memory comprises the Web page.

3. system according to claim 2, wherein said agency be configured to the Web page carry out one or more conversion and with the Web page stores of institute's conversion in (n+1) floor height speed cache memory.

4. system according to claim 1, wherein said system comprise at least one deck cache memory in addition, and wherein the caches object in each except that lowermost layer layer is all quoted cache memory in the lower level.

5. system according to claim 1, wherein (n+1) floor height speed cache memory is configured to be stored in the common shared information of all inlets in the corresponding cache memory.

6. system according to claim 1, wherein each (n+1) floor height speed cache memory is configured to store the multiple version corresponding to the Web page of single URL(uniform resource locator).

7. system according to claim 1, wherein said system are configured to allow so that all inlets in selected (n+1) floor height speed cache memory are carried out simultaneously with a kind of mode of operation they are handled.

8. a method that is used to retrieve desired object wherein has a plurality of key words relevant with this expectation object, and this method comprises:

Identification and the relevant a plurality of key words of expectation object;

N the inlet of n keyword search in n floor height speed cache memory in the key word that passes through to be discerned, wherein said n inlet comprises n key word and n the object in the key word of being discerned, and wherein said n object comprises the identifier of (n+1) floor height speed cache memory; And

Search (n+1) the individual inlet in (n+1) floor height speed cache memory, wherein said (n+1) individual inlet comprises (n+1) individual key word and (n+1) individual object of the key word of being discerned, wherein said (n+1) individual object comprises this desired object

Wherein, described n key word comprises URL(uniform resource locator);

Wherein, described (n+1) individual object comprises the Web page, and wherein, described (n+1) individual key word comprises the version identifier of the corresponding uniform resource locator and the described corresponding uniform resource locator of the described Web page.

9. method according to claim 8, wherein said method comprises by described URL(uniform resource locator) searches first inlet in the ground floor cache memory, and wherein said first inlet comprises as the URL(uniform resource locator) of corresponding key word with as the identifier of the second layer cache memory of corresponding objects; And search second inlet in second layer cache memory, wherein said second inlet comprises as the version of the URL(uniform resource locator) of corresponding key word and as the expectation Web page of corresponding objects.

10. method according to claim 8, wherein said cache memory comprise having more than two-layer cache memory system.

11. method according to claim 8, wherein (n+1) floor height speed cache memory is configured to store the common shared information of all inlets in (n+1) floor height speed cache memory.

12. method according to claim 8, wherein (n+1) floor height speed cache memory is configured to store the multiple version corresponding to the Web page of single URL(uniform resource locator).

13. method according to claim 8 further comprises so that all inlets in (n+1) floor height speed cache memory are carried out simultaneously with a kind of mode of operation they is handled.

14. method according to claim 8 further comprises web content is carried out one or more conversion and the web content after the conversion is deposited in (n+1) floor height speed cache memory.

Images